id stringlengths 4 10 | text stringlengths 4 2.14M | source stringclasses 2
values | created timestamp[s]date 2001-05-16 21:05:09 2025-01-01 03:38:30 | added stringdate 2025-04-01 04:05:38 2025-04-01 07:14:06 | metadata dict |
|---|---|---|---|---|---|
143795936 | Resolved Issue #4 in upstream
Added escape rule.
@bkarthikeyan28 Nice fix. Thanks :)
| gharchive/pull-request | 2016-03-27T10:36:42 | 2025-04-01T06:45:52.001104 | {
"authors": [
"bkarthikeyan28",
"ssundarraj"
],
"repo": "ssundarraj/commander",
"url": "https://github.com/ssundarraj/commander/pull/6",
"license": "apache-2.0",
"license_type": "permissive",
"license_source": "bigquery"
} |
782168804 | Problem with whitespaces
I have encountered the following issue:
When translating from Russian to Ukrainian, the resulting text lacks whitespaces at all. All words are translated correctly, but they are all merged as a single word.
So in order to translate from Russian to Ukrainian, i have to do RU -> EN -> UK translation.
The version of googletrans is googletrans==4.0.0rc1
It happens for CA -> ES too.
The workaround works like a charm! CA -> EN -> ES
Thanks @cryptoqube777 :)
It happens for CA -> ES too.
The workaround works like a charm! CA -> EN -> ES
Thanks @cryptoqube777 :)
I am using workaround:
in text replace "spaces" to then in translated text replace back to "spaces".
Although in translated text " " may be slightly modified, pay attention.
I am using workaround:
in text replace "spaces" to then in translated text replace back to "spaces".
Although in translated text " " may be slightly modified, pay attention.
Is there a clean workaround? The ones specified here don't really return the correct text
| gharchive/issue | 2021-01-08T14:46:24 | 2025-04-01T06:45:52.005376 | {
"authors": [
"SeerBird",
"cryptoqube777",
"igor-kremin",
"sod1497"
],
"repo": "ssut/py-googletrans",
"url": "https://github.com/ssut/py-googletrans/issues/269",
"license": "mit",
"license_type": "permissive",
"license_source": "bigquery"
} |
898898473 | All translation sources default to 'en' [Bug]
Googletrans version:
[ ] 4.0.0rc1
[x] 3.1.0a0
[ ] 3.0.0
[ ] 2.x
I'm submitting a ...
[x] bug report
[ ] feature request
Current behavior:
No matter the specified src, the returned translation specifies src=en.
Expected behavior:
The src should be whatever I specify it as. For example, src=tr should be a translation from Turkish.
Steps to reproduce:
test = translator.translate('merhaba', src='tr', dest='ar')
print(test)
Returns
Translated(src=en, dest=ar, text=merhaba, pronunciation=merhaba, extra_data="{'translat...")
Related code:
insert short code snippets here
See above
Other information:
I had the same problem. Upon investigation I noticed that I got a 429: Too many requests response instead of a 200. Is there a limit on requests that one can make?
I had the same problem. Upon investigation I noticed that I got a 429: Too many requests response instead of a 200. Is there a limit on requests that one can make?
Think this is the same issue: [https://github.com/ssut/py-googletrans/issues/121]
I have the same problem but it seems to only occur with large texts. I don't seem to get the 429: Too many requests
Apparently the length limit is 5000
| gharchive/issue | 2021-05-22T23:16:50 | 2025-04-01T06:45:52.012913 | {
"authors": [
"FurkanToprak",
"SeerBird",
"albertki",
"thomasmutter"
],
"repo": "ssut/py-googletrans",
"url": "https://github.com/ssut/py-googletrans/issues/292",
"license": "mit",
"license_type": "permissive",
"license_source": "bigquery"
} |
1870269107 | next parameter is missing from response when sortby is used in a query
The next parameter is missing in the links[0].bodyof the response when sortby is used in a query. I thought pystac-client had the bug, but it's clear what's happening when using postman. This causes clients to either loop forever or keep sending the same query for the first page again and again.
Should the fallback here be to use page on the client side, risking duplicate object being ingested while paging?
Related Issues:
#242 (see comments), #332, #354
Ways to reproduce this bug:
Easiest way to reproduce this is using stac-browser
Using cURL
This works, and you get a next token.
--header 'Content-Type: application/json' \
--data '{
"limit": 1
}'
Response:
{
"type": "FeatureCollection",
"stac_version": "1.0.0",
"stac_extensions": [],
"context": {
"limit": 1,
"matched": 61610747,
"returned": 1
},
"numberMatched": 61610747,
"numberReturned": 1,
"features": [
{
...
}
],
"links": [
{
"rel": "next",
"title": "Next page of Items",
"method": "POST",
"type": "application/geo+json",
"href": "https://earth-search.aws.element84.com/v1/search",
"merge": false,
"body": {
"limit": 1,
"next": "2023-08-25T23:36:41.463517Z,S1A_IW_GRDH_1SDV_20230825T233627_20230825T233654_050038_060542,sentinel-1-grd"
}
},
{
"rel": "root",
"type": "application/json",
"href": "https://earth-search.aws.element84.com/v1"
}
]
}
And this doesn’t…
curl --location 'https://earth-search.aws.element84.com/v1/search' \
--header 'Content-Type: application/json' \
--data '{
"sortby": "-id",
"limit": 1
}'
Response:
{
"type": "FeatureCollection",
"stac_version": "1.0.0",
"stac_extensions": [],
"context": {
"limit": 1,
"matched": 61610832,
"returned": 1
},
"numberMatched": 61610832,
"numberReturned": 1,
"features": [
{
...
}
],
"links": [
{
"rel": "next",
"title": "Next page of Items",
"method": "POST",
"type": "application/geo+json",
"href": "https://earth-search.aws.element84.com/v1/search",
"merge": false,
"body": {
"sortby": [
{
"field": "id",
"direction": "desc"
}
],
"limit": 1
}
},
{
"rel": "root",
"type": "application/json",
"href": "https://earth-search.aws.element84.com/v1"
}
]
}
Nothing in the links[0].body
This just hangs and keeps pinging the API for the next href and body (w/o a next page) forever.
stac-client search https://earth-search.aws.element84.com/v1/ -c sentinel-2-l2a --sortby properties.updated
CC: @philvarner, @dmcvicar
fix released in 3.5.0, will be deployed to Earth Search next week
deployed to Earth Search.
| gharchive/issue | 2023-08-28T18:43:16 | 2025-04-01T06:45:52.027473 | {
"authors": [
"philvarner",
"samsipe"
],
"repo": "stac-utils/stac-server",
"url": "https://github.com/stac-utils/stac-server/issues/589",
"license": "MIT",
"license_type": "permissive",
"license_source": "github-api"
} |
2488523485 | Generate SBOMs during build
Needed for https://github.com/stackabletech/issues/issues/614
This enables SBOM generation at build time for:
All Airflow versions
Druid 30.0.0
Hadoop 3.3.6 and 3.4.0
HBase 2.4.18
hbase-operator-tools 1.2.0
Phoenix 5.2.0
Hive 3.1.3
NiFi 1.27
Omid 1.1.2
Spark 3.5.1
All Superset versions
Trino 451
jmx-exporter 1.0.1
Zookeeper 3.9.2
I did not bother to create patches for non-LTS versions yet. We could extend that later if we need it. But currently we want to focus on fixing vulnerabilities in our LTS versions.
Products where SBOM generation did not work yet:
Kafka: Problem with the Gradle CycloneDX plugin and Scala (CycloneDX/cyclonedx-gradle-plugin#239)
OPA / statsd_exporter: cyclonedx-gomod expects the source code to be in a Git repo and obtains the version from the Git tag. Solutions: a) Patch cyclonedx-gomod b) Create a fake Git repo c) Clone the source code from Git instead of mirroring it in Nexus
We can look into fixing that at a later point in time. The main value of build time SBOM generation is currently that it helps detecting how a dependency is brought into an image. Since Kafka and the Go-based products are not heavily affected and the fixes for both did not seem to be trivial, I figured that we could skip them for now since SBOM generation for these products won't provide too much value anyway at the moment. We can improve this later, we'll have to adapt SBOM generation continuously for all upcoming product versions anyway.
Another PR related to this one will follow, since this one only places an SBOM as a file inside the image. The other one will enable using this SBOM file to extend our actual published SBOMs with the dependency information (those who are attached to images and shown e.g. in our SBOM browser).
I tested all builds successfully on my local machine and verified that the SBOMs are present.
I'd like to review this next week. If it has time I'd be grateful for that but this is no block or veto. I'll let you decide.
I think the builds for product versions where I did not create a patch file that generates the SBOM would fail right now. We can discuss next week if we want to create patch files for those versions as well or make the "move SBOM" steps in the Dockerfiles optional ("only move it if the file exists").
@lfrancke I fixed the issues mentioned in your comments and added patches for all product versions we currently build.
| gharchive/pull-request | 2024-08-27T07:28:07 | 2025-04-01T06:45:52.037415 | {
"authors": [
"dervoeti",
"lfrancke"
],
"repo": "stackabletech/docker-images",
"url": "https://github.com/stackabletech/docker-images/pull/814",
"license": "Apache-2.0",
"license_type": "permissive",
"license_source": "github-api"
} |
2170483620 | Change poll icon color when active (#680)
Very simple. Well done!
| gharchive/pull-request | 2024-03-06T02:07:26 | 2025-04-01T06:45:52.051159 | {
"authors": [
"huumn",
"jp-melanson"
],
"repo": "stackernews/stacker.news",
"url": "https://github.com/stackernews/stacker.news/pull/898",
"license": "MIT",
"license_type": "permissive",
"license_source": "github-api"
} |
1234065257 | Refactor upstream sync workflow
The upstream-sync.yml workflow has gone through a significant rewrite with the aim of removing third-party dependencies, making it easier to maintain and extend. Also this workflow now has the ability to create a release series branch if one does not already exist, for example stackhpc/yoga.
@m-bull, all requested changes have been made.
| gharchive/pull-request | 2022-05-12T14:24:19 | 2025-04-01T06:45:52.053214 | {
"authors": [
"jackhodgkiss"
],
"repo": "stackhpc/.github",
"url": "https://github.com/stackhpc/.github/pull/10",
"license": "Apache-2.0",
"license_type": "permissive",
"license_source": "github-api"
} |
2233698678 | ROX-23526: access insecure registry
Description
Scanner Skips TLS when fetching the image's manifest and image layer if Scanner GRPC client says so
Checklist
[x] Investigated and inspected CI test results
[x] Unit test and regression tests added
If any of these don't apply, please comment below.
Testing Performed
Deploy scanner v4 via helm:
helm upgrade --install -n stackrox stackrox-central-services --create-namespace ./stackrox-central-services-chart -f "$vals_dir/central_values3.yaml" -f "$vals_dir/secret_values.yaml"
2.Add uncertified registry and scan
oc new-project dave
# Create BuildConfig:
oc apply -f - << EOF
kind: BuildConfig
apiVersion: build.openshift.io/v1
metadata:
name: dave-bc
labels:
build: dave-bc
spec:
source:
type: Binary
binary: {}
strategy:
type: Docker
dockerStrategy: {}
output:
to:
kind: ImageStreamTag
name: 'dave-is:latest'
EOF
# Create ImageStream
oc apply -f - <<EOF
kind: ImageStream
apiVersion: image.openshift.io/v1
metadata:
name: dave-is
spec:
lookupPolicy:
local: false
EOF
# Build an Image:
cat > Dockerfile <<EOF
FROM nginx:1.24.0
EOF
oc start-build dave-bc -F --from-dir=.
# Create Image Integration
## first get user/pass
SEC_NAME=$(oc get secrets -o custom-columns=":metadata.name" --no-headers | grep "^default-docker")
oc get secrets $SEC_NAME -o json | jq -r '.data.".dockercfg"' | base64 -d | jq -r '."image-registry.openshift-image-registry.svc:5000" | "== User:\n\(.username)\n\n==Pass:\n\(.password)"'
## use the output above to create a docker image integration in ACS, will have to skip TLS verification
# Scan via roxctl
roxctl image scan --force --image=image-registry.openshift-image-registry.svc:5000/dave/dave-is:latest
Result:
{
"id": "sha256:cce7842f619c9d37ef13e8c6a59800d62fdd4a365367483682d3bfb10290c1af",
"name": {
"registry": "image-registry.openshift-image-registry.svc:5000",
"remote": "dave/dave-is",
"tag": "latest",
"fullName": "image-registry.openshift-image-registry.svc:5000/dave/dave-is:latest"
},
"names": [
{
"registry": "image-registry.openshift-image-registry.svc:5000",
"remote": "dave/dave-is",
"tag": "latest",
"fullName": "image-registry.openshift-image-registry.svc:5000/dave/dave-is:latest"
}
],
"metadata": {
"v1": {
"digest": "sha256:6ae9d92c3fa8938f62472baddcedea745b4307b39ad984d944e614742096f8df",
"created": "2024-04-24T14:02:56.862058723Z",
"layers": [
{
"instruction": "ADD",
"value": "file:5d6b639e8b6bcc01149b7486502558088f9816200063ca72b91a1f989bc8d85e in /",
"created": "2023-04-11T19:57:20Z"
},
{
"instruction": "CMD",
"value": "[\"bash\"]",
"created": "2023-04-11T19:57:20Z",
"empty": true
},
{
"instruction": "LABEL",
"value": "maintainer=NGINX Docker Maintainers \u003cdocker-maint@nginx.com\u003e",
"created": "2023-04-11T19:57:20Z",
"empty": true
},
{
"instruction": "ENV",
"value": "NGINX_VERSION=1.24.0",
"created": "2023-04-11T19:57:20Z",
"empty": true
},
{
"instruction": "ENV",
"value": "NJS_VERSION=0.7.12",
"created": "2023-04-11T19:57:20Z",
"empty": true
},
{
"instruction": "ENV",
"value": "PKG_RELEASE=1~bullseye",
"created": "2023-04-11T19:57:20Z",
"empty": true
},
{
"instruction": "RUN",
...
{
"instruction": "COPY",
"value": "20-envsubst-on-templates.sh /docker-entrypoint.d # buildkit",
"created": "2023-04-11T19:57:20Z"
},
{
"instruction": "COPY",
"value": "30-tune-worker-processes.sh /docker-entrypoint.d # buildkit",
"created": "2023-04-11T19:57:20Z"
},
{
"instruction": "ENTRYPOINT",
"value": "[\"/docker-entrypoint.sh\"]",
"created": "2023-04-11T19:57:20Z",
"empty": true
},
{
"instruction": "EXPOSE",
"value": "map[80/tcp:{}]",
"created": "2023-04-11T19:57:20Z",
"empty": true
},
{
"instruction": "STOPSIGNAL",
"value": "SIGQUIT",
"created": "2023-04-11T19:57:20Z",
"empty": true
},
{
"instruction": "CMD",
"value": "[\"nginx\" \"-g\" \"daemon off;\"]",
"created": "2023-04-11T19:57:20Z",
"empty": true
},
{
"instruction": "ENV",
"value": "\"OPENSHIFT_BUILD_NAME\"=\"dave-bc-5\" \"OPENSHIFT_BUILD_NAMESPACE\"=\"dave\"",
"created": "2024-04-24T14:02:56.786154736Z",
"empty": true
},
{
"instruction": "LABEL",
"value": "\"io.openshift.build.name\"=\"dave-bc-5\" \"io.openshift.build.namespace\"=\"dave\"",
"created": "2024-04-24T14:02:56.862899023Z",
"empty": true
}
],
"user": "root",
"command": [
"nginx",
"-g",
"daemon off;"
],
"entrypoint": [
"/docker-entrypoint.sh"
],
"labels": {
"io.buildah.version": "1.32.2",
"io.openshift.build.name": "dave-bc-5",
"io.openshift.build.namespace": "dave",
"maintainer": "NGINX Docker Maintainers \u003cdocker-maint@nginx.com\u003e"
}
},
"v2": {
"digest": "sha256:cce7842f619c9d37ef13e8c6a59800d62fdd4a365367483682d3bfb10290c1af"
},
"layerShas": [
"sha256:04e7578caeaa5a84ad5d534aabbb307a37c85f9444b94949d37544a1c69c8f15",
"sha256:57a1056ea48405cbdf64995d14dc4c4435097c6a28cbf941bd7f524eafe011f3",
"sha256:6989106bacf00ee358aaf2e9b291a45c3766142d008d02ffe09be6d8efd3d9a7",
"sha256:c00d1142b33106cf135e0a082f8be11b233ef95209c59a3f956bad6957072750",
"sha256:cb49393af980543dd9fd2b63dfd65e17d9ebd150a7197a6baed3b62d3c090e27",
"sha256:5ac1ebd8aebeead2594dc2a466e7eba1c08395263aad4315fd5ee5f40796abea"
],
"dataSource": {
"id": "1dad6eb6-204a-410a-a2d7-a49aac3964d6",
"name": "ocp2-test2"
}
},
"scan": {
"scanTime": "2024-04-24T14:38:36.790198636Z",
"components": [
{
"name": "dpkg",
"version": "1.20.13",
"layerIndex": 0,
"location": "var/lib/dpkg/status"
},
{
"name": "libbz2-1.0",
"version": "1.0.8-4",
"layerIndex": 0,
"location": "var/lib/dpkg/status"
},
{
"name": "librtmp1",
"version": "2.4+20151223.gitfa8646d.1-2+b2",
"layerIndex": 6,
"location": "var/lib/dpkg/status"
},
{
"name": "nginx-module-xslt",
"version": "1.24.0-1~bullseye",
"layerIndex": 6,
"location": "var/lib/dpkg/status"
},
{
"name": "libintl",
"version": "0.21",
"layerIndex": 6,
"source": "JAVA",
"location": "usr/share/java/libintl-0.21.jar"
},
{
"name": "gcc-10-base",
"version": "10.2.1-6",
"vulns": [
{
"cve": "CVE-2023-4039",
"cvss": 4.8,
"summary": "**DISPUTED**A failure in the -fstack-protector feature in GCC-based toolchains that target AArch64 allows an attacker to exploit an existing buffer overflow in dynamically-sized local variables in your application without this being detected. This stack-protector failure only applies to C99-style dynamically-sized local variables or those created using alloca(). The stack-protector operates as intended for statically-sized local variables. The default behavior when the stack-protector detects an overflow is to terminate your application, resulting in controlled loss of availability. An attacker who can exploit a buffer overflow without triggering the stack-protector might be able to change program flow control to cause an uncontrolled loss of availability or to go further and affect confidentiality or integrity. NOTE: The GCC project argues that this is a missed hardening bug and not a vulnerability by itself.",
"link": "https://security-tracker.debian.org/tracker/CVE-2023-4039",
"scoreVersion": "V3",
"cvssV3": {
"vector": "CVSS:3.1/AV:N/AC:H/PR:N/UI:N/S:U/C:L/I:L/A:N",
"exploitabilityScore": 2.2,
"impactScore": 2.5,
"attackVector": "ATTACK_NETWORK",
"attackComplexity": "COMPLEXITY_HIGH",
"confidentiality": "IMPACT_LOW",
"integrity": "IMPACT_LOW",
"score": 4.8,
"severity": "MEDIUM"
},
"publishedOn": "0001-01-01T00:00:00Z",
"vulnerabilityType": "IMAGE_VULNERABILITY",
"severity": "MODERATE_VULNERABILITY_SEVERITY"
}
],
"layerIndex": 0,
"location": "var/lib/dpkg/status",
"topCvss": 4.8,
"riskScore": 1.09075
},
...
"cvssV3": {
"vector": "CVSS:3.1/AV:L/AC:L/PR:N/UI:R/S:U/C:N/I:N/A:H",
"exploitabilityScore": 1.8,
"impactScore": 3.6,
"userInteraction": "UI_REQUIRED",
"availability": "IMPACT_HIGH",
"score": 5.5,
"severity": "MEDIUM"
},
"publishedOn": "0001-01-01T00:00:00Z",
"vulnerabilityType": "IMAGE_VULNERABILITY",
"severity": "MODERATE_VULNERABILITY_SEVERITY"
}
],
"layerIndex": 6,
"location": "var/lib/dpkg/status",
"topCvss": 5.5,
"riskScore": 1.09075
}
],
"operatingSystem": "debian:11",
"dataSource": {
"id": "a87471e6-9678-4e66-8348-91e302b6de07",
"name": "Scanner V4"
},
"notes": [
"OS_CVES_UNAVAILABLE",
"PARTIAL_SCAN_DATA"
],
"hash": "17396088734232535124"
},
"components": 143,
"cves": 133,
"fixableCves": 1,
"lastUpdated": "2024-04-24T14:38:36.940600379Z",
"riskScore": 6.8999996,
"topCvss": 10,
"notes": [
"MISSING_SIGNATURE",
"MISSING_SIGNATURE_VERIFICATION_DATA"
]
}
Here I tell how I validated my change
TODO(replace-me)
Use this space to explain how you validated that your change functions exactly how you expect it.
Feel free to attach JSON snippets, curl commands, screenshots, etc. Apply a simple benchmark: would the information you
provided convince any reviewer or any external reader that you did enough to validate your change.
It is acceptable to assume trust and keep this section light, e.g. as a bullet-point list.
It is acceptable to skip testing in cases when CI is sufficient, or it's a markdown or code comment change only.
It is also acceptable to skip testing for changes that are too taxing to test before merging. In such case you are
responsible for the change after it gets merged which includes reverting, fixing, etc. Make sure you validate the change
ASAP after it gets merged or explain in PR when the validation will be performed.
Explain here why you skipped testing in case you did so.
Have you created automated tests for your change? Explain here which validation activities you did manually and why so.
Reminder for reviewers
In addition to reviewing code here, reviewers must also review testing and request further testing in case the
performed one does not seem sufficient. As a reviewer, you must not approve the change until you understand the
performed testing and you are satisfied with it.
pkg/images/enricher: 2024/04/22 23:15:51.528782 enricher_impl.go:667: Debug: No signatures associated with image "image-registry.openshift-image-registry.svc:5000/dave/dave-is:latest" so no verification will be done
pkg/grpc/errors: 2024/04/22 23:15:51.528841 interceptor.go:60: Error: Internal error occurred: image enrichment error: error scanning image: image-registry.openshift-image-registry.svc:5000/dave/dave-is:latest error: scanning "image-registry.openshift-image-registry.svc:5000/dave/dave-is:latest" with scanner "Scanner V4": index and scan image report (reference: "image-registry.openshift-image-registry.svc:5000/dave/dave-is@sha256:d158bec5aa876c6699d18fca1bbedf6b07230e7cc07d2dfcb648bacaa8dad317"): get or create index: create index: rpc error: code = Internal desc = listing image layers (reference "image-registry.openshift-image-registry.svc:5000/dave/dave-is@sha256:d158bec5aa876c6699d18fca1bbedf6b07230e7cc07d2dfcb648bacaa8dad317"): Get "https://image-registry.openshift-image-registry.svc:5000/v2/": tls: failed to verify certificate: x509: certificate signed by unknown authority; GET http://image-registry.openshift-image-registry.svc:5000/v2/: unexpected status code 400 Bad Request: Client sent an HTTP request to an HTTPS server.
We probably should not change https to http in this case
For the test you ran, what happened prior to this PR? What was the error message? I'm just curious
For the test you ran, what happened prior to this PR? What was the error message? I'm just curious
It was something like this
ERROR: Scanning image failed: could not scan image: "image-registry.openshift-image-registry.svc:5000/dave/dave-is:latest": rpc error: code = Internal desc = image enrichment error: error scanning image: image-registry.openshift-image-registry.svc:5000/dave/dave-is:latest error: scanning "image-registry.openshift-image-registry.svc:5000/dave/dave-is:latest" with scanner "Scanner V4": index and scan image report (reference: "image-registry.openshift-image-registry.svc:5000/dave/dave-is@sha256:2aa55eb00150f8c6861bd6573e5c2d88603054272490f56833395f6071bb706f"): get or create index: create index: rpc error: code = Internal desc = listing image layers (reference "image-registry.openshift-image-registry.svc:5000/dave/dave-is@sha256:2aa55eb00150f8c6861bd6573e5c2d88603054272490f56833395f6071bb706f"): Get "https://image-registry.openshift-image-registry.svc:5000/v2/": tls: failed to verify certificate: x509: certificate signed by unknown authority; GET http://image-registry.openshift-image-registry.svc:5000/v2/: unexpected status code 400 Bad Request: Client sent an HTTP request to an HTTPS server.
No errors occur when using the “Stackrox Scanner” / “scanner-slim”
| gharchive/pull-request | 2024-04-09T15:17:37 | 2025-04-01T06:45:52.074948 | {
"authors": [
"daynewlee"
],
"repo": "stackrox/stackrox",
"url": "https://github.com/stackrox/stackrox/pull/10680",
"license": "Apache-2.0",
"license_type": "permissive",
"license_source": "github-api"
} |
2566199633 | 🛑 STAKR.space site is down
In 00ed69b, STAKR.space site (https://stakr.space) was down:
HTTP code: 0
Response time: 0 ms
Resolved: STAKR.space site is back up in 7952de2 after 44 minutes.
| gharchive/issue | 2024-10-04T11:57:01 | 2025-04-01T06:45:52.102294 | {
"authors": [
"stakrspace"
],
"repo": "stakrspace/upptime",
"url": "https://github.com/stakrspace/upptime/issues/231",
"license": "MIT",
"license_type": "permissive",
"license_source": "github-api"
} |
2618805471 | 🛑 STAKR.space site is down
In ea7e822, STAKR.space site (https://stakr.space) was down:
HTTP code: 0
Response time: 0 ms
Resolved: STAKR.space site is back up in 7ba6124 after 38 minutes.
| gharchive/issue | 2024-10-28T15:44:20 | 2025-04-01T06:45:52.104963 | {
"authors": [
"stakrspace"
],
"repo": "stakrspace/upptime",
"url": "https://github.com/stakrspace/upptime/issues/563",
"license": "MIT",
"license_type": "permissive",
"license_source": "github-api"
} |
1112437595 | ⚠️ GitLab Server has degraded performance
In 956e36f, GitLab Server (https://gitlab01.its-telekom.eu) experienced degraded performance:
HTTP code: 200
Response time: 6963 ms
Resolved: GitLab Server performance has improved in 92b6a1c.
| gharchive/issue | 2022-01-24T10:17:25 | 2025-04-01T06:45:52.121559 | {
"authors": [
"stamateas"
],
"repo": "stamateas/upptime",
"url": "https://github.com/stamateas/upptime/issues/3094",
"license": "MIT",
"license_type": "permissive",
"license_source": "github-api"
} |
919345415 | ⚠️ GitLab Server has degraded performance
In 8bed0bf, GitLab Server (https://gitlab01.its-telekom.eu) experienced degraded performance:
HTTP code: 200
Response time: 6218 ms
Resolved: GitLab Server performance has improved in 26f1907.
| gharchive/issue | 2021-06-12T01:03:05 | 2025-04-01T06:45:52.124774 | {
"authors": [
"stamateas"
],
"repo": "stamateas/upptime",
"url": "https://github.com/stamateas/upptime/issues/312",
"license": "MIT",
"license_type": "permissive",
"license_source": "github-api"
} |
1556243437 | 🛑 ReportPortal Server is down
In b9db333, ReportPortal Server (https://reporting.its-telekom.eu/) was down:
HTTP code: 0
Response time: 0 ms
Resolved: ReportPortal Server is back up in 0ca87de.
| gharchive/issue | 2023-01-25T08:28:57 | 2025-04-01T06:45:52.127192 | {
"authors": [
"stamateas"
],
"repo": "stamateas/upptime",
"url": "https://github.com/stamateas/upptime/issues/4068",
"license": "MIT",
"license_type": "permissive",
"license_source": "github-api"
} |
1649033193 | ⚠️ GitLab Server has degraded performance
In ce5e423, GitLab Server (https://gitlab01.its-telekom.eu) experienced degraded performance:
HTTP code: 200
Response time: 2719 ms
Resolved: GitLab Server performance has improved in b3a7833.
| gharchive/issue | 2023-03-31T09:36:02 | 2025-04-01T06:45:52.129668 | {
"authors": [
"stamateas"
],
"repo": "stamateas/upptime",
"url": "https://github.com/stamateas/upptime/issues/4145",
"license": "MIT",
"license_type": "permissive",
"license_source": "github-api"
} |
172602879 | Solve systems of nonlinear equations
Summary:
Support solving a system of non-linear equations using one of the better supported unsupported modules in Eigen. An alternative would be to use KINSOL in Sundials, but I think Eigen would be easier.
Description:
To start with, let's assume the user has written a user-defined function that inputs a vector of length n and outputs a vector of length n that contains the values of the equations that are supposed to be zero at the solution. Also, let's assume the user has written a user-defined function that inputs a vector of length n and outputs a square matrix of order n that represents the Jacobian, i.e. the partial derivative of the i-th equation with respect to the j-th input goes in the i,j cell of this matrix.
(This is easier if these two functions are defined in a local functions block such that data and transformed data are in scope, but I need to create another issue for that.)
The call in the transformed parameters (or model) block of the Stan program would look like
transformed parameters {
vector[n] solution;
{
vector[n] starting_values;
// fill starting_values with something intelligent
solution = dogleg(starting_values, equations_functor, jacobian_functor);
}
}
where the signature of the dogleg function (or we could call it powell or something else) is
vector dogleg(vector, functor, functor);
As an overview, the dogleg C++ function would do these steps:
Eigen::VectorXd theta = static_cast<double>(starting_values)
Instantiate a Eigen::HybridNonLinearSolver with a suitable hybrj_functor
Call the hybrj1 method of a Eigen::HybridNonLinearSolver with theta as the initial point
Return a var vector after using the implicit function theorem to figure out the derivatives
In detail, for step 2, see the example at
https://bitbucket.org/eigen/eigen/src/5a47e5a5b02e4d6ae1da98c2348f9c1cb01bdaf9/unsupported/test/NonLinearOptimization.cpp?at=default&fileviewer=file-view-default#NonLinearOptimization.cpp-245
The necessary hybrj_functor has an int operator()(const VectorXd &x, VectorXd &fvec) and an int df(const VectorXd &x, MatrixXd &fjac) that each return 0 on success and use the second argument to store the function values and Jacobian respectively. So, we need to create a hybrj_functor whose operator() calls the functor that is the second argument to dogleg and assigns the resulting vector to fvec while whose df() method calls the functor that is the third argument to dogleg and assigns the resulting matrix to fjac.
For step 4, it is just like
https://en.wikipedia.org/wiki/Implicit_function_theorem#Application:_change_of_coordinates
We need to evaluate the Jacobian at the solution and multiply its inverse by the negative of the solution vector to get the partial derivatives.
Reproducible Steps:
Does not currently exist
Current Output:
Does not currently exist
Expected Output:
A var vector such that if you pass that vector to equations_functor it returns a numerically zero vector and if you pass that vector to jacobian_function it returns a non-singular matrix.
Additional Information:
We need this for the models the Federal Reserve wants us to estimate and lots of economics models generally.
Current Version:
v2.11.0
I'm happy to code up the the language parts, but you don't want me anywhere near any Jacobian or linear algebra code that matters.
We want to coordinate with @charlesm93 on this because there are PK/PD applications we want to be sure to handle. At least I think this is the same feature (I'm really not that good with all this linear algebra and solving stuff).
This sounds like a good starting point for the kind of root finding functionality we would want for pharmacometrics (PMX) applications. Some initial thoughts:
The solver function would need to accept other parameters as arguments and return a var vector with gradients wrt those parameters.
I think we will also want a version that automatically generates the Jacobian.
The primary PMX application is the calculation of amounts in each compartment at a periodic steady-state resulting from multiple equal doses at equal intervals. This involves numerically solving a system of nonlinear equations that themselves involve the numerical solution of a system of ODEs.
The function should allow for the starting values to be parameters but coerce them to double. This permits intelligent automatic calculation of starting values based on model parameters. For example when calculating the steady-state solution for a pharmacokinetic model we might calculate initial estimates based on scaling of a single dose calculation.
I don't have any experience with Eigen::HybridNonLinearSolver, so I don't know how it compares with KINSOL wrt computational efficiency, robustness (e.g., sensitivity to initial estimates), etc.
I think we are saying the same things here. Having the Jacobian be automatic would be nice, but I think it would require fwd mode, which isn't fully in place yet. I haven't used KINSOL either but my sense is that Powell's method, which Eigen implements, is the consensus choice for this sort of thing.
Why would we need forward mode for an embedded Jacobian? In
the ODE solver, we just use nested reverse mode.
Forward mode's ready to go, though it could use a lot
of optimization. When it's written more efficiently, it
should be much faster to calculate N x N Jacobians with
N forward-mode calls rather than N backward-mode ones.
Eigen has a signature that omits the functor for the Jacobian and calculates the Jacobian by numerical differences. That might be less accurate than autodiff, but I think it is about as many flops as N reverse mode calls. In any event, we should start with the case where there is an analytical Jacobian and once that is done, the rest will be easy.
I'd think that would depend on how they do finite diffs to calculate a Jacobian. It could be O(N^2) rather than O(N) if they do each derivative separately.
@bob-carpenter Is this
https://github.com/stan-dev/math/compare/feature/dogleg
about what it needs to be on the Math side in order for you to generate the code for it on the Stan side?
Yes, that would enable a functor dogleg() to be written as
a special expression like integrate_ode() if that's what you're
asking.
It'll be a while before I can get to higher-order functions
within Stan!
Bob
On Aug 28, 2016, at 2:04 PM, bgoodri notifications@github.com wrote:
@bob-carpenter Is this
https://github.com/stan-dev/math/compare/feature/dogleg
about what it needs to be on the Math side in order for you to generate the code for it on the Stan side?
—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub, or mute the thread.
I wanted to ask what the status of the dogleg function was. From what I can tell, the math code has been written but not tested (didn't find unit tests), and we need to expose it to Stan's grammar, in a similar manner than was done for the ODE integrators. I'm happy to get my hands dirty with both tasks.
We can start with a function that requires an analytical jacobian, though in the long run we'll want an automatic approximation of the Jacobian for the function to have broader applications.
I think that is about the state of it. Doing an autodiffed Jacobian is easy
to implement in reverse mode because the .jacobian() method is already
implemented.
On Mon, Nov 7, 2016 at 5:10 PM, Charles Margossian <notifications@github.com
wrote:
I wanted to ask what the status of the dogleg function was. From what I
can tell, the math code has been written but not tested (didn't find unit
tests), and we need to expose it to Stan's grammar, in a similar manner
than was done for the ODE integrators. I'm happy to get my hands dirty with
both tasks.
We can start with a function that requires an analytical jacobian, though
in the long run we'll want an automatic approximation of the Jacobian for
the function to have broader applications.
—
You are receiving this because you authored the thread.
Reply to this email directly, view it on GitHub
https://github.com/stan-dev/stan/issues/2023#issuecomment-258979344, or mute
the thread
https://github.com/notifications/unsubscribe-auth/ADOrqjj0Xa_izX8cSVG71qoJKPl2sdpYks5q76HQgaJpZM4Jqgvr
.
Actually let's take a step back. We need a way to pass in parameters and data. We could use the following signature:
vector equation(vector, real[], real[], int[])
vector dogleg(vector, functor, functor, real[], real[], int[]);
where the additional arguments contain parameters, real data, and integer data. These arguments should also work for the jacobian, which should depend on the same variables as the equation function.
I personally would rather us implement the "local functions" block between
the transformed data and parameters blocks so that the functions would be
defined as part of the class and anything from the data and transformed
data blocks would be in scope. Going the route of how the signatures for
the integrate_ode_* functions are defined is really cumbersome if the
equations involve matrices.
On Mon, Nov 7, 2016 at 5:31 PM, Charles Margossian <notifications@github.com
wrote:
Actually let's take a step back. We need a way to pass in parameters and
data. We could use the following signature:
vector equation(vector, real[], real[], int[])
vector dogleg(vector, functor, functor, real[], real[], int[]);
where the additional arguments contain parameters, real data, and integer
data. These arguments should also work for the jacobian, which should
depend on the same variables as the equation function.
—
You are receiving this because you authored the thread.
Reply to this email directly, view it on GitHub
https://github.com/stan-dev/stan/issues/2023#issuecomment-258984545, or mute
the thread
https://github.com/notifications/unsubscribe-auth/ADOrqjNbqBqT33zL_0fq2y5ldFbsqqq6ks5q76aogaJpZM4Jqgvr
.
+1 regarding additional functors.
What do you mean by "defined as part of the class"? I take it you mean the generated C++? I think we should think design first, then constrain design with implementation details later.
On Nov 7, 2016, at 5:43 PM, bgoodri notifications@github.com wrote:
I personally would rather us implement the "local functions" block between
the transformed data and parameters blocks so that the functions would be
defined as part of the class and anything from the data and transformed
data blocks would be in scope. Going the route of how the signatures for
the integrate_ode_* functions are defined is really cumbersome if the
equations involve matrices.
On Mon, Nov 7, 2016 at 5:31 PM, Charles Margossian <notifications@github.com
wrote:
Actually let's take a step back. We need a way to pass in parameters and
data. We could use the following signature:
vector equation(vector, real[], real[], int[])
vector dogleg(vector, functor, functor, real[], real[], int[]);
where the additional arguments contain parameters, real data, and integer
data. These arguments should also work for the jacobian, which should
depend on the same variables as the equation function.
—
You are receiving this because you authored the thread.
Reply to this email directly, view it on GitHub
https://github.com/stan-dev/stan/issues/2023#issuecomment-258984545, or mute
the thread
https://github.com/notifications/unsubscribe-auth/ADOrqjNbqBqT33zL_0fq2y5ldFbsqqq6ks5q76aogaJpZM4Jqgvr
.
—
You are receiving this because you are subscribed to this thread.
Reply to this email directly, view it on GitHub, or mute the thread.
I thought we had a consensus on the design a few months ago:
Introduce a "local functions" block between transformed data and
parameters
Such functions would be member functions of the model class (like
log_prob and whatnot) rather than floating around in the namespace. Thus,
local functions can access objects declared in data and transformed data
without having to pass them as arguments. The user would define a local
function that inputs a vector of parameters and outputs a vector that is
numerically zero at the solution.
Then functors like dogleg() would only need a minimal number of
arguments, as in
vector dogleg_with_jacobian(function, function, vector)
vector dogleg(function, vector)
Where is the spec? I don't remember seeing an actual design. (I could have missed it and the decision.)
On Nov 7, 2016, at 5:53 PM, bgoodri notifications@github.com wrote:
I thought we had a consensus on the design a few months ago:
Introduce a "local functions" block between transformed data and
parameters
Such functions would be member functions of the model class (like
log_prob and whatnot) rather than floating around in the namespace. Thus,
local functions can access objects declared in data and transformed data
without having to pass them as arguments. The user would define a local
function that inputs a vector of parameters and outputs a vector that is
numerically zero at the solution.
Then functors like dogleg() would only need a minimal number of
arguments, as in
vector dogleg_with_jacobian(function, function, vector)
vector dogleg(function, vector)
—
You are receiving this because you commented.
Reply to this email directly, view it on GitHub, or mute the thread.
https://github.com/stan-dev/stan/wiki/Functionals-spec
On Mon, Nov 7, 2016 at 6:19 PM, Daniel Lee notifications@github.com wrote:
Where is the spec? I don't remember seeing an actual design. (I could have
missed it and the decision.)
On Nov 7, 2016, at 5:53 PM, bgoodri notifications@github.com wrote:
I thought we had a consensus on the design a few months ago:
Introduce a "local functions" block between transformed data and
parameters
Such functions would be member functions of the model class (like
log_prob and whatnot) rather than floating around in the namespace. Thus,
local functions can access objects declared in data and transformed data
without having to pass them as arguments. The user would define a local
function that inputs a vector of parameters and outputs a vector that is
numerically zero at the solution.
Then functors like dogleg() would only need a minimal number of
arguments, as in
vector dogleg_with_jacobian(function, function, vector)
vector dogleg(function, vector)
—
You are receiving this because you commented.
Reply to this email directly, view it on GitHub, or mute the thread.
—
You are receiving this because you authored the thread.
Reply to this email directly, view it on GitHub
https://github.com/stan-dev/stan/issues/2023#issuecomment-258995315, or mute
the thread
https://github.com/notifications/unsubscribe-auth/ADOrqryIH9s39WqwEGmYvMVmTDN8ydiUks5q77H1gaJpZM4Jqgvr
.
On Nov 7, 2016, at 5:53 PM, bgoodri notifications@github.com wrote:
I thought we had a consensus on the design a few months ago:
Yes, we do. I could work on this next. I don't think it'd
be that hard.
I'm realizing that a lot of the type inference we've been
doing in the main body of Stan programs falls down for
functions because of their templating. Can't tell when
things are double or not, for instance, at compile time.
It'd require reasoning about the instantiations of the
functions. That was the bug in the conditional operator,
by the way.
I've been thinking about adding the list/tuple type and
it's going to mess with a lot of the basic code which
assumes a type is (int|real|vector|row_vector|vector|matrix)
with a number of array dimensions. The constraints don't
play into the type system. But adding a tuple is different,
because we need to define the types of the elements.
I can start thinking about general functional programming.
It'd be awesome to add that. And with functors in C++, I
think it might be possible.
I didn't realize local functions were an option we were contemplating. I agree functionals would be awesome and, among other things, would help a lot for the generalized event handler.
implement the "local functions" block between
the transformed data and parameters blocks
You mean between the parameters and transformed parameters block, right? That way parameters can be passed in the function.
Two points:
Are we sticking with Powell's method? Michael suggested we would be better off with a fully gradient-based method since we'll be computing derivatives anyways. Powell's method is hybrid. We could go for Newton's method but it might not be stable enough for a lot of problems.
What is the time frame for implementing local functions? I'd rather do it "right" in the first go, but I'm shooting for a working prototype of the solver in Torsten before the end of January (deliverable for a grant), and I might create a quick and dirty working version.
No way we're going to get local functions by end of January. Too much
going on between now and then and they're going to involve a ton
of testing. But probably not long after that if it's the next big
thing I do after the AST and generator refactor.
Bob
On Dec 5, 2016, at 2:00 PM, Charles Margossian notifications@github.com wrote:
Two points:
• Are we sticking with Powell's method? Michael suggested we would be better off with a fully gradient-based method since we'll be computing derivatives anyways. Powell's method is hybrid. We could go for Newton's method but it might not be stable enough for a lot of problems.
• What is the time frame for implementing local functions? I'd rather do it "right" in the first go, but I'm shooting for a working prototype of the solver in Torsten before the end of January (deliverable for a grant), and I might create a quick and dirty working version.
—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub, or mute the thread.
As discussed in the meeting, I'm going ahead and developing a first version of the solver. I'll keep it modular, and I'll test with the dogleg method Ben began working on. Should I create a new branch or continue working on feature/dogleg?
Same branch
On Thu, Feb 2, 2017 at 1:53 PM, Charles Margossian <notifications@github.com
wrote:
As discussed in the meeting, I'm going ahead and developing a first
version of the solver. I'll keep it modular, and I'll test with the dogleg
method Ben began working on. Should I create a new branch or continue
working on feature/dogleg?
—
You are receiving this because you were assigned.
Reply to this email directly, view it on GitHub
https://github.com/stan-dev/stan/issues/2023#issuecomment-277047328, or mute
the thread
https://github.com/notifications/unsubscribe-auth/ADOrqnc9ajME1LVlV-bN017QHK2kzEc_ks5rYiYVgaJpZM4Jqgvr
.
Whatever's easiest in terms of branching.
From the meeting, we decided to go with an integrate_ode-like
interface for now, then later simplify by removing some arguments.
Ideally, there will be a clean call to the actual solver that
we can plug and play with different solvers, but the main goal's
to get one solver working and building and tested.
Bob
On Feb 2, 2017, at 1:53 PM, Charles Margossian notifications@github.com wrote:
As discussed in the meeting, I'm going ahead and developing a first version of the solver. I'll keep it modular, and I'll test with the dogleg method Ben began working on. Should I create a new branch or continue working on feature/dogleg?
—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub, or mute the thread.
I'm trying to figure out how to call dogleg. I seem to be doing something wrong with the functors (I wrote the functors as I would have written them for the ODE solver):
inline Eigen::VectorXd
algebraEq(const Eigen::VectorXd x) {
Eigen::VectorXd y(2);
y(0) = x(0) - 36;
y(1) = x(1) - 6;
return y;
}
struct algebraEq_functor {
inline Eigen::VectorXd
operator()(const Eigen::VectorXd x) const {
return algebraEq(x);
}
};
inline Eigen::MatrixXd
jacobian(const Eigen::VectorXd x) {
Eigen::MatrixXd y(2, 2);
y(0, 0) = 1;
y(0, 1) = 0;
y(1, 0) = 0;
y(1, 1) = 1;
return y;
}
struct jacobian_functor {
inline Eigen::MatrixXd
operator()(const Eigen::VectorXd x) const {
return jacobian(x);
}
};
TEST(MathMatrix, dogleg) {
Eigen::VectorXd x(2);
x << 32, 5;
Eigen::VectorXd theta;
theta = stan::math::dogleg(x, algebraEq_functor(), jacobian_functor());
}
The compiler produces the following error message:
error: no matching conversion for
functional-style cast from 'const Eigen::VectorXd' (aka 'const
Matrix<double, Dynamic, 1>') to 'algebraEq_functor'
and
error: no matching conversion for
functional-style cast from 'const Eigen::VectorXd' (aka 'const
Matrix<double, Dynamic, 1>') to 'jacobian_functor'
fjac = F2(x);
I'm guessing I'm passing the wrong arguments to dogleg.
It really helps to have line numbers for this, but I'm
going to take a wild guess that when you call this
theta = stan::math::dogleg(x, algebraEq_functor(), jacobian_functor());
}
and it says
The compiler produces the following error message:
error: no matching conversion for
functional-style cast from 'const Eigen::VectorXd' (aka 'const
Matrix<double, Dynamic, 1>') to 'algebraEq_functor'
that you have the arguments in the wrong order. I don't know
what branch you're on. But check out the signature for dogleg
and match it.
Bob
I just uploaded the code on feature/dogleg. The test is dogleg_test.cpp. I'm pretty sure I have the order right (changing the order did not fix the bug). I'll keep digging, but in the meantime, here's the full error message:
In file included from test/unit/math/prim/mat/fun/dogleg_test.cpp:2:
./stan/math/prim/mat/fun/dogleg.hpp:29:18: error: no matching conversion for
functional-style cast from 'const Eigen::VectorXd' (aka 'const
Matrix<double, Dynamic, 1>') to 'algebraEq_functor'
fvec = F1(x);
^~~~
./stan/math/prim/mat/fun/dogleg.hpp:26:14: note: in instantiation of member
function 'stan::math::dogleg(const Eigen::VectorXd &, const
algebraEq_functor, const jacobian_functor)::hybrj_functor::operator()'
requested here
struct hybrj_functor : NLOFunctor {
^
test/unit/math/prim/mat/fun/dogleg_test.cpp:43:23: note: in instantiation of
function template specialization 'stan::math::dogleg<algebraEq_functor,
jacobian_functor>' requested here
theta = stan::math::dogleg(x, algebraEq_functor(), jacobian_functor());
^
test/unit/math/prim/mat/fun/dogleg_test.cpp:14:8: note: candidate constructor
(the implicit copy constructor) not viable: no known conversion from
'const Eigen::VectorXd' (aka 'const Matrix<double, Dynamic, 1>') to
'const algebraEq_functor' for 1st argument
struct algebraEq_functor {
^
test/unit/math/prim/mat/fun/dogleg_test.cpp:14:8: note: candidate constructor
(the implicit default constructor) not viable: requires 0 arguments, but 1
was provided
In file included from test/unit/math/prim/mat/fun/dogleg_test.cpp:2:
./stan/math/prim/mat/fun/dogleg.hpp:33:18: error: no matching conversion for
functional-style cast from 'const Eigen::VectorXd' (aka 'const
Matrix<double, Dynamic, 1>') to 'jacobian_functor'
fjac = F2(x);
^~~~
./stan/math/prim/mat/fun/dogleg.hpp:26:14: note: in instantiation of member
function 'stan::math::dogleg(const Eigen::VectorXd &, const
algebraEq_functor, const jacobian_functor)::hybrj_functor::df' requested
here
struct hybrj_functor : NLOFunctor {
^
test/unit/math/prim/mat/fun/dogleg_test.cpp:43:23: note: in instantiation of
function template specialization 'stan::math::dogleg<algebraEq_functor,
jacobian_functor>' requested here
theta = stan::math::dogleg(x, algebraEq_functor(), jacobian_functor());
^
test/unit/math/prim/mat/fun/dogleg_test.cpp:31:8: note: candidate constructor
(the implicit copy constructor) not viable: no known conversion from
'const Eigen::VectorXd' (aka 'const Matrix<double, Dynamic, 1>') to
'const jacobian_functor' for 1st argument
struct jacobian_functor {
^
test/unit/math/prim/mat/fun/dogleg_test.cpp:31:8: note: candidate constructor
(the implicit default constructor) not viable: requires 0 arguments, but 1
was provided
2 errors generated.
make: *** [test/unit/math/prim/mat/fun/dogleg_test.o] Error 1
make test/unit/math/prim/mat/fun/dogleg_test failed
exit now (02/03/17 08:46:46 EST)
Are you sure it's on the branch? I'm not finding the test file in
stan-dev/math
On Fri, Feb 3, 2017 at 8:50 AM Charles Margossian notifications@github.com
wrote:
I just uploaded the code on feature/dogleg. The test is dogleg_test.cpp.
I'm pretty sure I have the order right (changing the order did not fix the
bug). I'll keep digging, but in the meantime, here's the full error message:
In file included from test/unit/math/prim/mat/fun/dogleg_test.cpp:2:
./stan/math/prim/mat/fun/dogleg.hpp:29:18: error: no matching conversion for
functional-style cast from 'const Eigen::VectorXd' (aka 'const
Matrix<double, Dynamic, 1>') to 'algebraEq_functor'
fvec = F1(x);
^~~~
./stan/math/prim/mat/fun/dogleg.hpp:26:14: note: in instantiation of member
function 'stan::math::dogleg(const Eigen::VectorXd &, const
algebraEq_functor, const jacobian_functor)::hybrj_functor::operator()'
requested here
struct hybrj_functor : NLOFunctor {
^
test/unit/math/prim/mat/fun/dogleg_test.cpp:43:23: note: in instantiation of
function template specialization 'stan::math::dogleg<algebraEq_functor,
jacobian_functor>' requested here
theta = stan::math::dogleg(x, algebraEq_functor(), jacobian_functor());
^
test/unit/math/prim/mat/fun/dogleg_test.cpp:14:8: note: candidate constructor
(the implicit copy constructor) not viable: no known conversion from
'const Eigen::VectorXd' (aka 'const Matrix<double, Dynamic, 1>') to
'const algebraEq_functor' for 1st argument
struct algebraEq_functor {
^
test/unit/math/prim/mat/fun/dogleg_test.cpp:14:8: note: candidate constructor
(the implicit default constructor) not viable: requires 0 arguments, but 1
was provided
In file included from test/unit/math/prim/mat/fun/dogleg_test.cpp:2:
./stan/math/prim/mat/fun/dogleg.hpp:33:18: error: no matching conversion for
functional-style cast from 'const Eigen::VectorXd' (aka 'const
Matrix<double, Dynamic, 1>') to 'jacobian_functor'
fjac = F2(x);
^~~~
./stan/math/prim/mat/fun/dogleg.hpp:26:14: note: in instantiation of member
function 'stan::math::dogleg(const Eigen::VectorXd &, const
algebraEq_functor, const jacobian_functor)::hybrj_functor::df' requested
here
struct hybrj_functor : NLOFunctor {
^
test/unit/math/prim/mat/fun/dogleg_test.cpp:43:23: note: in instantiation of
function template specialization 'stan::math::dogleg<algebraEq_functor,
jacobian_functor>' requested here
theta = stan::math::dogleg(x, algebraEq_functor(), jacobian_functor());
^
test/unit/math/prim/mat/fun/dogleg_test.cpp:31:8: note: candidate constructor
(the implicit copy constructor) not viable: no known conversion from
'const Eigen::VectorXd' (aka 'const Matrix<double, Dynamic, 1>') to
'const jacobian_functor' for 1st argument
struct jacobian_functor {
^
test/unit/math/prim/mat/fun/dogleg_test.cpp:31:8: note: candidate constructor
(the implicit default constructor) not viable: requires 0 arguments, but 1
was provided
2 errors generated.
make: *** [test/unit/math/prim/mat/fun/dogleg_test.o] Error 1
make test/unit/math/prim/mat/fun/dogleg_test failed
exit now (02/03/17 08:46:46 EST)
—
You are receiving this because you are subscribed to this thread.
Reply to this email directly, view it on GitHub
https://github.com/stan-dev/stan/issues/2023#issuecomment-277250866, or mute
the thread
https://github.com/notifications/unsubscribe-auth/AAfA6YUCsWxm59925Oj_MiI6xScwxAigks5rYzC8gaJpZM4Jqgvr
.
Yes, positive.
https://github.com/stan-dev/math/blob/feature/dogleg/test/unit/math/prim/mat/fun/dogleg_test.cpp
Line 29 in
https://github.com/stan-dev/math/blob/feature/dogleg/stan/math/prim/mat/fun/dogleg.hpp
F1 is being treated as the type name not as the argument name (that would
be the functor) so it tries to do a cast rather than producing an Eigen
vector like you want. I'm not sure why the arguments are not named since
that would solve the problem but maybe there's more magic going on than I
realize here. In any case, what's happening on line 29 is understandable
and not what you want. K
On Fri, Feb 3, 2017 at 9:05 AM Charles Margossian notifications@github.com
wrote:
Yes, positive.
https://github.com/stan-dev/math/blob/feature/dogleg/test/unit/math/prim/mat/fun/dogleg_test.cpp
—
You are receiving this because you commented.
Reply to this email directly, view it on GitHub
https://github.com/stan-dev/stan/issues/2023#issuecomment-277253926, or mute
the thread
https://github.com/notifications/unsubscribe-auth/AAfA6WMYiEHvmPG_PhupWXRB7rx79AOFks5rYzQtgaJpZM4Jqgvr
.
It was intended to work as in Eigen's tests
https://bitbucket.org/eigen/eigen/src/5430cad5531274d51024d3950d2fa511aeec5055/unsupported/test/NonLinearOptimization.cpp?at=default&fileviewer=file-view-default
but apparently something went wrong.
I changed it locally so that the last two arguments were named, but now I
get an error that says
./stan/math/prim/mat/fun/dogleg.hpp:29:18: error: reference to local
variable 'fvec_functor' declared in enclosing function
'stan::math::dogleg<algebraEq_functor, jacobian_functor>'
fvec = fvec_functor(x);
^
./stan/math/prim/mat/fun/dogleg.hpp:26:14: note: in instantiation of member
function 'stan::math::dogleg(const Eigen::VectorXd &, const
algebraEq_functor, const
jacobian_functor)::hybrj_functor::operator()' requested here
struct hybrj_functor : NLOFunctor {
^
test/unit/math/prim/mat/fun/dogleg_test.cpp:43:23: note: in instantiation
of function template specialization 'stan::math::dogleg<algebraEq_functor,
jacobian_functor>' requested here
theta = stan::math::dogleg(x, algebraEq_functor(), jacobian_functor());
^
./stan/math/prim/mat/fun/dogleg.hpp:24:47: note: 'fvec_functor' declared
here
dogleg(const Eigen::VectorXd& x, const F1 fvec_functor, const F2
fjac_functor) {
^
So, how do we pass the user's functions to the solver?
On Fri, Feb 3, 2017 at 9:24 AM, Krzysztof Sakrejda <notifications@github.com
wrote:
Line 29 in
https://github.com/stan-dev/math/blob/feature/dogleg/stan/
math/prim/mat/fun/dogleg.hpp
F1 is being treated as the type name not as the argument name (that would
be the functor) so it tries to do a cast rather than producing an Eigen
vector like you want. I'm not sure why the arguments are not named since
that would solve the problem but maybe there's more magic going on than I
realize here. In any case, what's happening on line 29 is understandable
and not what you want. K
On Fri, Feb 3, 2017 at 9:05 AM Charles Margossian <
notifications@github.com>
wrote:
Yes, positive.
https://github.com/stan-dev/math/blob/feature/dogleg/test/
unit/math/prim/mat/fun/dogleg_test.cpp
—
You are receiving this because you commented.
Reply to this email directly, view it on GitHub
https://github.com/stan-dev/stan/issues/2023#issuecomment-277253926,
or mute
the thread
<https://github.com/notifications/unsubscribe-auth/AAfA6WMYiEHvmPG_
PhupWXRB7rx79AOFks5rYzQtgaJpZM4Jqgvr>
.
—
You are receiving this because you were assigned.
Reply to this email directly, view it on GitHub
https://github.com/stan-dev/stan/issues/2023#issuecomment-277258329, or mute
the thread
https://github.com/notifications/unsubscribe-auth/ADOrqk6YRY5VuxLeGUQHRb7SKIAt4t3-ks5rYzingaJpZM4Jqgvr
.
Ben, are we pre-c++11 here or post-c++11 here? Looks like you can't use a
struct declared inside a function as as template parameter pre c++11 (
http://stackoverflow.com/a/36766405) but you can with c++11. I haven't
checked the details on the c++11 usage.
On Fri, Feb 3, 2017 at 10:46 AM bgoodri notifications@github.com wrote:
I changed it locally so that the last two arguments were named, but now I
get an error that says
./stan/math/prim/mat/fun/dogleg.hpp:29:18: error: reference to local
variable 'fvec_functor' declared in enclosing function
'stan::math::dogleg<algebraEq_functor, jacobian_functor>'
fvec = fvec_functor(x);
^
./stan/math/prim/mat/fun/dogleg.hpp:26:14: note: in instantiation of member
function 'stan::math::dogleg(const Eigen::VectorXd &, const
algebraEq_functor, const
jacobian_functor)::hybrj_functor::operator()' requested here
struct hybrj_functor : NLOFunctor {
^
test/unit/math/prim/mat/fun/dogleg_test.cpp:43:23: note: in instantiation
of function template specialization 'stan::math::dogleg<algebraEq_functor,
jacobian_functor>' requested here
theta = stan::math::dogleg(x, algebraEq_functor(), jacobian_functor());
^
./stan/math/prim/mat/fun/dogleg.hpp:24:47: note: 'fvec_functor' declared
here
dogleg(const Eigen::VectorXd& x, const F1 fvec_functor, const F2
fjac_functor) {
^
So, how do we pass the user's functions to the solver?
On Fri, Feb 3, 2017 at 9:24 AM, Krzysztof Sakrejda <
notifications@github.com
wrote:
Line 29 in
https://github.com/stan-dev/math/blob/feature/dogleg/stan/
math/prim/mat/fun/dogleg.hpp
F1 is being treated as the type name not as the argument name (that would
be the functor) so it tries to do a cast rather than producing an Eigen
vector like you want. I'm not sure why the arguments are not named since
that would solve the problem but maybe there's more magic going on than I
realize here. In any case, what's happening on line 29 is understandable
and not what you want. K
On Fri, Feb 3, 2017 at 9:05 AM Charles Margossian <
notifications@github.com>
wrote:
Yes, positive.
https://github.com/stan-dev/math/blob/feature/dogleg/test/
unit/math/prim/mat/fun/dogleg_test.cpp
—
You are receiving this because you commented.
Reply to this email directly, view it on GitHub
https://github.com/stan-dev/stan/issues/2023#issuecomment-277253926,
or mute
the thread
<https://github.com/notifications/unsubscribe-auth/AAfA6WMYiEHvmPG_
PhupWXRB7rx79AOFks5rYzQtgaJpZM4Jqgvr>
.
—
You are receiving this because you were assigned.
Reply to this email directly, view it on GitHub
https://github.com/stan-dev/stan/issues/2023#issuecomment-277258329,
or mute
the thread
<
https://github.com/notifications/unsubscribe-auth/ADOrqk6YRY5VuxLeGUQHRb7SKIAt4t3-ks5rYzingaJpZM4Jqgvr
.
—
You are receiving this because you commented.
Reply to this email directly, view it on GitHub
https://github.com/stan-dev/stan/issues/2023#issuecomment-277282163, or mute
the thread
https://github.com/notifications/unsubscribe-auth/AAfA6TTnfetjG5UUfqRbKcx_XdItnd58ks5rY0vcgaJpZM4Jqgvr
.
I guess I'm confused about why we're even bothering with struct hybrj_functor, couldn't we just ask the user to supply one functor like that or make it not be local to the function? Either of those would make this simpler.
That is a possibility, although I think we will be post-C++11 by the time
this is live.
On Fri, Feb 3, 2017 at 11:40 AM, Krzysztof Sakrejda <
notifications@github.com> wrote:
I guess I'm confused about why we're even bothering with struct
hybrj_functor, couldn't we just ask the user to supply one functor like
that or make it not be local to the function? Either of those would make
this simpler.
—
You are receiving this because you were assigned.
Reply to this email directly, view it on GitHub
https://github.com/stan-dev/stan/issues/2023#issuecomment-277296940, or mute
the thread
https://github.com/notifications/unsubscribe-auth/ADOrqvV2NYL7ZUn5wb_gBw6rfrxMVZOdks5rY1iEgaJpZM4Jqgvr
.
are we pre-c++11 here or post-c++11 here?
If we want to expose this to the stan language, we have to work with pre-C++11.
I guess I'm confused about why we're even bothering with struct hybrj_functor, couldn't we just ask the user to supply one functor like that or make it not be local to the function?
From a user's perpective, it probably makes more sense to supply fvec (and fjac, the Jacobian). I suppose hybrj_functor doesn't have to be local; instead, we could create a constructor that requires fvec and fjac as two of its arguments. I'll try an implementation. Currently working through NonLinearOptimization to see if I can get the original function working.
Ok, if you do figure out the original I'd love to hear about it, I'm not
sure I understand the intent there.
On Fri, Feb 3, 2017, 2:05 PM Charles Margossian notifications@github.com
wrote:
are we pre-c++11 here or post-c++11 here?
If we want to expose this to the stan language, we have to work with
pre-C++11.
I guess I'm confused about why we're even bothering with struct
hybrj_functor, couldn't we just ask the user to supply one functor like
that or make it not be local to the function?
From a user's perpective, it probably makes more sense to supply fvec (and
fjac, the Jacobian). I suppose hybrj_functor doesn't have to be local;
instead, we could create a constructor that requires fvec and fjac as two
of its arguments. I'll try an implementation. Currently working through
NonLinearOptimization to see if I can get the original function working.
—
You are receiving this because you commented.
Reply to this email directly, view it on GitHub
https://github.com/stan-dev/stan/issues/2023#issuecomment-277333473, or mute
the thread
https://github.com/notifications/unsubscribe-auth/AAfA6XsF2oFFyIKn1KnOjyKkdXdVgXR3ks5rY3pagaJpZM4Jqgvr
.
Hi Krzysztof,
I updated the "unit test" I started working on. In it I solve a system of equations using Eigen's function and following closely what's on the link Ben sent us. There is a bit of work in progress in the dogleg function, so apologies if this appears confusing.
I got a working prototype of the dogleg function! Adding more difficult unit tests.
Nice! It looks like you had to do more than just name the arguments?
On Fri, Feb 3, 2017, 4:44 PM Charles Margossian notifications@github.com
wrote:
I got a working prototype of the dogleg function! Adding more difficult
unit tests.
—
You are receiving this because you commented.
Reply to this email directly, view it on GitHub
https://github.com/stan-dev/stan/issues/2023#issuecomment-277370374, or mute
the thread
https://github.com/notifications/unsubscribe-auth/AAfA6QU7golv0rOkIzHRGHDkODEVubJEks5rY5-jgaJpZM4Jqgvr
.
Yes. I moved the hybrj_functor structure outside the scope of the function, and I gave it two templated members (f1 and f2, which correspond to the functors respectively returning the y(x) and the Jacobian). I then created a constructor to allow the dogleg function to pass in f1 and f2 as defined by the user.
I also changed the NLOFunctor, mostly following the Eigen's example. I think there are ways of simplifying it.
@charlesm93 --- is there a branch somewhere? I didn't see one with the issue number on it, which is our usual convention.
Currently, the name of the branch is feature/dogleg. I can switch to a new branch named feature/issue-2023-dogleg.
Assertion failed: (n==qtb.size()), function dogleg, file lib/eigen_3.2.9/unsupported/Eigen/src/NonLinearOptimization/dogleg.h, line 30.
What is n and what is qtb and why are their sizes required to be the same? Sorry for asking the obvious, but that's what the assertion says the problem is.
This is actually not obvious at all and requires a bit of digging in eigen's code. In dogleg.h we have:
n = qrfac.cols()
where qrfac is the triangular matrix obtained when doing the QR decomposition of the Jacobian, i.e. J=QR, where Q is a unary matrix and R a triangular matrix.
qtb is given by transpose(Q) * y, where y(x0, x1) is the function we want the roots of. Ok, so let's check the dimensions. In the unit test, y is a vector of length 1, that takes in an input vector of length 2.
Thus dim(J) = 1 x 2, dim(Q) = 1 x 1, and dim(R) = 1 x 2.
Thus n = 2 (the number of columns in R), and size(qtb) = (1 x 1) * (1 * 1) = 1 (where I treated y as row-major vector).
... which makes the assert statement fail.
Let's look at a more abstract case, where we have a rectangular Jacobian.
Let size(x) = k, and size(y) = l.
Thus dim(J) = l x k, dim(Q) = l x l, and dim(R) = l x k.
Thus n = k and size(qtb) = (l x l) * (l * 1) = l.
Once again, the assert statement would fail.
Inside the code for dogleg, qtb[i] appears in for loops that range from 0 to n, so to avoid a segmentation fault, the assert statement needs to be verified. But I haven't yet connected this constraint in the code to a mathematical constraint.
I started working on a variation of dogleg for the function we will eventually expose to the Stan language:
algebra_solver(algebraEq_functor, jacobian_functor, x, parms, dat, dat_int)
@bgoodri Do we want to use this function instead of dogleg? For now, I'm keeping both. One version may work better for C++ users and spare them the awkward step of packing and unpacking values in the parms, dat, and dat_int containers. I didn't exactly construct algebra_solver as a wrapper around dogleg, although the architecture is pretty similar (I had to do a bit more gymnastic with functors).
Also, with Ben's ok, I'll rename the branch, as requested by @bob-carpenter, feature/issue-2023-algebra-solver.
I'm starting a post on discourse to discuss how we'll compute the Jacobian of the solutions with respect to parms.
The next step is to compute the Jacobian of the solutions with respect to parms. @bgoodri suggests using the implicit function theorem, which, admittedly, I am still wrapping my head around.
Here's a scheme I propose implementing:
We know (or can compute) J_f(x), the Jacobian of the function w.r.t the unknowns x.
We could also, by properly manipulating functors, compute J_f(p), the Jacobian of the function w.r.t the parameters p.
Applying the chain rule, and using the inverse function theorem, I get
J_x(p) = J_f(p) * J_x(f) = J_f(p) * inverse[ J_f(x) ]
Note we assume the functions are differentiable around the points of interest. I think this should work, although, I'm wondering if I overlooked a subtlety when applying the chain rule.
@bgoodri Did you have something similar in mind?
I already started the case where the Jacobian is unknown and has to be done
by autodiff. But it does not yet have the implicit function stuff once the
solution is found.
On Wed, Feb 15, 2017 at 10:09 AM, Charles Margossian <
notifications@github.com> wrote:
The next step is to compute the Jacobian of the solutions with respect to
parms. @bgoodri https://github.com/bgoodri suggests using the implicit
function theorem, which, admittedly, I am still wrapping my head around.
Here's a scheme I propose implementing:
We know (or can compute) J_f(x), the Jacobian of the function w.r.t the
unknowns x.
We could also, by properly manipulating functors, compute J_f(p), the
Jacobian of the function w.r.t the parameters p.
Applying the chain rule, and using the inverse function theorem, I get
J_x(p) = J_f(p) * J_x(f) = J_f(p) * inverse[ J_f(x) ]
Note we assume the functions are differentiable around the points of
interest. I think this should work, although, I'm wondering if I overlooked
a subtlety when applying the chain rule.
@bgoodri https://github.com/bgoodri Did you have something similar in
mind?
—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub
https://github.com/stan-dev/stan/issues/2023#issuecomment-280036290, or mute
the thread
https://github.com/notifications/unsubscribe-auth/ADOrqmATUFM7xJbaCecPaw5n8BLso9mlks5rcxU4gaJpZM4Jqgvr
.
I uploaded a working prototype of algebra_solver with a simple unit test (all in the rev regime). The function finds the solutions and propagates the gradient!
A couple of things:
I got an error when I included both the dogleg under the rev and the prim directories, due to redefinition of the functors (as a temporary solution, I did not include rev/.../dogleg.hpp in rev/mat.hpp; the easy fix would be to rename the functors in one of the files).
I wanted to use value_of to pass parms as an eigen vector of doubles instead of var, but got an error regarding converting vars to double. I created a new value function to do that. We could also overload value_of to convert var to doubles, but I'm not sure if this would be a desirable feature.
Ok, the next steps involve: checks and error messages (there are a few constraints posed by the calculation of the Jacobian), fwd regime, more unit tests.
Awesome. Can't wait to try it out on some icdf's! K
On Wed, Feb 22, 2017, 2:01 PM Charles Margossian notifications@github.com
wrote:
I uploaded a working prototype of algebra_solver with a simple unit test
(all in the rev regime). The function finds the solutions and propagates
the gradient!
A couple of things:
I got an error when I included both the dogleg under the rev and the
prim directories, due to redefinition of the functors (as a temporary
solution, I did not include rev/.../dogleg.hpp in rev/mat.hpp; the easy fix
would be to rename the functors in one of the files).
I wanted to use value_of to pass parms as an eigen vector of doubles
instead of var, but got an error regarding converting vars to double. I
created a new value function to do that. We could also overload
value_of to convert var to doubles, but I'm not sure if this would be
a desirable feature.
Ok, the next steps involve: checks and error messages (there are a few
constraints posed by the calculation of the Jacobian), fwd regime, more
unit tests.
—
You are receiving this because you commented.
Reply to this email directly, view it on GitHub
https://github.com/stan-dev/stan/issues/2023#issuecomment-281767260, or mute
the thread
https://github.com/notifications/unsubscribe-auth/AAfA6atwbKb4PsRN64hccv87Yfz3K7F-ks5rfIXxgaJpZM4Jqgvr
.
On Feb 22, 2017, at 2:56 PM, Krzysztof Sakrejda notifications@github.com wrote:
Awesome. Can't wait to try it out on some icdf's! K
On Wed, Feb 22, 2017, 2:01 PM Charles Margossian notifications@github.com
wrote:
I uploaded a working prototype of algebra_solver with a simple unit test
(all in the rev regime). The function finds the solutions and propagates
the gradient!
A couple of things:
I got an error when I included both the dogleg under the rev and the
prim directories, due to redefinition of the functors (as a temporary
solution, I did not include rev/.../dogleg.hpp in rev/mat.hpp; the easy fix
would be to rename the functors in one of the files).
A header guard doesn't work?
I wanted to use value_of to pass parms as an eigen vector of doubles
instead of var, but got an error regarding converting vars to double. I
created a new value function to do that. We could also overload
value_of to convert var to doubles, but I'm not sure if this would be
a desirable feature.
You should be able to compute var to double using value_of.
You just need to make sure to include the header for it from rev.
Ok, the next steps involve: checks and error messages (there are a few
constraints posed by the calculation of the Jacobian), fwd regime, more
unit tests.
Sounds good.
Bob
Hi,
Any ideas for particularly hard algebra systems I should throw at the solver in the unit tests?
Linear systems with analytic solutions so you can verify accuracy,
for example a linear matrix system.
Non-linear systems with analytic solutions so you can verify accurac,
for example third-order polynomials with known roots.
Systems with no solutions to check error conditions.
Systems with degenerate solutions to check error conditions.
On Mar 27, 2017, at 11:50 AM, Charles Margossian notifications@github.com wrote:
Hi,
Any ideas for particularly hard algebra systems I should throw at the solver in the unit tests?
—
You are receiving this because you are subscribed to this thread.
Reply to this email directly, view it on GitHub https://github.com/stan-dev/stan/issues/2023#issuecomment-289496284, or mute the thread https://github.com/notifications/unsubscribe-auth/ABdNlnmCQzaagovrIxHpwP2AZ9S4Xlv-ks5rp9rJgaJpZM4Jqgvr.
I threw two unsolvable systems at the solver, and didn't get any error message. I expected Eigen's dogleg function would react when it can't solve a system. Instead, it either gives me the wrong result or nan. I could implement a check inside algebra_solver that rejects the current proposal when the solution is "bad" (but according to what metric? Do we require the system to be 0 +/- some error? Seems like the algorithm should be able to put a bound on the error of its approximation.)
The Eigen function HybridNonLinearSolver is based on one of the nonlinear equation solvers in MINPACK, HYBRJ or HYBRJ1 I am guessing. Based on an incomplete read of the MINPACK documentation those functions attempt to find the minimum of the L2 norm of the user-specified functions and leave it to the user to determine if that minimum is a root. In your example it looks like the function correctly found that minimum. We need to add code to check if the result is a root.
The Eigen function HybridNonLinearSolver is based on one of the nonlinear equation solvers in MINPACK, HYBRJ or HYBRJ1 I am guessing
Yes, that seems right, see https://eigen.tuxfamily.org/dox/unsupported/group__NonLinearOptimization__Module.html.
We need to add code to check if the result is a root.
Ok, that sends us down the "check and accept/reject metropolis proposal" route. I'll require the system to be 0 +/- e-10 and leave it open to discussion as to whether we want to change the error, give the user control over it, etc.
Unit tests work. ;-)
If Eigen doesn’t throw on a failure to converge to a root then we’ll have
to do that by hand. Just throw an exception and the sampler will take
care of the accept/reject part.
On Mar 29, 2017, at 4:10 PM, Charles Margossian notifications@github.com wrote:
The Eigen function HybridNonLinearSolver is based on one of the nonlinear equation solvers in MINPACK, HYBRJ or HYBRJ1 I am guessing
Yes, that seems right, see https://eigen.tuxfamily.org/dox/unsupported/group__NonLinearOptimization__Module.html https://eigen.tuxfamily.org/dox/unsupported/group__NonLinearOptimization__Module.html.
We need to add code to check if the result is a root.
Ok, that sends us down the "check and accept/reject metropolis proposal" route. I'll require the system to be 0 +/- e-10 and leave it open to discussion as to whether we want to change the error, give the user control over it, etc.
—
You are receiving this because you commented.
Reply to this email directly, view it on GitHub https://github.com/stan-dev/stan/issues/2023#issuecomment-290210688, or mute the thread https://github.com/notifications/unsubscribe-auth/ABdNlg8fKCEGdAsJYPGb5ZGO0vtiRtmYks5rqrq_gaJpZM4Jqgvr.
We'll probably want a version with a sensible default and one with user control. That's what we did with the ODE solvers. If not, then just the one with user controls (along with at least recommended defaults).
Usually there's both an absolute tolerance and relative tolerance and some logic requiring both tolerances to be met.
Per Thursday's conversation, I added a test demonstrating how initial guesses can influence the solution in a degenerate case. The Jacobian "adapts" to the solution -- which is pretty neat!
Usually there's both an absolute tolerance and relative tolerance and some logic requiring both tolerances to be met.
@bob-carpenter: currently, the code throws an exception when f(solution) doesn't go to 0 +/- abs_tol. I don't think a relative error is formally defined in this case; the only thing that comes to my mind would be an error relative to the initial guess (x(i) * rel_tol) -- but does testing for 0 +/- (x(i) * rel_tol) really make sense?
We'll then have two signatures:
algebra_solver(x, y, dat, dat_int, abs_tol, rel_tol)
algebra_solver(x, y, dat, dat_int)
Next, implementation in Stan. My only concern is translating the system of equations into a functor. To get the Jacobian with respect to either x or y, I constructed the operator of the functor as follow:
operator()(const Eigen::Matrix<T, Eigen::Dynamic, 1> x) const {
if (x_is_dv_)
return degenerateEq(x, y_, dat_, dat_int_);
else
return degenerateEq(x_, x, dat_, dat_int_);
}
The idea is that either x or y can be the independent variables, with respect to which we compute the jacobian. Which is which is determined by the member x_is_dv (a boolean).
I'm looking at the way a regular function gets translated from Stan to C++. I assume this is what I'll have to deal with -- or can we tweak how the function passed to the algebraic solver gets translated?
...
@bob-carpenter: currently, the code throws an exception when f(solution) doesn't go to 0 +/- abs_tol. I don't think a relative error is formally defined in this case; the only thing that comes to my mind would be an error relative to the initial guess (x(i) * rel_tol) -- but does testing for 0 +/- (x(i) * rel_tol) really make sense?
Relative tolerance is defined relative to where you're at. So if the solution is 1e8, we don't want to just use an absolute tolerance of 1e-10 as the difference is beyond double-precision floating point capacity (which is about 1e-16).
Where this gets tricky is for solutions near zero. If the actual solution's zero, no relative tolerance is going to work because you keep getting closer to zero then reducing the relative tolerance.
I'd suggest looking at how the optimizers built into Stan use tolerances.
If you don't have relative tolerance, then you absolutely are going to need a user-specified absolute tolerance.
We'll then have two signatures:
• algebra_solver(x, y, dat, dat_int, abs_tol, rel_tol)
• algebra_solver(x, y, dat, dat_int)
Next, implementation in Stan. My only concern is translating the system of equations into a functor. To get the Jacobian with respect to either x or y, I constructed the operator of the functor as follow:
operator()(const Eigen::Matrix<T, Eigen::Dynamic, 1> x) const {
if (x_is_dv_)
return degenerateEq(x, y_, dat_, dat_int_);
else
return degenerateEq(x_, x, dat_, dat_int_);
}
The idea is that either x or y can be the independent variables, with respect to which we compute the jacobian.
Can they both be independent?
Which is which is determined by the member x_is_dv (a boolean).
You want to calculate x_is_dv statically, not as a member variable. Check out how the code for the distributions (e.g., normal_lpdf) is written.
The ode_integrator functions aren't actually functions---they're special expression constructs. This allows them to take in names of functions (our ordinary functions don't take functions as arguments) and also constrain some of their arguments (like solution times and data input) to be double only.
I'm looking at the way a regular function gets translated from Stan to C++. I assume this is what I'll have to deal with -- or can we tweak how the function passed to the algebraic solver gets translated?
Write one with a couple arguments and look at the output in the model file .hpp generated by stanc (you can use CmdStan's makefile, just cd to cmdstan, then do
$ make target
where target is the path foo/bar if the Stan file is in foo/bar.stan.
Relative tolerance is defined relative to where you're at. So if the solution is 1e8, we don't want to just use an absolute tolerance of 1e-10 as the difference is beyond double-precision floating point capacity (which is about 1e-16).
We currently test the solution by plugging it in the system. We want to check z(x) = 0. The tricky thing is we're not directly directly measuring an error in x, but in z. So we need to propagate the error. The "final relative tolerance" would be z(x * rel_tol). This gives us how far away from 0 z can be.
Can they [the arguments x and y] both be independent?
No. When you call Jacobian, f is expected to have exactly one dependent argument.
Write one with a couple arguments and look at the output in the model file .hpp generated by stanc
Helpful. A bit of functor gymnastic did the trick, now I need to update all the tests.
That's tricky not because it's a function z(x), but because the function's value is zero, z(x) = 0. Then relative error in z(x) doesn't really make sense.
How do the systems you're using specify tolerances?
That's tricky not because it's a function z(x), but because the function's value is zero, z(x) = 0. Then relative error in z(x) doesn't really make sense.
Ah yes, I get your point. To quote the MINPACK manual: "F-convergence is not tested for systems of nonlinear equations where F(x*) = 0 is the expected result." They propose another convergence test, I'm currently reading through it.
After reading through MINPACK (see attached) and dissecting Eigen's code, I found the following tuning parameters:
xtol: allows to test for convergence; the root finder stops when (delta < xtol * xnorm)
maxfev: the maximum number of function evaluations; the root finder stops when it reaches maxfev.
Eigen does not throw a message when maxfev is reached which may be an issue.
In addition, I propose adding an absolute tolerance test on F(x*).
ftol: if ||F(x*)|| > ftol, throw an exception.
This gives three tuning parameters to the user: xtol, maxfev, and ftol, which are analogous, though not equivalent, to the relative_tolerance, the max_num_steps, and the absolute_tolerance in the ODE integrator.
The user should understand the root finder stops when it reaches acceptable convergence to the solution (as determined by xtol and ||F(x)||) or when it reaches the maximum number of iterations. If the result is not acceptable, an exception is thrown when looking at the final value of F(x*). The error message suggests decreasing xtol or increasing maxfev.
We could try to get an exception when maxfev is reached but this might require creating a stan version of HybridNonLinear.h (since we do not edit Eigen's code).
ANL8074a.pdf
RE: what does the std::ostream* pstream_ argument do in integrate_ode?
That stream is for print or reject statements inside the user-defined
function being integrated.
On Fri, Apr 7, 2017 at 9:42 AM, Charles Margossian <notifications@github.com
wrote:
RE: what does the std::ostream* pstream_ argument do in integrate_ode?
—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub
https://github.com/stan-dev/stan/issues/2023#issuecomment-292539569, or mute
the thread
https://github.com/notifications/unsubscribe-auth/ADOrqiF-Xi5mtlChmrpbFnggRnVVw8XJks5rtj1LgaJpZM4Jqgvr
.
Switch from using the feature/dogleg branch to the feature/issue-2023-algebra-solver branch to respect dev norms.
I'm ready to submit a pull request, but I've only "finished" (in quotes because I expect we'll do revisions as we review the code) the algebra_solver for the rev case. Is there interest in the fwd case? Even if there is, I'd rather first add the rev version and then work on fwd case.
For these higher-order things we can go with just reverse. We don't have a forward-mode ODE solver yet!
Are the arguments all sufficiently general that we'll be able to plug into Stan with double values wherever rev-mode vars are allowed? Otherwise, we tend to need to overpromote the double values to var.
On Apr 7, 2017, at 3:32 PM, Charles Margossian notifications@github.com wrote:
I'm ready to submit a pull request, but I've only "finished" (in quotes because I expect we'll do revisions as we review the code) the algebra_solver for the rev case. Is there interest in the fwd case? Even if there is, I'd rather first add the rev version and then work on fwd case.
—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub, or mute the thread.
Are the arguments all sufficiently general that we'll be able to plug into Stan with double values wherever rev-mode vars are allowed?
No. One of the arguments of the function, y, is expected to contain var (it is similar to parm in the ODE solver). Statements, such as value_of(y) and y_[i] = y(i).vi_ put constraints on the type of y. The user does pass a dat (and dat_int) argument. But if he or she wants to call the algebraic solver on only data, he or she would have to rewrite the functor that gets passed in, by replacing all dependencies on y with dependencies on dat.
If we want y to be a double or var, I think we need to overload the algebraic solver (so create one version under prim/)... which is straightforward! The question is how valuable would this be to users? Would there be a model where the same system gets solved with data and parameters at another point?
There are certainly uses one could imagine in transformed data
and the generated quantities block, so yes it’s worth having a
prim version that then gets specialized for the current rev
implementation.
On Apr 10, 2017, at 12:27 PM, Charles Margossian notifications@github.com wrote:
Are the arguments all sufficiently general that we'll be able to plug into Stan with double values wherever rev-mode vars are allowed?
No. One of the arguments of the function, y, is expected to contain var (it is similar to parm in the ODE solver). Statements, such as value_of(y) and y_[i] = y(i).vi_ put constraints on the type of y. The user does pass a dat (and dat_int) argument. But if he or she wants to call the algebraic solver on only data, he or she would have to rewrite the functor that gets passed in, by replacing all dependencies on y with dependencies on dat.
If we want y to be a double or var, I think we need to overload the algebraic solver (so create one version under prim/)... which is straightforward! The question is how valuable would this be to users? Would there be a model where the same system gets solved with data and parameters at another point?
—
You are receiving this because you commented.
Reply to this email directly, view it on GitHub https://github.com/stan-dev/stan/issues/2023#issuecomment-293003442, or mute the thread https://github.com/notifications/unsubscribe-auth/ABdNlmh7xnbAiTF_X_P9cWLZS6f3JTzcks5rulh8gaJpZM4Jqgvr.
It's easier to document that way, and if it's part of some generative process, then it could be useful for simulating data.
The parameters in the ode_integrate function are allowed to be double values for just this reason. The only restriction is that the solution times and real data must be data. Generally anything that can be a var should be allowed to be data, too.
@betanalpha, @bob-carpenter : ok, I created a prim version of the algebraic solver. I realized over lunch break I would also need for Torsten. The rev version now calls the prim solver, and then builds the vari object on top.
I submitted a pull-request. I removed files which were extraneous for the request (such as the prototype fwd version of dogleg) -- but all these files are still saved on the feature/dogleg branch.
| gharchive/issue | 2016-08-23T03:11:48 | 2025-04-01T06:45:52.347671 | {
"authors": [
"betanalpha",
"bgoodri",
"billgillespie",
"bob-carpenter",
"charlesm93",
"sakrejda",
"syclik"
],
"repo": "stan-dev/stan",
"url": "https://github.com/stan-dev/stan/issues/2023",
"license": "bsd-3-clause",
"license_type": "permissive",
"license_source": "bigquery"
} |
327127181 | Feature/issue 2462 effective_sample_size interface for rstan/pystan
… old effective_sample_size to new stan-dev/math autocovariance fns
Submission Checklist
[ ] Run unit tests: ./runTests.py src/test/unit
[x] Run cpplint: make cpplint
[x] Declare copyright holder and open-source license: see below
Summary
Add new interface to effective_sample_size for RStan and PyStan.
Replace https://github.com/stan-dev/stan/pull/2477
Intended Effect
New interface.
How to Verify
make clean && ./runTests.py -j2 src/test/unit/mcmc/chains_test.cpp
Side Effects
None intended.
Documentation
Inline as function doc.
Copyright and Licensing
Please list the copyright holder for the work you are submitting (this will be you or your assignee, such as a university or company):
California State University, Chico
By submitting this pull request, the copyright holder is agreeing to license the submitted work under the following licenses:
Code: BSD 3-clause (https://opensource.org/licenses/BSD-3-Clause)
Documentation: CC-BY 4.0 (https://creativecommons.org/licenses/by/4.0/)
@bob-carpenter, when you have some time, would you please provide some feedback on the following items? Thanks.
The (old) method double effective_sample_size(const Eigen::Matrix<Eigen::VectorXd, Dynamic, 1> samples) const no longer copies the chains twice. However, the new interface double effective_sample_size(const std::vector<const double*>& chains, int num_draws) const makes one copy, which I can't seem to avoid.
If I redo the calculations within the new interface, instead of reusing the code as in this PR, I can avoid that one copy. Which would you prefer: a) duplicate code so as to avoid one copy, or b) reuse code and suffer a single copy, for easy maintenance?
Surely, I'm missing something, but how exactly is RStan/PyStan supposed to access this new private method double effective_sample_size(const std::vector<const double*>& chains, int num_draws) const?
The function that enables testing is named effective_sample_size_new, but surely there's a better name than this.
There is still a namespace level using. Every using has to be encapsulated in a function.
Does the version currently in develop copy? If not, the new version should not do any copying, even if it requires code duplication. [Sorry---couldn't follow what you meant by old and new here.]
They should access the new method however they accessed the old method. It doesn't need to be private if the privacy is the issue.
effective_sample_size_new shouldn't be in the API. What was the goal of implementing the same function with different names?
Stepping back, if all these functions are under the hood, why are we updating them? Is this being done at the request of the interface developers?
Good thought, stepping back. The intent of issue https://github.com/stan-dev/stan/issues/2462 is for RStan and PyStan to be able to outsource their effective sample size calculations. As it stands both RStan and PyStan call stan::math::autocovariance and then do the remaining calculations internally.
When I first wrote this PR, it seemed logical to overload effective_sample_size within the chains class since this issue came up in Stan. This no longer seems like a good idea to me, because I don't see how RStan/PyStan will access a member (public or private) of the chains class without instantiating the entire class. Instantiating the chains class works for CmdStan, but forcing this class on RStan or PyStan is too much.
All this is to say that Michael's analysis API idea seems like the right move, inclusive of his not piecemeal suggestion. Since it's clear that I've only recently understood the broader issue here, I'm going to step back again, see if I can grok more of the I/O concerns contained in the analysis API thread on discourse, and then try to bring forward a solution to the analysis API idea.
If you'd like me to close this PR, in an attempt to reduce clutter, please let me know.
I'll close the PR here.
I completely agree there should be some kind of posterior analysis API that lives in stan-dev/stan.
Are there functions that make sense to implement in stan-dev/math? These could then be used by an analysis API. It'd also be great to have stable implementations of correlation, covariance, and autocorrelation exposed through the Stan language.
| gharchive/pull-request | 2018-05-28T22:31:20 | 2025-04-01T06:45:52.364213 | {
"authors": [
"bob-carpenter",
"roualdes"
],
"repo": "stan-dev/stan",
"url": "https://github.com/stan-dev/stan/pull/2530",
"license": "bsd-3-clause",
"license_type": "permissive",
"license_source": "bigquery"
} |
1155806044 | [Low Priority/Optional] allow export to support more than two archs
Description
the cpu env doesn't build on an m1 macbook with native (non-rosetta'd), so i just unlocked the arch parameter. I dunno if we want to do this, or to commit an m1 file, but figured i'd toss it out there.
yeah i don't know what the right solution is. Docker is probably the most reproducible and requirements.txt is probably the most friendly. My intuition says requirements.txt for normal use + docker for actual reproducibility.
| gharchive/pull-request | 2022-03-01T20:55:34 | 2025-04-01T06:45:52.393218 | {
"authors": [
"dlwh"
],
"repo": "stanford-crfm/mistral",
"url": "https://github.com/stanford-crfm/mistral/pull/110",
"license": "Apache-2.0",
"license_type": "permissive",
"license_source": "github-api"
} |
2413623035 | IndexError: list index out of range
我在本地语料库（您提供的axciv_data.csv），大模型我使用vllm中兼容openai的接口替换原本的openai的接口，运行的参数是：
但是一直报错：
经过排查发现是knowledge_curation.py 172line中 searched_results: List[StormInformation] = self.retriever.retrieve(list(set(queries)), exclude_urls=[ground_truth_url])无法查到searched_results，searched_results=[]
继续更深一步的排查，发现是rm.py 396line中，related_docs = self.qdrant.similarity_search_with_score(query, k=self.k)，related_docs 为空，请问这是因为什么。
Traceback:
File "/Users/admin/miniforge3/envs/storm/lib/python3.11/site-packages/streamlit/runtime/scriptrunner/script_runner.py", line 535, in run_script
exec(code, module.dict)
File "/Users/admin/AI/storm/frontend/demo_light/storm.py", line 64, in
main()
File "/Users/admin/AI/storm/frontend/demo_light/storm.py", line 60, in main
CreateNewArticle.create_new_article_page()
File "/Users/admin/AI/storm/frontend/demo_light/pages_util/CreateNewArticle.py", line 78, in create_new_article_page
st.session_state["runner"].run(topic=st.session_state["page3_topic"], do_research=False,
File "/Users/admin/AI/storm/src/storm_wiki/engine.py", line 305, in run
draft_article = self.run_article_generation_module(outline=outline,
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/Users/admin/AI/storm/src/interface.py", line 376, in wrapper
result = func(*args, **kwargs)
^^^^^^^^^^^^^^^^^^^^^
File "/Users/admin/AI/storm/src/storm_wiki/engine.py", line 196, in run_article_generation_module
draft_article = self.storm_article_generation.generate_article(
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/Users/admin/AI/storm/src/storm_wiki/modules/article_generation.py", line 56, in generate_article
information_table.prepare_table_for_retrieval()
File "/Users/admin/AI/storm/src/storm_wiki/modules/storm_dataclass.py", line 161, in prepare_table_for_retrieval
self.encoded_snippets = self.encoder.encode(self.collected_snippets, show_progress_bar=False)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/Users/admin/miniforge3/envs/storm/lib/python3.11/site-packages/sentence_transformers/SentenceTransformer.py", line 565, in encode
if all_embeddings[0].dtype == torch.bfloat16:
~~~~~~~~~~~~~~^^^_
I'm wondering if what's causing the error is that you are initializing an empty "./vector_store" without filling it with the documents in the csv file. Please try adding "--update-vector-store" as one of the arguments. This will tell the model to add the documents in the csv file to the offline vector store.
I'm getting this error either running the GUI with streamlit and in CLI. Tried to remove and reinstall Pytorch, Sentence Transformers etc., but nothing works :(
It is the same error as this one - with Sentence Transformers. Now I'm unable to run it using any of the options (local, GTP, Claude).
Did you check if it’s cause by rm.py line 396 returns empty result? And check @AMMAS1 response above. Does that solve your problem?
I tried running it with --update-vector-store argument with no help. How would i know if rm.py return empty result?
I tried running it with --update-vector-store argument with no help. How would i know if rm.py return empty result? EDIT. The only option which is actually working for me is with this command : python examples/run_storm_wiki_gpt_with_VectorRM.py --output-dir /Users/admin/AI/storm --vector-db-mode offline --offline-vector-db-dir /Users/admin/AI/storm --update-vector-store --csv-file-path /Users/admin/Downloads/polished_literature_articles_detailed.csv --do-research --do-generate-outline --do-generate-article --do-polish-article
Awesome! Glad to know it worked in the end. Would you mind sharing your first command (the one with "--update-vector-store") that didn't work so we can analyze where things went wrong?
For your question, if you're using your own data (as in using VectorRM class from rm.py), rm.py should not return an empty result as long as the vector store is not empty and search_top_k parameter is not 0, which is not the default.
In order to check if your vector store has data, you can call VectorRM.get_vector_count(), and it should return the number of chunks (your documents after chunking them) in the vector store. If that is zero, then rm.py would return an empty list.
Hope that helps!
I would like to use GUI tho. But getting the same error here too:
(storm) admin@teodor--macbook-pro demo_light % streamlit run storm.py
You can now view your Streamlit app in your browser.
Local URL: http://localhost:8501
Network URL: http://172.20.10.4:8501
root : ERROR : Error occurs when searching query What is a sentinel node biopsy?: 'hits'
root : ERROR : Error occurs when searching query How is sentinel node biopsy used in colon cancer?: 'hits'
root : ERROR : Error occurs when searching query How does sentinel node biopsy help in the treatment of colon cancer?: 'hits'
root : ERROR : Error occurs when searching query What is sentinel node biopsy in colon cancer?: 'hits'
root : ERROR : Error occurs when searching query What is a sentinel node biopsy?: 'hits'
root : ERROR : Error occurs when searching query What is sentinel node biopsy?: 'hits'
root : ERROR : Error occurs when searching query How does a sentinel node biopsy work?: 'hits'
root : ERROR : Error occurs when searching query How is sentinel node biopsy used in colon cancer treatment?: 'hits'
root : ERROR : Error occurs when searching query What is the purpose of sentinel node biopsy in colon cancer treatment?: 'hits'
root : ERROR : Error occurs when searching query What is the purpose of a sentinel node biopsy?: 'hits'
root : ERROR : Error occurs when searching query What is the significance of a sentinel node biopsy in cancer treatment?: 'hits'
root : ERROR : Error occurs when searching query Why is sentinel node biopsy important in colon cancer?: 'hits'
root : ERROR : Error occurs when searching query Sentinel node biopsy in colon cancer: 'hits'
root : ERROR : Error occurs when searching query Definition of sentinel node biopsy: 'hits'
root : ERROR : Error occurs when searching query sentinel lymph node identification: 'hits'
root : ERROR : Error occurs when searching query What is a sentinel node biopsy?: 'hits'
root : ERROR : Error occurs when searching query What is a sentinel node biopsy?: 'hits'
root : ERROR : Error occurs when searching query Sentinel node biopsy in colon cancer: 'hits'
root : ERROR : Error occurs when searching query sentinel node biopsy colon cancer: 'hits'
root : ERROR : Error occurs when searching query How is a sentinel node biopsy performed?: 'hits'
root : ERROR : Error occurs when searching query Purpose of sentinel node biopsy in medical practice: 'hits'
root : ERROR : Error occurs when searching query colon cancer surgery: 'hits'
root : ERROR : Error occurs when searching query What is sentinel node biopsy?: 'hits'
root : ERROR : Error occurs when searching query What is the purpose of a sentinel node biopsy?: 'hits'
root : ERROR : Error occurs when searching query Explanation of sentinel node biopsy: 'hits'
root : ERROR : Error occurs when searching query Sentinel node biopsy in colon cancer: 'hits'
root : ERROR : Error occurs when searching query What is a sentinel node biopsy and how is it used in colon cancer diagnosis?: 'hits'
root : ERROR : Error occurs when searching query Comparative studies of sentinel node biopsy in colon cancer and other cancers: 'hits'
root : ERROR : Error occurs when searching query Definition of sentinel node biopsy: 'hits'
root : ERROR : Error occurs when searching query Benefits of sentinel node biopsy in colon cancer: 'hits'
root : ERROR : Error occurs when searching query What are the benefits and risks of a sentinel node biopsy compared to other types of biopsies for colon cancer?: 'hits'
root : ERROR : Error occurs when searching query Sentinel node biopsy in colon cancer: 'hits'
root : ERROR : Error occurs when searching query What is a sentinel node biopsy?: 'hits'
root : ERROR : Error occurs when searching query Procedure for sentinel node biopsy in colon cancer: 'hits'
root : ERROR : Error occurs when searching query How is a sentinel node biopsy performed in patients with colon cancer?: 'hits'
root : ERROR : Error occurs when searching query Current evidence supporting use of sentinel node biopsy in colon cancer: 'hits'
knowledge_storm.interface : INFO : run_knowledge_curation_module executed in 11.3001 seconds
knowledge_storm.interface : INFO : run_outline_generation_module executed in 12.6983 seconds
sentence_transformers.SentenceTransformer : INFO : Use pytorch device_name: mps
sentence_transformers.SentenceTransformer : INFO : Load pretrained SentenceTransformer: paraphrase-MiniLM-L6-v2
2024-07-18 11:57:34.601 Uncaught app exception
Traceback (most recent call last):
File "/Users/admin/miniforge3/envs/storm/lib/python3.11/site-packages/streamlit/runtime/scriptrunner/script_runner.py", line 535, in _run_script
exec(code, module.dict)
File "/Users/admin/AI/storm/frontend/demo_light/storm.py", line 60, in
main()
File "/Users/admin/AI/storm/frontend/demo_light/storm.py", line 56, in main
CreateNewArticle.create_new_article_page()
File "/Users/admin/AI/storm/frontend/demo_light/pages_util/CreateNewArticle.py", line 78, in create_new_article_page
st.session_state["runner"].run(topic=st.session_state["page3_topic"], do_research=False,
File "/Users/admin/miniforge3/envs/storm/lib/python3.11/site-packages/knowledge_storm/storm_wiki/engine.py", line 315, in run
draft_article = self.run_article_generation_module(outline=outline,
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/Users/admin/miniforge3/envs/storm/lib/python3.11/site-packages/knowledge_storm/interface.py", line 376, in wrapper
result = func(*args, **kwargs)
^^^^^^^^^^^^^^^^^^^^^
File "/Users/admin/miniforge3/envs/storm/lib/python3.11/site-packages/knowledge_storm/storm_wiki/engine.py", line 197, in run_article_generation_module
draft_article = self.storm_article_generation.generate_article(
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/Users/admin/miniforge3/envs/storm/lib/python3.11/site-packages/knowledge_storm/storm_wiki/modules/article_generation.py", line 57, in generate_article
information_table.prepare_table_for_retrieval()
File "/Users/admin/miniforge3/envs/storm/lib/python3.11/site-packages/knowledge_storm/storm_wiki/modules/storm_dataclass.py", line 162, in prepare_table_for_retrieval
self.encoded_snippets = self.encoder.encode(self.collected_snippets, show_progress_bar=False)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/Users/admin/miniforge3/envs/storm/lib/python3.11/site-packages/sentence_transformers/SentenceTransformer.py", line 565, in encode
if all_embeddings[0].dtype == torch.bfloat16:
~~~~~~~~~~~~~~^^^
IndexError: list index out of range
I'm wondering if what's causing the error is that you are initializing an empty "./vector_store" without filling it with the documents in the csv file. Please try adding "--update-vector-store" as one of the arguments. This will tell the model to add the documents in the csv file to the offline vector store. # @AMMAS1
谢谢您，如您所说，解决了这个问题
| gharchive/issue | 2024-07-17T13:25:43 | 2025-04-01T06:45:52.428280 | {
"authors": [
"AMMAS1",
"Yucheng-Jiang",
"tstanek390",
"yaojianchao"
],
"repo": "stanford-oval/storm",
"url": "https://github.com/stanford-oval/storm/issues/82",
"license": "MIT",
"license_type": "permissive",
"license_source": "github-api"
} |
2387976352 | Add filter for QdrantRM
Add filter support for QdrantRM to support qdrant search filter like
https://qdrant.tech/documentation/concepts/filtering/
uses
Thanks @hawktang !
| gharchive/pull-request | 2024-07-03T08:11:12 | 2025-04-01T06:45:52.430821 | {
"authors": [
"arnavsinghvi11",
"hawktang"
],
"repo": "stanfordnlp/dspy",
"url": "https://github.com/stanfordnlp/dspy/pull/1239",
"license": "MIT",
"license_type": "permissive",
"license_source": "github-api"
} |
2169089246 | add 'k' as argument to FaissRM.forward()
add 'k' as argument to FaissRM.forward()
LGTM! tagging @isaacbmiller as it seems like the ruff fix workflow is erroring out. Is this something on the workflow end? or does it need to be handled within the PR?
If you run ruff check --fix-only, then commit, it will pass CI.
LGTM after that
If you run ruff check --fix-only, then commit, it will pass CI.
@software-artisan
| gharchive/pull-request | 2024-03-05T12:29:58 | 2025-04-01T06:45:52.433302 | {
"authors": [
"arnavsinghvi11",
"isaacbmiller",
"software-artisan"
],
"repo": "stanfordnlp/dspy",
"url": "https://github.com/stanfordnlp/dspy/pull/559",
"license": "MIT",
"license_type": "permissive",
"license_source": "github-api"
} |
2229815275 | evaluate
Fix ModuleNotFoundError: No module named 'evaluate' after !pip install git+https://github.com/frankaging/pyreft.gitby adding this step : !pip install evaluate in the readme. It's raised a error that can cause a disgreements to the user.
ty! requirements.txt is updated! i also updated README to show verified conda env setup. it should work now with a from-scratch conda env with python3.10.
| gharchive/issue | 2024-04-07T15:17:50 | 2025-04-01T06:45:52.435041 | {
"authors": [
"frankaging",
"jeaneigsi"
],
"repo": "stanfordnlp/pyreft",
"url": "https://github.com/stanfordnlp/pyreft/issues/34",
"license": "Apache-2.0",
"license_type": "permissive",
"license_source": "github-api"
} |
2391110410 | Fix the way the UTMs are being sent
The change in the way the UTMs are being sent to align with the documentation. Discussed also with Mixpanel's support team.
@mbrucki Thank you for the PR!
| gharchive/pull-request | 2024-07-04T15:32:41 | 2025-04-01T06:45:52.436235 | {
"authors": [
"kHorozhanov",
"mbrucki"
],
"repo": "stape-io/mixpanel-tag",
"url": "https://github.com/stape-io/mixpanel-tag/pull/6",
"license": "Apache-2.0",
"license_type": "permissive",
"license_source": "github-api"
} |
447918048 | Adding lightning icon
Adding lightning icon:
looks good to me :) just one thing to ensure - was this icon verified by designers? @codepretty
| gharchive/pull-request | 2019-05-23T23:40:39 | 2025-04-01T06:45:52.437461 | {
"authors": [
"kuzhelov",
"notandrew"
],
"repo": "stardust-ui/react",
"url": "https://github.com/stardust-ui/react/pull/1385",
"license": "MIT",
"license_type": "permissive",
"license_source": "github-api"
} |
1521264471 | (WIP/Blocked) Update Ubuntu version from deprecated 18.x for "coordinator-test" GH Action (v1)
What this PR does:
Updates ref to deprecated Ubuntu version (18.04) with newer (20.04)
Which issue(s) this PR fixes:
n/a
Checklist
[x] Changes manually tested
[ ] Automated Tests added/updated
[ ] Documentation added/updated
[ ] CLA Signed: DataStax CLA
Clsoing in favor of https://github.com/stargate/stargate/pull/2378.
@tatu-at-datastax OK?
@ivansenic yes, makes sense. Thank you!
| gharchive/pull-request | 2023-01-05T19:25:11 | 2025-04-01T06:45:52.440793 | {
"authors": [
"ivansenic",
"tatu-at-datastax"
],
"repo": "stargate/stargate",
"url": "https://github.com/stargate/stargate/pull/2344",
"license": "Apache-2.0",
"license_type": "permissive",
"license_source": "github-api"
} |
2151742850 | Using motion-blinds on HA running in VBoxVM
I've a setup where HA (HAOS image) runs in a VBoxVM. I've added the support for a gate driver that I own in a new PR and I've 2 questions:
After this PR is merged what would be the necessary steps for me to make a component in HA use it?
Do you have any recommendations how to configure multicast forwarding from host to that VM?
@kopasiak I have merged your PR, released a new version of motionblinds and made the nessesary Pull Requests in HomeAssistant: https://github.com/home-assistant/core/pull/111265
I expect those to be merged before the beta cut of Home Assistant 2023.3.0.
So all you have to do is wait untill HomeAssistant 2023.3.0 will be released and update to that version.
The motionblinds integration should then start to work with your Krispol garage gate.
I am not an expert with network setup, so unfortunetely I cant help you with configuring the multicast correctly.
If you appreciate the Motionblinds integration and want to support its development, please consider sponsering the upstream library.
| gharchive/issue | 2024-02-23T20:22:43 | 2025-04-01T06:45:52.444005 | {
"authors": [
"kopasiak",
"starkillerOG"
],
"repo": "starkillerOG/motion-blinds",
"url": "https://github.com/starkillerOG/motion-blinds/issues/42",
"license": "MIT",
"license_type": "permissive",
"license_source": "github-api"
} |
2113162681 | refactor(execution, native_blockifier): make TransactionExecutor.fina…
…lize() work without Py objects
This change is
Codecov Report
All modified and coverable lines are covered by tests :white_check_mark:
:exclamation: No coverage uploaded for pull request base (barak/block_executor/transaction_executor_execute@7372c3d). Click here to learn what that means.
:exclamation: Current head c84133b differs from pull request most recent head 63b3728. Consider uploading reports for the commit 63b3728 to get more accurate results
Additional details and impacted files
@@ Coverage Diff @@
## barak/block_executor/transaction_executor_execute #1418 +/- ##
====================================================================================
Coverage ? 70.77%
====================================================================================
Files ? 60
Lines ? 7869
Branches ? 7869
====================================================================================
Hits ? 5569
Misses ? 1869
Partials ? 431
:umbrella: View full report in Codecov by Sentry.
:loudspeaker: Have feedback on the report? Share it here.
crates/native_blockifier/src/py_block_executor.rs line 125 at r1 (raw file):
Previously, elintul (Elin) wrote…
Can you explain the change?
Done.
I replace iter() with into_iter(). into_iter() takes ownership, as you suggested (:
Code quote:
.into_iter()
| gharchive/pull-request | 2024-02-01T18:26:45 | 2025-04-01T06:45:52.469550 | {
"authors": [
"barak-b-starkware",
"codecov-commenter"
],
"repo": "starkware-libs/blockifier",
"url": "https://github.com/starkware-libs/blockifier/pull/1418",
"license": "Apache-2.0",
"license_type": "permissive",
"license_source": "github-api"
} |
2758145915 | feat: add boolean operators to Result type
Adds and, and_then, or, or_else, unwrap_or_else to ResultTrait along with tests.
note: the module documentation is taken from Rust's. the item-level doc was adapted to Cairo.
This change is
| gharchive/pull-request | 2024-12-24T17:17:36 | 2025-04-01T06:45:52.471655 | {
"authors": [
"cairolover",
"reviewable-StarkWare"
],
"repo": "starkware-libs/cairo",
"url": "https://github.com/starkware-libs/cairo/pull/6921",
"license": "Apache-2.0",
"license_type": "permissive",
"license_source": "github-api"
} |
2658023332 | chore: add remote servers struct
Stack:
#2041
#2040 ⬅
#2035
#2034
#2033
#2032
#2030
#2029
#2028
⚠️ Part of a stack created by spr. Do not merge manually using the UI - doing so may have unexpected results.
This change is
✓ Commit merged in pull request #2041
| gharchive/pull-request | 2024-11-14T08:36:04 | 2025-04-01T06:45:52.475386 | {
"authors": [
"nadin-Starkware",
"reviewable-StarkWare"
],
"repo": "starkware-libs/sequencer",
"url": "https://github.com/starkware-libs/sequencer/pull/2040",
"license": "Apache-2.0",
"license_type": "permissive",
"license_source": "github-api"
} |
1401918731 | Error when uploading photo using UploadFile
Minimal reproducible example on this repo.
When posting a photo to a route with an UploadFile parameter, an exception is thrown:
@post("/")
async def hello_world(
data: UploadFile = Body(media_type=RequestEncodingType.MULTI_PART),
) -> dict[str, Any]:
form_data = await data.read()
print(len(form_data))
return {"hello": "world"}
with open("flower.jpg", "rb") as f:
data = f.read()
with httpx.Client() as client:
client.post("http://localhost:8000/", files={"data": data})
Traceback (most recent call last):
File "/home/mitch/uni/starlite-multipart-error/.venv/lib/python3.9/site-packages/starlite/middleware/exceptions.py", line 49, in __call__
await self.app(scope, receive, send)
File "/home/mitch/uni/starlite-multipart-error/.venv/lib/python3.9/site-packages/starlite/routes/http.py", line 72, in handle
response = await self._get_response_for_request(
File "/home/mitch/uni/starlite-multipart-error/.venv/lib/python3.9/site-packages/starlite/routes/http.py", line 116, in _get_response_for_request
response = await self._call_handler_function(
File "/home/mitch/uni/starlite-multipart-error/.venv/lib/python3.9/site-packages/starlite/routes/http.py", line 144, in _call_handler_function
response_data = await self._get_response_data(
File "/home/mitch/uni/starlite-multipart-error/.venv/lib/python3.9/site-packages/starlite/routes/http.py", line 168, in _get_response_data
kwargs["data"] = await request_data
File "/home/mitch/uni/starlite-multipart-error/.venv/lib/python3.9/site-packages/starlite/kwargs.py", line 540, in _get_request_data
form_data = await request.form()
File "/home/mitch/uni/starlite-multipart-error/.venv/lib/python3.9/site-packages/starlite/connection/request.py", line 140, in form
form_values = await parser()
File "/home/mitch/uni/starlite-multipart-error/.venv/lib/python3.9/site-packages/starlite_multipart/parser.py", line 84, in __call__
parse_result.extend(await self._parse_chunk())
File "/home/mitch/uni/starlite-multipart-error/.venv/lib/python3.9/site-packages/starlite_multipart/parser.py", line 65, in _parse_chunk
items.append((field_name, data.decode(self.charset)))
UnicodeDecodeError: 'utf-8' codec can't decode byte 0xad in position 2: invalid start byte
Please confirm if this issue is resolved @devmitch
All fixed, thanks @Goldziher @jtraub
| gharchive/issue | 2022-10-08T11:37:18 | 2025-04-01T06:45:52.478314 | {
"authors": [
"Goldziher",
"devmitch"
],
"repo": "starlite-api/starlite-multipart",
"url": "https://github.com/starlite-api/starlite-multipart/issues/3",
"license": "MIT",
"license_type": "permissive",
"license_source": "github-api"
} |
268002987 | cannt uninstall on windows 10 home, showing access denied.
please help me.
Antivirus? Add to exclusions?
sir,
i dont have any idea of this software. yes i have McAfee antivirus.
If you have no idea how to add something to exclusions, then get a rid of antivirus software.
sir,
i uninstalled the antivirus, still it is showing no access.
can you please connect to my PC via teamviewer and resolve my problem.
Ok, let's see, give me IDs
ID : 227 102 156
password: 1031
Thankyou soooo much sir.
Issue resolved, what was done:
Disable McAfee antivirus
Download the release again and unpack to different folder
Install rfxvmt.dll (see https://github.com/stascorp/rdpwrap/issues/194#issuecomment-325627235)
Re-install RDP Wrapper
thankyou soooooo much sir
sir,
i want to uninstall it from my computer.
and also want to delete the folder, but its not getting deleted
I think you'll be able to delete it after reboot.
complete OS reboot..?
Yes. Also try disabling McAfee for longer period (so it doesn't turn on when OS boots).
If that doesn't help, try granting access and changing permissions on blocked files.
will i have to re install OS again.?
No.
then i just have to re_start right...?
Yes. Restart == Reboot
okay after restart what should i do..?
Use uninstall.bat from that second folder.
okay,
i will restart the system and then give you my teamviewer id again.
id : 227 102 156
password : 7123
thankyou sir.
Wiped out clean.
| gharchive/issue | 2017-10-24T11:56:33 | 2025-04-01T06:45:52.494710 | {
"authors": [
"binarymaster",
"possaashton"
],
"repo": "stascorp/rdpwrap",
"url": "https://github.com/stascorp/rdpwrap/issues/332",
"license": "apache-2.0",
"license_type": "permissive",
"license_source": "bigquery"
} |
614804541 | Add REACT_APP_HUB_DESTINATION to production-ropsten env
Was missing from #1714. We may want some sort of "required env vars" check for these files before CI passes? cc @kerzhner
It's a good question of where runtime configuration should be validated. Right before a release is probably a good time.
| gharchive/pull-request | 2020-05-08T15:12:27 | 2025-04-01T06:45:52.527493 | {
"authors": [
"kerzhner",
"snario"
],
"repo": "statechannels/monorepo",
"url": "https://github.com/statechannels/monorepo/pull/1742",
"license": "MIT",
"license_type": "permissive",
"license_source": "github-api"
} |
751839826 | replaceAddressesAndBigNumberify does not allow controllable order
https://github.com/statechannels/statechannels/blob/68eb445a35e9be52a06b65c19427505c84163f9e/packages/nitro-protocol/test/test-helpers.ts#L219
If I pass in {B: 10, A: 10} it might return:
[{destination: B, amount: 10}, {destination: A, amount: 10}]
[{destination: A, amount: 10}, {destination: B, amount: 10}]
I should be able to control the order of the destinations.
Actually, this just modifies an outcome shorthand. It's on the user to properly construct the outcome from the shorthand.
| gharchive/issue | 2020-11-26T21:47:55 | 2025-04-01T06:45:52.529593 | {
"authors": [
"andrewgordstewart"
],
"repo": "statechannels/statechannels",
"url": "https://github.com/statechannels/statechannels/issues/2979",
"license": "MIT",
"license_type": "permissive",
"license_source": "github-api"
} |
677194595 | [server-wallet] Add getSigningAddress to Wallet api
getSigningAddress has the following behavior:
if at least one signing address exists, return it.
otherwise, create a signing address (and private key), store it, and return it.
On second thought, I don't think the transactions (as used) give protection against generating duplicate signing addresses given the default Postgres isolation level of Read Committed
On second thought, I don't think the transactions (as used) give protection against generating duplicate signing addresses given the default Postgres isolation level of Read Committed
Fixed in fb1c07f14.
| gharchive/pull-request | 2020-08-11T21:01:25 | 2025-04-01T06:45:52.532365 | {
"authors": [
"kerzhner"
],
"repo": "statechannels/statechannels",
"url": "https://github.com/statechannels/statechannels/pull/2453",
"license": "MIT",
"license_type": "permissive",
"license_source": "github-api"
} |
2752264309 | Convert GrpcSerializer to use buildbuf_es instead of timostamm-protobuf-ts
As I understand it buildbuf_es has the ability to use transports other than grpc. Looking forward to allowing this to run in the web we need a transport that can be used there, namely a connect transport. Every attempt has been made to preserve backwards compatibility but there are some unavoidable differences:
The transport needs to be constructed and needs a cert & key to communicate with a tls server. To this end the kernelServer.ts was modified as well to expose this data with the goal of avoiding code duplication.
A force cast was needed in saveCellExecution.ts, presumably due to graphql.ts, with the suspected data types it contains coming from the timostamm proto bindings. This file exists in a __generated-platform__ directory but I was unable to find how this file was generated and investigate further.
Let me know what you think.
Thanks!
https://github.com/stateful/runme/issues/641 documents the motivation to migrate to bufbuild_es
The transport needs to be constructed and needs a cert & key to communicate with a tls server. To this end the kernelServer.ts was modified as well to expose this data with the goal of avoiding code duplication.
I think this is temporary because right now the Executor is still using timostamm-ts; so we need to construct two different transports 1) one using timostamm-ts and 2) one using es libraries.
This should go away as soon as we migrate the Executor to use the es libraries and then we can go back to the pattern of having the Server just return a transport.
it seems these tests are failing for the same reason that my prior PR failed for. Is this your assessment as well? I ran the tests locally and got all green lights.
Hi @hotpocket, the issue is caused by the required permissions needed to run the E2E tests for the GitHub integration.
I’ve opened this PR to skip those tests specifically for forked repositories.
Let's see what @sourishkrout thinks about this solution.
it seems these tests are failing for the same reason that my prior PR failed for. Is this your assessment as well? I ran the tests locally and got all green lights.
This is a false alarm. This one 2e2 test fails because branches outside the Stateful org won't have access to a secret required to make this tests pass. Safe to ignore.
Let's see what @sourishkrout thinks about this solution.
Merged into main which will definitely help avoid false positives. Thank you!
| gharchive/pull-request | 2024-12-20T09:00:51 | 2025-04-01T06:45:52.539174 | {
"authors": [
"hotpocket",
"jlewi",
"pastuxso",
"sourishkrout"
],
"repo": "stateful/vscode-runme",
"url": "https://github.com/stateful/vscode-runme/pull/1876",
"license": "Apache-2.0",
"license_type": "permissive",
"license_source": "github-api"
} |
523339699 | next day error
the next day of expression "0 15 10 ? * 6#3" is error
Thanks for the report! Unfortunately I don't have free cycles to dig it. If this is actually a bug - I can happily accept a PR for it.
Lib does not allow for cron expression with Year and L、W、# as specified in Quartz
Support for "W" is something we actually really need.
This project is no longer maintained, see its updated and extended fork in exander77/supertinycron repo, closing the issue to archive the repo.
| gharchive/issue | 2019-11-15T08:56:49 | 2025-04-01T06:45:52.581558 | {
"authors": [
"niondir",
"staticlibs",
"xunmenging"
],
"repo": "staticlibs/ccronexpr",
"url": "https://github.com/staticlibs/ccronexpr/issues/26",
"license": "Apache-2.0",
"license_type": "permissive",
"license_source": "github-api"
} |
1149946464 | Add git hook to format and lint staged files
Make life easier by not having to think about running prettier.
bors merge
| gharchive/pull-request | 2022-02-25T01:36:55 | 2025-04-01T06:45:52.597878 | {
"authors": [
"D4nte"
],
"repo": "status-im/js-waku",
"url": "https://github.com/status-im/js-waku/pull/578",
"license": "Apache-2.0",
"license_type": "permissive",
"license_source": "github-api"
} |
366833494 | How to get posts from people I follow?
Hi guys,
I'm about to graduate. I'm developing a project about blockchain and Steemit.
I'm trying to find a way to get posts from users I follow. I've tried to use this method, but without any success. I'd like to put in the tag field 'USERNAME' a list of users, not only one username. Can someone help me?
Thank you in advance!!!
The method is :
var ul = document.getElementById('result');
var query = {
tag: USERNAME,
limit: 100
};
steem.api.getDiscussionsByFeed(query, function (err, discussions) {
//console.log(err, discussions);
if (!err) {
discussions.map(function (discussion) {
var li = document.createElement('li');
li.innerHTML = discussion.title;
ul.appendChild(li);
});
}
});
steem.api.getDiscussionsByFeed already gets the posts from users you follow if you put the username as your own.
const query = {tag: 'pjau', limit: 100};
steem.api.getDiscussionsByFeed(query, (err, response) => {
console.log(err, response);
});
You would need to filter it though, because it includes reblogs. Below will do that, but it will also remove any you follow if someone else you follow reblogs them. lol
response.forEach((r) => {
// remove any that are reblogs
if (r.reblogged_by.length == 0) {
console.log(r);
}
});
Other than that, I have no idea, maybe create an array of the accounts you want and filter out them from it?
const accounts = ['isnochys','arcange'];
const query = {tag: 'pjau', limit: 100};
steem.api.getDiscussionsByFeed(query, (err, response) => {
if (response) {
response.forEach((r) => {
// remove any that are not in the accounts array
if (accounts.includes(r.author)) {
console.log(r);
}
});
}
});
There's probably any easier way, but I don't know of it. But hopefully this helps.
Thank you Paul for your suggestions: i've tried to use your code, but whitout any success.
I can't be able to get back on my web page the posts' list...
You might have to mix-and-match different methods:
https://developers.steem.io/tutorials-javascript/get_follower_and_following_list#query
https://developers.steem.io/tutorials-javascript/blog_feed
| gharchive/issue | 2018-10-04T15:00:42 | 2025-04-01T06:45:52.944298 | {
"authors": [
"carolabianchi",
"inertia186",
"pjanaway"
],
"repo": "steemit/steem-js",
"url": "https://github.com/steemit/steem-js/issues/409",
"license": "MIT",
"license_type": "permissive",
"license_source": "github-api"
} |
455821281 | Connect to MQTT Broker with username and password
How can i enter my username and password for my MQTT Broker in the sketch?
Hey @TimMakes ! Check out the Adafruit MQTT lib it has two optional parameters for this:
Adafruit_MQTT_Client(Client *client, const char *server, uint16_t port, const char *user="", const char *pass="")
| gharchive/issue | 2019-06-13T15:52:39 | 2025-04-01T06:45:52.979303 | {
"authors": [
"TimMakes",
"stelioskat"
],
"repo": "stelioskat/esp8266-led-dimmer",
"url": "https://github.com/stelioskat/esp8266-led-dimmer/issues/2",
"license": "mit",
"license_type": "permissive",
"license_source": "bigquery"
} |
1393372664 | unblock lumendex.io
hello,
our project lumendex.io has been added to black list. this is a coordinated attack from our competitors that are reporting us for scam and phishing, which is totally untrue and erroneous.
our domain providers have been attacked with spam emails stating that we are phishing for security keys. this is totally false, nonetheless they have blocked our domains without checking the code.
this group of people are reporting fake issues.
i also dont understand how could you blacklist us without checking the reports to see if they are true?!
i am attaching
Screencast_from_10-01-2022_12_17_31_PM.webm
screen recording showing that none of the requests are going to our server.
we are just a UI for the decentralized exchange on horizon.stellar.org
please inspect the website
lumendex.online, inspect every pixel of it and unlock our lumendex.io domain and unblock our issueing wallet.
Thank you,
Best Regards!
After consultation with several community members we came to consensus regarding your project. The domain and issuing accounts will remain in the blocklist.
have you got any arguments for this?
There are claims to connections to real world assets that are not supported. Do you have factual information to support the claim that these tokens have actual connections to the real world assets?
@lumendex One of the biggest legal problems occurs when a token infringes on the Brand, Trademark, Logo of real world assets.
@lumendex There are claims to connections to real world assets that are not supported. Do you have factual information to support the claim that these tokens have actual connections to the real world assets?
can you provide evidence with those claims? We did not copy any brand or any other entity. If you claim so, pease provide evidence, otherwise you are just dumping words with no facts.
@lumendex One of the biggest legal problems occurs when a token infringes on the Brand, Trademark, Logo of real world assets.
disclaimers are posted everywhere on our website. we do not support or recommend any assets, except our own token LDEX.
| gharchive/issue | 2022-10-01T11:53:45 | 2025-04-01T06:45:53.058126 | {
"authors": [
"CryptoHobbyist",
"lumendex",
"orbitlens"
],
"repo": "stellar-expert/public-directory",
"url": "https://github.com/stellar-expert/public-directory/issues/134",
"license": "MIT",
"license_type": "permissive",
"license_source": "github-api"
} |
455359176 | services/friendbot: Fix race conditions
Race detector (enabled in #1398) found some race conditions in services/friendbot: https://circleci.com/gh/stellar/go/1985
Closed by #1401
| gharchive/issue | 2019-06-12T18:24:21 | 2025-04-01T06:45:53.059706 | {
"authors": [
"bartekn",
"debnil"
],
"repo": "stellar/go",
"url": "https://github.com/stellar/go/issues/1400",
"license": "Apache-2.0",
"license_type": "permissive",
"license_source": "github-api"
} |
164891924 | Connection reset when submitting transaction and ledger close time exceeds 60 seconds
I think it should return 202 Accepted instead. Developer would have to resubmit a transaction to check it's status.
But I think scott's point was that the person developing against horizon is going to have to handle the time out anyway (in cases of network failure etc) so why bother to have this extra error state?
Because it's confusing. Connection reset may be caused by problems with Horizon but also by problems within the network of organization that's using Horizon. 202 Accepted is a clear indication that the transaction was accepted but another request must be sent to check it's status.
I don't understand why you need to know that horizon accepted the transaction. The fact that it accepted the transaction is zero guarantee that it will show up in the ledger. While I might be convinced to presenting a 500-class error on timeout at horizon occurs because of consensus lag, I am adamantly against returning something that is in the 200-class: Look at most code that checks for "was the request a success" and you'll see something naive like:
if code >= 200 && code < 300 {
doSomeThingOnSuccess()
}
All that code will start behaving incorrectly if we return a 202 for transactions that will potentially fail.
It's worth noting that the whole point of horizon's transaction posting subsystem is to simplify integration... If you prefer to manage the state yourself, just post directly to stellar core and deal with all the complexity/power.
Yeah, maybe 202 Accepted is not a right choice. But I think returning 500-class error can be a good idea. My point is that connection reset gives you no idea what happened: ledger close delays/internal network issues/external network issues/horizon crash and so on.
Yep, I think that is valid. It will reduce the instances of feedback-less errors while keeping the integration requirements the same. I'll look into what it takes to fix later today. IIRC, the choice to let the golang http layer time the connection out was made because that was the default behavior and required no work.
| gharchive/issue | 2016-07-11T17:33:25 | 2025-04-01T06:45:53.063837 | {
"authors": [
"bartekn",
"jedmccaleb",
"nullstyle"
],
"repo": "stellar/horizon",
"url": "https://github.com/stellar/horizon/issues/292",
"license": "apache-2.0",
"license_type": "permissive",
"license_source": "bigquery"
} |
2209028421 | display the correct soroban network
our current laboratory isn't showing the correct network for futurenet. It displays testnet.
Before:
After:
Preview is available here:http://laboratory-pr796.previews.kube001.services.stellar-ops.com/
| gharchive/pull-request | 2024-03-26T18:37:06 | 2025-04-01T06:45:53.066686 | {
"authors": [
"jeesunikim",
"stellar-jenkins"
],
"repo": "stellar/laboratory",
"url": "https://github.com/stellar/laboratory/pull/796",
"license": "Apache-2.0",
"license_type": "permissive",
"license_source": "github-api"
} |
1551447971 | contractimport fails when importing custom enums with discriminants
What version are you using?
0.4.3 but same with older versions (oldest tested 0.3.2)
What did you do?
Create a contract using a custom type enum with custom discriminants.
i.e.
#[contracttype]
#[derive(Debug, Copy, Clone, Eq, PartialEq, Ord, PartialOrd)]
#[repr(u32)]
pub enum MyCoolType {
ThisTypeIsSadBecauseTheNameIsTooLong = 0,
Type = 1
}
Import the generated WASM file using the soroban_sdk::contractimport macro.
What did you expect to see?
Everything should compile just fine.
What did you see instead?
$ RUSTFLAGS="-Z macro-backtrace" cargo +nightly test
error[E0507]: cannot move out of `*self` which is behind a shared reference
--> buggy/src/lib.rs:4:5
|
4 | / soroban_sdk::contractimport!(
5 | | file = "../target/wasm32-unknown-unknown/release/soroban_error_types.wasm"
6 | | );
| | ^
| | |
| |_____move occurs because `*self` has type `MyCoolType`, which does not implement the `Copy` trait
| in this procedural macro expansion
|
::: /home/kai/.cargo/registry/src/github.com-1ecc6299db9ec823/soroban-sdk-macros-0.3.2/src/lib.rs:123:1
|
123 | pub fn contracttype(metadata: TokenStream, input: TokenStream) -> TokenStream {
| ----------------------------------------------------------------------------- in this expansion of `#[soroban_sdk::contracttype]`
Addition Information
An example implementation for this bug is given here https://github.com/Smephite/soroban_tests.
It seems to likely be related to #758 .
The bug seems to be fixed when adding the Copy trait to the derived values on the contract types.
This issue will be resolved by https://github.com/stellar/rs-soroban-sdk/issues/630.
| gharchive/issue | 2023-01-20T21:03:01 | 2025-04-01T06:45:53.071149 | {
"authors": [
"Smephite",
"leighmcculloch"
],
"repo": "stellar/rs-soroban-sdk",
"url": "https://github.com/stellar/rs-soroban-sdk/issues/834",
"license": "Apache-2.0",
"license_type": "permissive",
"license_source": "github-api"
} |
2563938839 | Changes to support the API changes upcoming in rpc v22
The rpc client needs updating to support the following changes that are upcoming in rpc v22:
createdAt field in getTransaction response is now encoded as string instead of int64
The fields in getVersionInfo response were changed to Camel-case naming:
commitHash
buildTimestamp
captiveCoreVersion
protocolVersion
The legacy cost object has been removed from simulateTransaction response.
The correct resource costs can now be retrieved from the transaction XDR instead.
Remove pagingToken from getEvents response and replace it with Cursor.
Instead of being present in all event objects, the Cursor will be at the top level of the response similar to getTransactions. This makes it easier to use and brings it in sync with how Cursor is used in other RPC endpoints
Deprecated getLedgerEntry endpoint is now removed.
Users can use the more powerful getLedgerEntries endpoint to get the same result.
Add transaction hash to getTransactions response.
Each transaction object in the response now also includes a hex-encoded transaction hash string.
Ref: https://gist.github.com/aditya1702/8a0e3e05689217009692517d744c7f10
Related discussion:
https://stellarfoundation.slack.com/archives/C0445GF868N/p1727957758063739?thread_ts=1727909447.005259&cid=C0445GF868N
The above list of changes include breaking changes. Some of the breaking changes are changes to the response fields, and there is not time period for where both old and new respond fields will be present to allow clients to adopt just one or the other at any one time, so when updating the client to support the new fields the client must be able to support both the old and new fields.
This may require rewriting the way the client parses the JSON, introducing custom parsers, or parsing some fields into dynamic json values, and then parsing them manually to one of the types we expect.
This change is a dependency of releasing the stellar-cli for protocol 22.
See #https://github.com/stellar/rs-stellar-rpc-client/pull/12
@willemneal Missed with Molly. Platform team is unavailable next week as the whole team is offsite. So I think given you already opened the PR, I assigned this to you.
I'm confused why need backwards compatibility. When pointing at an RPC endpoint that supports v22, why would we need to support the old fields?
A developer might be building for mainnet and testnet at the same time. They'll have a CLI installed, and new CLI's should be backwards compatible with the current network in use. Otherwise we end up with a poor developer experience where developers need to have multiple versions installed, actively switch between them.
Also, operators will need to deploy the new RPC ahead of network upgrades, and operators of those RPC shouldn't need to coordinate with users of the CLI. Users should be able to grab the new CLI, then at some point operators upgrade the RPC and the user is largely unaffected or doesn't care.
If needed couldn't the old client be used with an older RPC endpoint?
Most of the breaking changes are changes only in the response, which means the CLI can't know which client to use until it gets a response back and need to parse it.
A developer might be building for mainnet and testnet at the same time.
That's like saying prod is running Python 3 but staging is running Python 2 but you want the same python binary to support both 🧐 It's pretty standard to use different binaries for different targets - expecting a v22 and v21 CLI if there are different protocols on each one seems okay to me.
operators of those RPC shouldn't need to coordinate with users of the CLI
Except we expect this for every protocol upgrade? If we (SDF) deploy a v22 RPC, it's well-understood that everyone needs to update their SDKs according. Why is the CLI different?
new CLI's should be backwards compatible with the current network in use
Wouldn't this logic apply to all clients and SDKs, which certainly isn't the case?
| gharchive/issue | 2024-10-03T12:32:11 | 2025-04-01T06:45:53.082391 | {
"authors": [
"Shaptic",
"janewang",
"leighmcculloch",
"willemneal"
],
"repo": "stellar/rs-stellar-rpc-client",
"url": "https://github.com/stellar/rs-stellar-rpc-client/issues/16",
"license": "Apache-2.0",
"license_type": "permissive",
"license_source": "github-api"
} |
2546842061 | Add conversions from AccountId and PublicKey to MuxedAccount
What
Add conversions from AccountId and PublicKey to MuxedAccount.
Why
The types are largely compatible and while MuxedAccounts are used in transactions applications often only support G addresses and convenient conversions from the AccountId/PublicKey type used for G addresses to the MuxedAccount type for use in transactions would be helpful.
@willemneal What other conversions would be helpful iyo?
| gharchive/pull-request | 2024-09-25T03:58:30 | 2025-04-01T06:45:53.084559 | {
"authors": [
"leighmcculloch"
],
"repo": "stellar/rs-stellar-xdr",
"url": "https://github.com/stellar/rs-stellar-xdr/pull/396",
"license": "Apache-2.0",
"license_type": "permissive",
"license_source": "github-api"
} |
657640435 | Add show/hide functionality to JSON in logs
The SEP24 Demo Client uses a show/hide link to keep the JSON shown in logs neat - we need to port this over to the Stellar demo wallet.
We use the renderjson package for this in the demo client, should work fine here too.
const Renderjson = require("renderjson");
Renderjson.set_show_to_level(1);
div.appendChild(Renderjson(params));
https://github.com/stellar/sep24-demo-client/blob/master/src/ui/ui-actions.js#L110
Looks like this actually already happened.
| gharchive/issue | 2020-07-15T20:40:52 | 2025-04-01T06:45:53.091912 | {
"authors": [
"JakeUrban",
"lydiat",
"msfeldstein"
],
"repo": "stellar/stellar-demo-wallet",
"url": "https://github.com/stellar/stellar-demo-wallet/issues/24",
"license": "Apache-2.0",
"license_type": "permissive",
"license_source": "github-api"
} |
1950695387 | SDP-902 - Release 1.0.0 to main
What
Release 1.0.0-rc2 to main.
Checklist
PR Structure
[x] This PR has reasonably narrow scope (if not, break it down into smaller PRs).
[x] This PR does not mix refactoring changes with feature changes (split into two PRs otherwise).
[X] This PR's title starts with the name of the package, area, or subject affected by the change.
Thoroughness
[x] This PR adds tests for the new functionality or fixes.
[x] This PR contains the link to the Jira ticket it addresses.
Configs and Secrets
[ ] No new CONFIG variables are required -OR- the new required ones were added to the helmchart's values.yaml file.
[ ] No new CONFIG variables are required -OR- the new required ones were added to the deployments (pr-preview, dev, demo, prd).
[ ] No new SECRETS variables are required -OR- the new required ones were mentioned in the helmchart's values.yaml file.
[ ] No new SECRETS variables are required -OR- the new required ones were added to the deployments (pr-preview secrets, dev secrets, demo secrets, prd secrets).
Release
[ ] This is not a breaking change.
[ ] This is ready for production.. If your PR is not ready for production, please consider opening additional complementary PRs using this one as the base. Only merge this into develop or main after it's ready for production!
Deployment
[ ] Does the deployment work after merging?
stellar-disbursement-platform-backend-preview is available here:SDP: https://sdp-backend-pr74.previews.kube001.services.stellar-ops.com/healthAP: https://sdp-ap-pr74.previews.kube001.services.stellar-ops.com/healthFrontend: https://sdp-backend-dashboard-pr74.previews.kube001.services.stellar-ops.com
| gharchive/pull-request | 2023-10-18T21:28:25 | 2025-04-01T06:45:53.101874 | {
"authors": [
"marwen-abid",
"stellar-jenkins"
],
"repo": "stellar/stellar-disbursement-platform-backend",
"url": "https://github.com/stellar/stellar-disbursement-platform-backend/pull/74",
"license": "Apache-2.0",
"license_type": "permissive",
"license_source": "github-api"
} |
1326398966 | Registrate v0.1.0
@JuliaRegistrator register
@JuliaRegistrator register
Error while trying to register: "Tag with name v0.1.0 already exists and points to a different commit"
@JuliaRegistrator register branch=v0.1.0
@JuliaRegistrator register
Registration pull request created: JuliaRegistries/General/65513
After the above pull request is merged, it is recommended that a tag is created on this repository for the registered package version.
This will be done automatically if the Julia TagBot GitHub Action is installed, or can be done manually through the github interface, or via:
git tag -a v0.1.0 -m "<description of version>" c484fbd1517ae89e8656ab571f26ced87662a634
git push origin v0.1.0
@JuliaRegistrator register
Registration pull request updated: JuliaRegistries/General/65513
After the above pull request is merged, it is recommended that a tag is created on this repository for the registered package version.
This will be done automatically if the Julia TagBot GitHub Action is installed, or can be done manually through the github interface, or via:
git tag -a v0.1.0 -m "<description of version>" 481d7f716a1f0ad92f20af2e4d8f4d43af962a4c
git push origin v0.1.0
| gharchive/issue | 2022-08-02T21:04:40 | 2025-04-01T06:45:53.156505 | {
"authors": [
"JuliaRegistrator",
"stepanzh"
],
"repo": "stepanzh/ScoredTests.jl",
"url": "https://github.com/stepanzh/ScoredTests.jl/issues/1",
"license": "MIT",
"license_type": "permissive",
"license_source": "github-api"
} |
1556743403 | Unable to flash OTA
Hi all,
I had my board flashed and working fine with my E4 door. I saw there were changes and I decided to flash the board OTA, like always. I made a mistake with the config.h and now the board doesn’t work, but, I can either find the way to reflash it OTA, all the ports are close now. I guess I will need to solder the pins to have the serial working, but, what I did wrong? I remember, when I had this issues before, the board can enter in a recovery mode if it doesn’t detect the wifi, but not anymore… Any advice?
Thanks,
Hi @i5Js,
are you sure you got it working with an E4? If really you would be the first one ever. See #15
Which recovery mode, I don't know any recovery mode? If OTA isn't working you have to flash it via serial port.
Are you sure you opend your issue in the right Github project?
Ok, now all make sense. Thomas code and this code is different. His firmware had a recovery mode, and indeed, it worked in my E4. I will solder the serial port and will put his firmware back.
Thanks
Enviado desde dispositivo móvil
De: stephan192 @.>
Enviado: Wednesday, January 25, 2023 5:00:23 PM
Para: stephan192/hoermann_door @.>
Cc: i5Js @.>; Mention @.>
Asunto: Re: [stephan192/hoermann_door] Unable to flash OTA (Issue #17)
Hi @i5Jshttps://github.com/i5Js,
are you sure you got it working with an E4? If really you would be the first one ever. See #15https://github.com/stephan192/hoermann_door/issues/15
Which recovery mode, I don't know any recovery mode? If OTA isn't working you have to flash it via serial port.
Are you sure you opend your issue in the right Github project?
—
Reply to this email directly, view it on GitHubhttps://github.com/stephan192/hoermann_door/issues/17#issuecomment-1403843903, or unsubscribehttps://github.com/notifications/unsubscribe-auth/AAIFKG5WUCYKY5DOZ2SF3QTWUFEZPANCNFSM6AAAAAAUGMGKKE.
You are receiving this because you were mentioned.Message ID: @.***>
| gharchive/issue | 2023-01-25T14:23:18 | 2025-04-01T06:45:53.163577 | {
"authors": [
"i5Js",
"stephan192"
],
"repo": "stephan192/hoermann_door",
"url": "https://github.com/stephan192/hoermann_door/issues/17",
"license": "MIT",
"license_type": "permissive",
"license_source": "github-api"
} |
1062463506 | Compilation error on Raspberry Pi 4 / CM4 : cannot convert ‘const long long int*’ to ‘const time_t*’ {aka ‘const long int*’}
I stumbled on this compilation fatal error (output.txt) while trying to compile Darkhelp on Compute Module 4 (armv7l Raspbian).
This error occurred in 2 files : first in DarkHelpServer.cpp and secondly in DarkHelpCli.cpp
I bypassed this error by editing line 230 in DarkHelpServer.cpp and line 1046 in DarkHelpCli.cpp https://github.com/stephanecharette/DarkHelp/blob/318e33d4fe7b97ec02e1086bac7c3adfd73ec74d/src-tool/DarkHelpServer.cpp#L230
https://github.com/stephanecharette/DarkHelp/blob/318e33d4fe7b97ec02e1086bac7c3adfd73ec74d/src-tool/DarkHelpCli.cpp#L1046
I wrote in both files:
std::time_t t = seconds;
const auto lt = std::localtime(&t);
Can you check if my update is correct and update these two files if it complies with your code ?
Thank you in advance.
PS : Thank you for your job done on this project, it's incredible how it's easy to use darknet with your help.
Thanks. I made the change in the latest version. Is this a 32-bit distro?
Nice, thank you !
Yes, it is a 32-bit distro !
That explains the problem. On my RPI 4 I'm running 64-bit Ubuntu, which is why I never saw this problem.
Ok, good to know !
You can mark this issue as resolved with your last commit if you want.
| gharchive/issue | 2021-11-24T14:00:34 | 2025-04-01T06:45:53.171968 | {
"authors": [
"QuietLullaby",
"stephanecharette"
],
"repo": "stephanecharette/DarkHelp",
"url": "https://github.com/stephanecharette/DarkHelp/issues/19",
"license": "MIT",
"license_type": "permissive",
"license_source": "github-api"
} |
2627577018 | 🛑 BISPM is down
In 85d2931, BISPM (http://www.bispm.co.uk) was down:
HTTP code: 403
Response time: 1087 ms
Resolved: BISPM is back up in 552a042 after 9 minutes.
| gharchive/issue | 2024-10-31T18:38:31 | 2025-04-01T06:45:53.178160 | {
"authors": [
"stephenmott"
],
"repo": "stephenmott/webmonitor",
"url": "https://github.com/stephenmott/webmonitor/issues/63",
"license": "MIT",
"license_type": "permissive",
"license_source": "github-api"
} |
449505439 | Support circular type dependencies
BREAKING - type config option required to be a thunk
:tada: This PR is included in version 6.0.0 :tada:
The release is available on:
npm package (@latest dist-tag)
GitHub release
Your semantic-release bot :package::rocket:
| gharchive/pull-request | 2019-05-28T22:27:43 | 2025-04-01T06:45:53.180398 | {
"authors": [
"stephentuso"
],
"repo": "stephentuso/ts-graphql",
"url": "https://github.com/stephentuso/ts-graphql/pull/21",
"license": "MIT",
"license_type": "permissive",
"license_source": "github-api"
} |
1397935304 | Update SpectaQL example
Closes #49
Re-generate example static GraphQL API documentation using spectaql@1.5.4.
@royderks
Hi @daniel-shuy could you tell me what GraphQL API you've used? It would be great to add the GraphQL files to the example or deploy it on a public endpoint
@royderks the GraphQL files are already in the example (in the /stepzen folder), I simply reused them.
To re-generate the static GraphQL API documentation I deployed it to https://publica1703e03faa4e8b6.stepzen.net/api/frankfurter/__graphql
@royderks the GraphQL files are already in the example (in the /stepzen folder), I simply reused them.
To re-generate the static GraphQL API documentation I deployed it to https://publica1703e03faa4e8b6.stepzen.net/api/frankfurter/__graphql
Hey that's awesome, my bad completely overlooked this. Great work!
| gharchive/pull-request | 2022-10-05T14:54:52 | 2025-04-01T06:45:53.183760 | {
"authors": [
"daniel-shuy",
"royderks"
],
"repo": "stepzen-dev/examples",
"url": "https://github.com/stepzen-dev/examples/pull/53",
"license": "MIT",
"license_type": "permissive",
"license_source": "github-api"
} |
2232576842 | setup function is nil
Describe the bug
A clear and concise description of what the bug is.
To Reproduce
Steps to reproduce the behavior:
Add gkeep.nvim example lazy config to nvim 0.10
Start nvim and sync/update lazy
Restart nvim
receive message:
Failed to run `config` for gkeep.nvim
.../Local/nvim-data/lazy/lazy.nvim/lua/lazy/core/loader.lua:373: attempt to call field 'setup' (a nil value)
# stacktrace: - init.lua:260
Expected behavior
gkeep would load without error
Version information
OS: [e.g. linux, mac, windows]
Windows 11 Pro
Neovim: [output of nvim -v]
NVIM v0.10.0-dev-2826+gf6dcc464f
Gkeep: [output of :checkhealth gkeep]
gkeep: health#gkeep#check
gkeep ~
- OK gkeepapi 0.15.1 installed
- OK gpsoauth 1.0.4 installed
- OK urllib3 1.26.18 installed
- OK keyring installed
- ERROR Remote plugin not found
- ADVICE:
- Try running :UpdateRemotePlugins and restart
Additional context
I'm not sure if this is related to the lack of setup, but I don't get any plugins listed in the rplugin.vim after running UpdateRemotePlugins
You are probably using lazy.nvim and have set config = true or opts = {}. Either of those values causes lazy to call require("gkeep").setup(), but setup is not a function that exists as all configuration is done by other means in this plugin.
ya, that was it. 😞
Sorry about the erroneous bug report.
| gharchive/issue | 2024-04-09T04:26:20 | 2025-04-01T06:45:53.190557 | {
"authors": [
"stevearc",
"tanj"
],
"repo": "stevearc/gkeep.nvim",
"url": "https://github.com/stevearc/gkeep.nvim/issues/19",
"license": "MIT",
"license_type": "permissive",
"license_source": "github-api"
} |
2254834 | Error in require
I am trying to test my generator using the gem but I get this error when generator_spec/test_case is required:
/Users/andrea/.rvm/gems/ruby-1.9.2-p180-patched@global/gems/rspec-rails-2.6.1/lib/rspec/rails/extensions/active_record/base.rb:26:in <top (required)>': uninitialized constant ActiveModel (NameError) from /Users/andrea/.rvm/gems/ruby-1.9.2-p180-patched@global/gems/rspec-rails-2.6.1/lib/rspec/rails/extensions.rb:1:in require'
from /Users/andrea/.rvm/gems/ruby-1.9.2-p180-patched@global/gems/rspec-rails-2.6.1/lib/rspec/rails/extensions.rb:1:in <top (required)>' from /Users/andrea/.rvm/gems/ruby-1.9.2-p180-patched@global/gems/rspec-rails-2.6.1/lib/rspec/rails.rb:8:in require'
from /Users/andrea/.rvm/gems/ruby-1.9.2-p180-patched@global/gems/rspec-rails-2.6.1/lib/rspec/rails.rb:8:in <top (required)>' from /Users/andrea/.rvm/gems/ruby-1.9.2-p180-patched/gems/generator_spec-0.8.4/lib/generator_spec/test_case.rb:1:in require'
from /Users/andrea/.rvm/gems/ruby-1.9.2-p180-patched/gems/generator_spec-0.8.4/lib/generator_spec/test_case.rb:1:in <top (required)>' from /Users/andrea/dev/kemen/virgilio/spec/generators/babel_add_field_generator_spec.rb:1:in require'
from /Users/andrea/dev/kemen/virgilio/spec/generators/babel_add_field_generator_spec.rb:1:in `<top (required)>'
Any suggestion?
The issue is that active model isn't loaded when rspec-rails expects it to be. What are you requiring at the top of application.rb? It's normally something like require 'rails/all'...
| gharchive/issue | 2011-11-16T10:41:32 | 2025-04-01T06:45:53.215427 | {
"authors": [
"andreacfm",
"stevehodgkiss"
],
"repo": "stevehodgkiss/generator_spec",
"url": "https://github.com/stevehodgkiss/generator_spec/issues/5",
"license": "mit",
"license_type": "permissive",
"license_source": "bigquery"
} |
1741957029 | hooks: explicitly handle axios errors
We were doing mostly OK by letting axios just throw exceptions, but that eats the error messages and does not disambiguate between errors on a response, errors sending the request, and errors creating the reqeust object. Axios could just publish a helpful error, but whatever.
Fixes #281
Obsoletes #284
| gharchive/pull-request | 2023-06-05T14:54:54 | 2025-04-01T06:45:53.216654 | {
"authors": [
"stevekuznetsov"
],
"repo": "stevekuznetsov/avalanche-forecast",
"url": "https://github.com/stevekuznetsov/avalanche-forecast/pull/286",
"license": "MIT",
"license_type": "permissive",
"license_source": "github-api"
} |
61156162 | add basic test for preid
@stevelacy wrote a super basic test, let me know if you would want to do more
Oh my we wrote them at the same time!
I'll merge it in
haha! thanks man, thanks for the responsiveness! let me know when you bump so i can point back at the standard package!
@scamden published to 0.3.0
thanks!
| gharchive/pull-request | 2015-03-13T17:44:53 | 2025-04-01T06:45:53.218356 | {
"authors": [
"scamden",
"stevelacy"
],
"repo": "stevelacy/gulp-bump",
"url": "https://github.com/stevelacy/gulp-bump/pull/37",
"license": "mit",
"license_type": "permissive",
"license_source": "bigquery"
} |
586427062 | Some United States Region data is Nil
The United States Region breakdown data is null for Recovered, and Deaths columns.
Currently there are no reliable sources for recovered numbers, but I'll add number of deaths soon.
Data for deaths have been added.
| gharchive/issue | 2020-03-23T18:31:53 | 2025-04-01T06:45:53.233631 | {
"authors": [
"papahabla",
"stevenliuyi"
],
"repo": "stevenliuyi/covid19",
"url": "https://github.com/stevenliuyi/covid19/issues/10",
"license": "MIT",
"license_type": "permissive",
"license_source": "github-api"
} |
215259191 | EvictSomeItems does not count leaf node in numEvicted
If I understand correctly, how it works, it should be:
diff --git a/collection.go b/collection.go
index 6e92191..1a91b0c 100644
--- a/collection.go
+++ b/collection.go
@@ -245,6 +245,7 @@ func (t *Collection) EvictSomeItems() (numEvicted uint64) {
})
if i != nil && err != nil {
t.store.ItemDecRef(t, i)
+ numEvicted++
}
return numEvicted
}
Hmm, I am not sure that this is correct. The actual eviction happens in CompareAndSwapPointer, not in ItemDecRef.
| gharchive/issue | 2017-03-19T11:58:57 | 2025-04-01T06:45:53.251869 | {
"authors": [
"starius"
],
"repo": "steveyen/gkvlite",
"url": "https://github.com/steveyen/gkvlite/issues/14",
"license": "mit",
"license_type": "permissive",
"license_source": "bigquery"
} |
1588400721 | Similarity matrix shape does not match the shape of the mask
Hello,
I was currently testing the implementation when an error occured: The shape of the mask [512, 512] at index 0 does not match the shape of the indexed tensor [2, 2] at index 0.
My batch size is 256.
The error occurs in this part of the code:
similarity_matrix = torch.matmul(features, features.T)
mask = torch.eye(labels.shape[0], dtype=torch.bool).to(device)
labels = labels[~mask].view(labels.shape[0], -1)
similarity_matrix = similarity_matrix[~mask].view(similarity_matrix.shape[0], -1)
I'm wondering if this something I'm doing wrong and how do I match the shape of tensors?
Thanks in advance!
same question~
The issue is because of the following line of code
https://github.com/sthalles/SimCLR/blob/1848fc934ad844ae630e6c452300433fe99acfd9/simclr.py#L28
which assumes that the number of features you have is a multiple of the batch size but this is not always true.
For example, consider a dataset with only a 100 elements and a batch size of 256. In that case it will create a labels of size (256, 256) even though it should only be (200, 200) (assuming you are using n_views of 2). The way to resolve this is by updating the above line to:
labels = torch.cat([torch.arange(int(features.size(0)/2)) for i in range(self.hparams["num_views"])], dim=0)
| gharchive/issue | 2023-02-16T21:41:26 | 2025-04-01T06:45:53.269357 | {
"authors": [
"hugofigueiras",
"laiyingxin2",
"sarda-devesh"
],
"repo": "sthalles/SimCLR",
"url": "https://github.com/sthalles/SimCLR/issues/60",
"license": "MIT",
"license_type": "permissive",
"license_source": "github-api"
} |
1729467485 | Fix lingering class-loading issues with CableReady::Updatable concern
Bugfix
Description
There's still some weirdness with the way the MemoryCacheDebounceAdapter is being loaded in CableReady::Updatable concern. Notes on the previous fix here: https://github.com/stimulusreflex/cable_ready/pull/274#issuecomment-1565767491
It seems to be resolved now, I'm not seeing the intermittent errors after this tweak :+1:
Checklist
[x] My code follows the style guidelines of this project
[x] Checks (StandardRB & Prettier-Standard) are passing
Ah! I think I was switching between too many branches when testing and one of them was still on 5.0.0-rc2 so the MemoryCacheDebounceAdapter was still under /lib... I'll take another look.
Yeah I was on the wrong branch. Sorry for the false alarm!
Thanks for doubling-checking, @Matt-Yorkley!
| gharchive/pull-request | 2023-05-28T13:25:34 | 2025-04-01T06:45:53.279748 | {
"authors": [
"Matt-Yorkley",
"marcoroth"
],
"repo": "stimulusreflex/cable_ready",
"url": "https://github.com/stimulusreflex/cable_ready/pull/277",
"license": "MIT",
"license_type": "permissive",
"license_source": "github-api"
} |
2584032851 | Imported component with bottomsheet has no closing animation slide
When creating a component with a bottomsheet in it and then importing that component into another component, the sliding animation that happens onClose is not happening, instead the bottomSheet just disappears instantly. the opening animation works correctly tho, Any ideas on this ? Thanks
The issue was because the component Bundles with BottomSheet was wrapped inside a conditional rendering like this {isBundlesOpen && ( <Bundles onClose={CloseBundles} isOpen={isBundlesOpen} title="You're out of taps" // Pass the title as prop /> )} removing the unecessary isBundlesOpen solved the issue.
| gharchive/issue | 2024-10-13T14:55:41 | 2025-04-01T06:45:53.283155 | {
"authors": [
"rodboro777"
],
"repo": "stipsan/react-spring-bottom-sheet",
"url": "https://github.com/stipsan/react-spring-bottom-sheet/issues/301",
"license": "MIT",
"license_type": "permissive",
"license_source": "github-api"
} |
2202111497 | Use different browser provider
Hey there! Ive modifed the code to use serper's api but it returns the result witrh a URL But for some reaosn it returns None when it searches the query correctly could someone guide me to modify it again
it's a serper api problem. check their docs.
| gharchive/issue | 2024-03-22T09:51:09 | 2025-04-01T06:45:53.284161 | {
"authors": [
"ARajgor",
"criticalswateam1"
],
"repo": "stitionai/devika",
"url": "https://github.com/stitionai/devika/issues/41",
"license": "MIT",
"license_type": "permissive",
"license_source": "github-api"
} |
1789551974 | Fix typos in GPIO docs
GPIO doc incorrectly states that in PullDown mode the input is connected to high.
bors r+
| gharchive/pull-request | 2023-07-05T13:21:53 | 2025-04-01T06:45:53.288847 | {
"authors": [
"burrbull",
"jorpic"
],
"repo": "stm32-rs/stm32f1xx-hal",
"url": "https://github.com/stm32-rs/stm32f1xx-hal/pull/465",
"license": "Apache-2.0",
"license_type": "permissive",
"license_source": "github-api"
} |
126090516 | Feature Suggestion: Go back to being original user.
Nice job thanks!
You're already storing the original user.pk in the session, it would be great to be able to go back to being who you were without logging out. If there's no objection I'd be happy to add a view for that.
Then perhaps a context processor to put html for a button in the context when the user is being impersonated. It would be up to the application developer to decide where to place the button so that anyone will be able to see it, since that would vary for every application.
Hmm, this is going to allow privilege escalation if a user somehow manages to store that flag in their session, so I'm not thrilled about it. However, this is useful functionality, so I'd be okay with it if it were optional, off by default and clearly communicated in the documentation that it's potentially insecure.
What do you think?
How about signing the session information using django.core.signing? Then it would be as secure as any other django login.. the difficulty of forging the session would be the same as the difficulty of forging the login cookie itself.
But if, even with that, you'd rather have it disabled by default, I'm totally ok with that.
That's actually a great idea. Can you make sure to include the original user ID, the target user ID and a timestamp in the signature? It should be something like:
request.session[session_flag] = "%s|%s|%s|%s" % (original_user_pk, user.pk, time.time(), signature)
And then, to validate:
original_id, user_id, timestamp, signature = request.session[session_flag].split("|")
if timestamp < time.time() - 24 * 3600 or user_id != request.user.id: <deny>
Something like that would probably provide enough security that there wouldn't be problems. I'd still prefer it if this were off by default, since it's really easy to turn it on (and the user would probably have to set up the view anyway), but it's not a requirement.
Ok, I'm not sure that the time aspect is really needed.. the session already has a timeout that can be configured. This is an admin function, at some point I feel it is up to the user to take responsibility...
The expiration isn't for the data, it's for the signature itself. It's so someone can't steal the signature and replay it at a later time. It's probably overkill, but it's so simple to add that it's probably not worth omitting.
Hi, sorry I went silent.. things got busy here but I still need this feature and going to get to it in the next week or two.
Sure, no problem. I think implementing the feedback above is enough for me to merge and release, thanks!
Yes, would greatly appreciate if this functionality is added and released :) Oh look, someone added it :)
Looks good overall, there are some minor issues I've left comments on. If you can fix those and the failing tests, I can merge, thank you!
Merged #31, thanks @benjaoming!
| gharchive/issue | 2016-01-12T03:44:06 | 2025-04-01T06:45:53.296712 | {
"authors": [
"benjaoming",
"benmehlman",
"skorokithakis"
],
"repo": "stochastic-technologies/django-loginas",
"url": "https://github.com/stochastic-technologies/django-loginas/issues/29",
"license": "bsd-3-clause",
"license_type": "permissive",
"license_source": "bigquery"
} |
1756056571 | xturing-0.1.4 BaseModel.create error
There is no problem when using xturing-0.1.3, but the following error occurs when upgrading to xturing-0.1.4
from xturing.datasets.instruction_dataset import InstructionDataset
from xturing.models.base import BaseModel
instruction_dataset = InstructionDataset("alpaca_data")
model = BaseModel.create("bloom_lora")
It gives an error:
/home/tl_admin/xturing_test/train.py:5 in <module> │
│ │
│ 2 from xturing.models.base import BaseModel │
│ 3 │
│ 4 instruction_dataset = InstructionDataset("alpaca_data") │
│ ❱ 5 model = BaseModel.create("bloom_lora") │
│ 6 │
│ 7 finetuning_config = model.finetuning_config() │
│ 8 │
│ │
│ /home/tl_admin/.local/lib/python3.8/site-packages/xturing/registry.py:14 in create │
│ │
│ 11 │ │
│ 12 │ @classmethod │
│ 13 │ def create(cls, class_key, *args, **kwargs): │
│ ❱ 14 │ │ return cls.registry[class_key](*args, **kwargs) │
│ 15 │ │
│ 16 │ @classmethod │
│ 17 │ def __getitem__(cls, key): │
│ │
│ /home/tl_admin/.local/lib/python3.8/site-packages/xturing/models/bloom.py:28 in __init__ │
│ │
│ 25 │ config_name: str = "bloom_lora" │
│ 26 │ │
│ 27 │ def __init__(self, weights_path: Optional[str] = None): │
│ ❱ 28 │ │ super().__init__(BloomLoraEngine.config_name, weights_path) │
│ 29 │
│ 30 │
│ 31 class BloomInt8(CausalInt8Model): │
│ │
│ /home/tl_admin/.local/lib/python3.8/site-packages/xturing/models/causal.py:222 in __init__ │
│ │
│ 219 │ │ model_name: Optional[str] = None, │
│ 220 │ │ target_modules: Optional[List[str]] = None, │
│ 221 │ ): │
│ ❱ 222 │ │ super().__init__( │
│ 223 │ │ │ engine, │
│ 224 │ │ │ weights_path=weights_path, │
│ 225 │ │ │ model_name=model_name, │
│ │
│ /home/tl_admin/.local/lib/python3.8/site-packages/xturing/models/causal.py:34 in __init__ │
│ │
│ 31 │ │ model_name: Optional[str] = None, │
│ 32 │ │ target_modules: Optional[List[str]] = None, │
│ 33 │ ): │
│ ❱ 34 │ │ self.engine = BaseEngine.create( │
│ 35 │ │ │ engine, │
│ 36 │ │ │ weights_path=weights_path, │
│ 37 │ │ │ model_name=model_name, │
│ │
│ /home/tl_admin/.local/lib/python3.8/site-packages/xturing/registry.py:14 in create │
│ │
│ 11 │ │
│ 12 │ @classmethod │
│ 13 │ def create(cls, class_key, *args, **kwargs): │
│ ❱ 14 │ │ return cls.registry[class_key](*args, **kwargs) │
│ 15 │ │
│ 16 │ @classmethod │
│ 17 │ def __getitem__(cls, key): │
╰──────────────────────────────────────────────────────────────────────────────────────────────────╯
TypeError: __init__() got an unexpected keyword argument 'model_name'
Hi @xupruvu,
Please use direct import for bloom loading
model = BloomLora()
Hi again, @xupruvu
Release is now in main branch, thanks for feedback!
| gharchive/issue | 2023-06-14T04:59:28 | 2025-04-01T06:45:53.301154 | {
"authors": [
"StochasticRomanAgeev",
"xupruvu"
],
"repo": "stochasticai/xTuring",
"url": "https://github.com/stochasticai/xTuring/issues/213",
"license": "Apache-2.0",
"license_type": "permissive",
"license_source": "github-api"
} |
191205873 | Added Blob attachment support for when syncing from remote server
This PR is to address problems with BLOB attachments during the PouchDB syncing process with a remote server like CouchDB.
The current process is to just global.btoa(attachment.data) which will always return the same string since encoding a Blob and not the content within. This means attachments just don't work if they come as type BLOB.
The solution is to run the Blob through the FileReader and then unwrap the ArrayBuffer returned back through btoa. This when provides a base64 string to be stored locally.
NOTE: This does not provide support to store BLOB attachments locally but instead support receiving remote BLOB attachments from a remote server. All BLOB attachments will be stored as base64 locally in AsyncStorage.
Before I fix the above tests, I just want to ask what the expectation and requirements of storing attachments in AsyncStorage are?
Is it required that attachments are stored as binary? I noticed the existing code takes a base64 then converts it using btoa before storage?
Sorry, I don't use attachments so no idea
For the moment, I'm going to close this PR as I've realised I might have jumped the gun a little early as although this solution works fine when React Native is in DEBUG mode, in RELEASE mode it causes an issue which I'm still tracking.
I'll re-raise once I get a solution and think a little more about the attachment question. I might ask the PouchDB guys for some suggestions
@ddolheguy No need to close the PR since it hurts discoverability. Imagine someone wants to start using this module and has attachments as a requirement. It's better that they see the problems on a surface without inspecting all the closed issues / pull requests. Closed should mean "already implemented / never will be implemented"
Same applies to #44 Actually, if a pull request has a string like "fixes #issue_id", the issue would be closed automatically on the pull request merge. See more at https://github.com/blog/1506-closing-issues-via-pull-requests
No problem, I've re-opened this. Just to give everyone an update, I haven't found an easy way to convert the BLOB types to base64 without saving them first to disk on the device then natively converting them.
@ddolheguy what was the issue that you found in release mode?
@jaredly I wasn't able to easily find a solution to convert the BLOB types to base64 on the device. Weirdly, it works when running debug with the browser but I believe it's using the engine there instead which is why the issue.
i will close this, i'd like to see a list of attachment Problems, and solve them
| gharchive/pull-request | 2016-11-23T07:46:33 | 2025-04-01T06:45:53.307147 | {
"authors": [
"ddolheguy",
"ewnd9",
"jaredly",
"stockulus"
],
"repo": "stockulus/pouchdb-react-native",
"url": "https://github.com/stockulus/pouchdb-react-native/pull/45",
"license": "mit",
"license_type": "permissive",
"license_source": "bigquery"
} |
247460623 | Avoid floating-point types in Ieee32::new and Ieee64::new.
This eliminates the need for unsafe code in code that uses Cretonne, a few
instances of unsafe code in Cretonne itself, and eliminates the only instance
of floating point in Cretonne.
As proposed here, this is an API change, requiring Cretonne users to update. I've prepared a patch here for wasm2cretonne.
See https://doc.rust-lang.org/1.0.0/style/style/naming/README.html for constructor names: new, with_details, or from_some_type.
| gharchive/pull-request | 2017-08-02T17:12:27 | 2025-04-01T06:45:53.320930 | {
"authors": [
"stoklund",
"sunfishcode"
],
"repo": "stoklund/cretonne",
"url": "https://github.com/stoklund/cretonne/pull/130",
"license": "apache-2.0",
"license_type": "permissive",
"license_source": "bigquery"
} |
279858822 | Different Font Spacing Issue
Essentially the problem at hand is that I'd like to mid-sentence or even mid-word change fonts. However, when I put the style in for the new font it give it a brand new line and then any text after also does this. Is there a way around this?
As an example I included a shared font with a show of what I mean. I want the TESTNOSPACE to be on the same line as the different font.
Share Link : http://homebrewery.naturalcrit.com/share/BkWRDpsr-z
After messing around with it, it seems using rather than or solved it. So I'll close this and hope anyone else with the same issue finds it.
| gharchive/issue | 2017-12-06T18:29:18 | 2025-04-01T06:45:53.323535 | {
"authors": [
"Kuhal"
],
"repo": "stolksdorf/homebrewery",
"url": "https://github.com/stolksdorf/homebrewery/issues/569",
"license": "mit",
"license_type": "permissive",
"license_source": "bigquery"
} |
1318988776 | rollback nodeselector and toleratiions
Signed-off-by: Wei Liu liuweixa@redhat.com
/assign @zhiweiyin318
/approve
/lgtm
| gharchive/pull-request | 2022-07-27T04:01:05 | 2025-04-01T06:45:53.324978 | {
"authors": [
"skeeey",
"zhiweiyin318"
],
"repo": "stolostron/klusterlet-addon-controller",
"url": "https://github.com/stolostron/klusterlet-addon-controller/pull/148",
"license": "Apache-2.0",
"license_type": "permissive",
"license_source": "github-api"
} |
1283198074 | 页眉相关问题
Discussed in https://github.com/stone-zeng/fduthesis/discussions/262
Originally posted by Aphroditefang June 24, 2022
您好，关于页眉有两个问题：
1.
目录这边的页眉怎么去掉呢？
这里参考文献能不能移到右边去呢？
是onside模式，然后利用
\makeatletter
\ExplSyntaxOn
\fancyhead[L]{\small\nouppercase{\fdu@kai \l__fdu_info_title_tl }}
\fancyhead[R]{\small \nouppercase{ \fdu@kai \csname l__fdu_header_center_mark_tl\endcsname\leftmark}}
\ExplSyntaxOff
\makeatother
设置页眉为左侧是文章名，右侧是章节名。
正文部分都是对的，参考文献和摘要部分就变成了左侧是文章名字，中间是章节名。
不要开多个 issue/discussion 提问
提问要说清模板版本
目录、参考文献没有出现页眉，请提供 MWE
1.好的，不小心同时弄了issue和discussion。
2.模板版本是最新的，克隆到本地后，生成thesis文件加并上传到overleaf。目录和参考文献首页没有页眉，但是第二页开始就有页眉了的。
目录。在 \tableofcontents 后面加上 \pagestyle{fancy}，在导言区加上\__fdu_patch_cmd:Nnn \tableofcontents
{
\__fdu_chapter_no_toc:V \contentsname
}
{
\__fdu_chapter_no_toc:V \contentsname \pagestyle{plain}
}
让所有章节在右边显示\cs_set_protected:Npn \__fdu_chapter_header:n #1
{
\bool_if:NTF \g__fdu_twoside_bool
{ \markboth {#1} {#1} }
{ \markboth { \hfill #1 } { } }
}
\makeatletter
\ExplSyntaxOn
\cs_set_protected:Npn \__fdu_chapter_header:n #1
{ \markboth {#1} { } }
\fancyhead[L]{\small\nouppercase{\fdu@kai\l__fdu_info_title_tl}}
\fancyhead[R]{\small\nouppercase{\fdu@kai\csname l__fdu_header_center_mark_tl\endcsname\leftmark}}
\ExplSyntaxOff
\makeatother
除每章首页外其余都用「左论文标题、右章名字」的格式作为页眉。
好的，所有问题都解决了。非常感谢个为耐心的解答，后续应该没什么疑问了
| gharchive/issue | 2022-06-24T03:00:00 | 2025-04-01T06:45:53.336330 | {
"authors": [
"Aphroditefang",
"sikouhjw",
"stone-zeng"
],
"repo": "stone-zeng/fduthesis",
"url": "https://github.com/stone-zeng/fduthesis/issues/263",
"license": "LPPL-1.3c",
"license_type": "permissive",
"license_source": "github-api"
} |
1027172382 | Use the right namespace for components
Previously operator used cluster object's namespace instead of storageos.
This change breaks desired behavior, so wrong at all.
| gharchive/pull-request | 2021-10-15T07:46:57 | 2025-04-01T06:45:53.338817 | {
"authors": [
"mhmxs"
],
"repo": "storageos/operator",
"url": "https://github.com/storageos/operator/pull/35",
"license": "MIT",
"license_type": "permissive",
"license_source": "github-api"
} |
1409227254 | refactor: sb sidebar template
Pull request type
Jira Link: INT-
[ ] Bugfix
[ ] Feature
[ ] Code style update (formatting, renaming)
[x] Refactoring (no functional changes, no api changes)
[ ] Other (please describe):
How to test this PR
Compare the storybook preview:
https://blok.ink/?path=/story/design-system-components-sbsidebar--default
https://storyblok-design-system-git-refactor-sb-si-4d777b-storyblok-com.vercel.app/?path=/story/design-system-components-sbsidebar--default
Test the app in the storyfront:
`cd storyblok-design-system && git checkout refactor/sb-sidebar-template``
yarn serve
The sidebar should work everywhere the same as before
What is the new behavior?
Other information
I tested in the app and it was working fine :)
| gharchive/pull-request | 2022-10-14T11:46:17 | 2025-04-01T06:45:53.364330 | {
"authors": [
"lisilinhart"
],
"repo": "storyblok/storyblok-design-system",
"url": "https://github.com/storyblok/storyblok-design-system/pull/322",
"license": "MIT",
"license_type": "permissive",
"license_source": "github-api"
} |
2722298424 | Support React 19
Description
Support React19
Suggested solution or improvement
Support React19
Additional context
No response
Validations
[X] Follow our Code of Conduct
Hi @The-End-Hero, can you please explain with more details what is the issue you are facing, a way to reproduce it, your suggested solution and some context? That would help us understanding what we could improve, thank you
React19 and Next15 have already been released, so expect to remove these warnings
@The-End-Hero you can safely ignore these warnings, and we are working on a new release with the upgrade to React 19
| gharchive/issue | 2024-12-06T07:43:07 | 2025-04-01T06:45:53.367508 | {
"authors": [
"The-End-Hero",
"edodusi"
],
"repo": "storyblok/storyblok-react",
"url": "https://github.com/storyblok/storyblok-react/issues/1276",
"license": "MIT",
"license_type": "permissive",
"license_source": "github-api"
} |
1355643069 | change host to localhost
📦 Published PR as canary version: 0.0.3-canary.4.cc1f164.0
:sparkles: Test out this PR locally via:
npm install @storybook/addon-bench@0.0.3-canary.4.cc1f164.0
# or
yarn add @storybook/addon-bench@0.0.3-canary.4.cc1f164.0
:rocket: PR was released in v0.0.3-next.0 :rocket:
| gharchive/pull-request | 2022-08-30T11:50:06 | 2025-04-01T06:45:53.369473 | {
"authors": [
"ndelangen",
"shilman"
],
"repo": "storybookjs/addon-bench",
"url": "https://github.com/storybookjs/addon-bench/pull/4",
"license": "MIT",
"license_type": "permissive",
"license_source": "github-api"
} |
306519855 | addon-options: Make shortcuts in storybook optional
Issue: #3196
What I did
I've added an option for enabling/disabling keyboard shortcuts.
How to test
One can try this feature by updating the storybook config.js as:
setOptions({
...
enableShortcuts: false, // true by default
})
This will disable the shortcuts.
Should we remove shortcuts help button and popup when shortcuts are disabled?
Yeah, that makes sense to disable it. Or we can add a text saying that shortcuts have been disabled.
Hey @Hypnosphi, Removed the shortcut help button when enableShortcuts: false.
| gharchive/pull-request | 2018-03-19T15:48:29 | 2025-04-01T06:45:53.418403 | {
"authors": [
"Hypnosphi",
"PunitGr"
],
"repo": "storybooks/storybook",
"url": "https://github.com/storybooks/storybook/pull/3237",
"license": "MIT",
"license_type": "permissive",
"license_source": "github-api"
} |
651962056 | UnhandledPromiseRejectionWarning when trying to login.
client.login(id, pass, uuid);
error message:
(node:11808) UnhandledPromiseRejectionWarning: #<Object>
(node:11808) UnhandledPromiseRejectionWarning: Unhandled promise rejection. This error originated either by throwing inside of an async function without a catch block, or by rejecting a promise which was not handled with .catch(). To terminate the node process on unhandled promise rejection, use the CLI flag `--unhandled-rejections=strict` (see https://nodejs.org/api/cli.html#cli_unhandled_rejections_mode). (rejection id: 2)
(node:11808) [DEP0018] DeprecationWarning: Unhandled promise rejections are deprecated. In the future, promise rejections that are not handled will terminate the Node.js process with a non-zero exit code.
When I add catch() to client.login,
const login = client.login(id, pass, uuid).catch(e=>{
console.log(e);
});;
I get the following error:
{
status: -100,
message: undefined,
userId: undefined,
countryISO: undefined,
countryCode: undefined,
accountId: undefined,
logonServerTime: undefined,
resetUserData: undefined,
accessToken: undefined,
refreshToken: undefined,
tokenType: undefined,
autoLoginEmail: undefined,
displayAccountId: undefined,
mainDevice: undefined,
mainDeviceAppVersion: undefined
}
What should I do?
Register device first. closing as invalid
duplicate issue #12
| gharchive/issue | 2020-07-07T03:19:41 | 2025-04-01T06:45:53.420941 | {
"authors": [
"blue-lang",
"storycraft"
],
"repo": "storycraft/node-kakao",
"url": "https://github.com/storycraft/node-kakao/issues/47",
"license": "MIT",
"license_type": "permissive",
"license_source": "github-api"
} |
329469029 | One rewind data instance per block when close to chain tip
Flush coinview when consensus tip is close to the chain tip to ensure that we have 1 rewind data per block.
Previously we were creating 1 rewind data for batch of blocks so during a reorg at the tip we were able to call a rewind on consensus tip and it would rewind a lot of blocks (up to 5000 if reorg happened when we just synced).
This is PR 2 from the list of PRs for the big refasctoring
Ah I did not run tests yet (or sync chains).
Ah I did not run tests yet (or sync chains).
Kk.
Ps
For me tests are passing (except for those tests that fail for 2 weeks already), stratis main and test synced.
| gharchive/pull-request | 2018-06-05T13:57:40 | 2025-04-01T06:45:53.484400 | {
"authors": [
"dangershony",
"noescape00"
],
"repo": "stratisproject/StratisBitcoinFullNode",
"url": "https://github.com/stratisproject/StratisBitcoinFullNode/pull/1475",
"license": "mit",
"license_type": "permissive",
"license_source": "bigquery"
} |
352606294 | Update the UI to be compatible with recent wallet API changes
The full node have modification to the wallet API which requires full node UI to be modified.
Consider taking the PR submitted with this issue into the repo, at the point in time when the git submodule of fullnode daemon is updated.
Task is to add the new passphrase parameter to API call.
Feature suggestion: https://github.com/stratisproject/StratisBitcoinFullNode/issues/1677
I'll take this one @bokobza
Done in #140 and #142
| gharchive/issue | 2018-08-21T15:54:49 | 2025-04-01T06:45:53.486355 | {
"authors": [
"dev0tion",
"sondreb"
],
"repo": "stratisproject/StratisCore",
"url": "https://github.com/stratisproject/StratisCore/issues/79",
"license": "MIT",
"license_type": "permissive",
"license_source": "github-api"
} |
1161633903 | Crash logs
[DEBUG]2022/03/08 00:18:50 client.go:51: on connect
[DEBUG]2022/03/08 00:18:50 conn.go:131: read start
panic: runtime error: invalid memory address or nil pointer dereference
[signal SIGSEGV: segmentation violation code=0x1 addr=0x78 pc=0x47541e2]
goroutine 3865 [running]:
github.com/stratosnet/sds/framework/client/cf.(*ClientConn).GetName(0x4c329d0)
/Users/espoir/Desktop/work/sds/framework/client/cf/conn.go:177 +0x22
github.com/stratosnet/sds/pp/peers.checkSingleSpLatency({0xc0001c8270, 0x12}, 0x80)
/Users/espoir/Desktop/work/sds/pp/peers/sp_peers.go:152 +0x2e5
github.com/stratosnet/sds/pp/peers.SendLatencyCheckMessageToSPList()
/Users/espoir/Desktop/work/sds/pp/peers/sp_peers.go:129 +0x127
created by github.com/stratosnet/sds/pp/event.ReqHBLatencyCheckSpList
/Users/espoir/Desktop/work/sds/pp/event/heart.go:44 +0xe5
Fixed by https://github.com/stratosnet/sds/pull/134
| gharchive/issue | 2022-03-07T16:20:12 | 2025-04-01T06:45:53.490514 | {
"authors": [
"espoir1989",
"mdumaisqsnetwork"
],
"repo": "stratosnet/sds",
"url": "https://github.com/stratosnet/sds/issues/128",
"license": "Apache-2.0",
"license_type": "permissive",
"license_source": "github-api"
} |
239682412 | How to use with pre-compiled template?
Hi there!
In the nunjucks documentations, it's written that you should pre-compile your templates before distribution/production.
I just did to test it, but this package doesn't render the js properly.
How should we setup this?
Thanks
Hi Extarys. I believe the precompiled option only applies if you are using Nunjucks on the browser. When using Nunjucks on Node.js, it compiles it once, then will serve the cached version for all subsequent requests (unless you set noCache to true). So as long as you're only using your templates server-side, you shouldn't need to worry about precompiling. Hope that helps.
Ohhh I see thank you very much!
| gharchive/issue | 2017-06-30T05:41:34 | 2025-04-01T06:45:53.496631 | {
"authors": [
"Extarys",
"strawbrary"
],
"repo": "strawbrary/koa-nunjucks-2",
"url": "https://github.com/strawbrary/koa-nunjucks-2/issues/27",
"license": "mit",
"license_type": "permissive",
"license_source": "bigquery"
} |
1914110189 | removed heartbeats - doesn't seem like we need them
This incorporates @blinktag's fixes to sdk-go-client as well
Forgot to update heartbeat-related tests. Doing that now.
| gharchive/pull-request | 2023-09-26T18:52:00 | 2025-04-01T06:45:53.507597 | {
"authors": [
"dselans"
],
"repo": "streamdal/snitch-server",
"url": "https://github.com/streamdal/snitch-server/pull/50",
"license": "Apache-2.0",
"license_type": "permissive",
"license_source": "github-api"
} |
651108167 | [Issue 126] Remove all the old dead code.
Fixes #126
Motivation
The project started off with hand coding the ampq protocol code. But
switched over to using qpid later. The old code wasn't deleted and left
around. This commit removes all of that code.
Modifications
This PR removes the entire io.streamnative.pulsar.handlers.ampq.frame package. That's where all the old code was.
Verifying this change
[ ] Make sure that the change passes the CI checks.
This change is a trivial rework / code cleanup without any test coverage.
Does this pull request potentially affect one of the following parts:
Dependencies (does it add or upgrade a dependency): no
The public API: yes
The schema: no
The default values of configurations: no
The wire protocol: no
The rest endpoints: no
The admin cli options: no
Anything that affects deployment: no
Documentation
Does this pull request introduce a new feature? no
If yes, how is the feature documented? not applicable
If a feature is not applicable for documentation, explain why?
If a feature is not documented yet in this PR, please create a followup issue for adding the documentation
Note: the frame package exists in the text code, but it isn't testing anything that was deleted. They tested parts of the io.streamnative.pulsar.handlers.amqp package.
I moved them up one package level.
@gaoran10
| gharchive/pull-request | 2020-07-05T17:12:55 | 2025-04-01T06:45:53.536444 | {
"authors": [
"davidlanouette"
],
"repo": "streamnative/aop",
"url": "https://github.com/streamnative/aop/pull/127",
"license": "Apache-2.0",
"license_type": "permissive",
"license_source": "github-api"
} |
1828925215 | X,y of click event in Image component
Hi
Thanks for the fantastic work. I really like this tool, but I have one suggestion about features, the ability to read the xy coordinates of a click event on the image component would be useful. Maybe should be passed to function by context parameter.
Hi, thanks for the kind words. I'm glad you're enjoying Streamsync. Events will soon get a big upgrade; this will be supported in the next release.
Hi
Thanks for letting me know. I will be waiting.
Happy to hear that, thanks for your patience!
| gharchive/issue | 2023-07-31T10:51:37 | 2025-04-01T06:45:53.549529 | {
"authors": [
"lincoln2491",
"ramedina86"
],
"repo": "streamsync-cloud/streamsync",
"url": "https://github.com/streamsync-cloud/streamsync/issues/66",
"license": "Apache-2.0",
"license_type": "permissive",
"license_source": "github-api"
} |
2279648496 | Add Bengali and see additional requested languages/dialects
A provider representative is requesting these most commonly used languages at provider locations other than Spanish and some french dialects
Pulaar (Guinea)
Pulaar (Mauritania)
Hassānīya (Mauritania)
Pulaar (Mali)
Wolof
Arabic (Mauritania)
Arabic (Sudan)
Arabic
@shakilhossain1 could you check and see if Google Translate has these dialects?
We can make a decision as to whether to implement some of these based on our ability to copyedit them.
Adding Russian to the list.
@shakilhossain1 just lifting up to get doobneek's russian fixes on test.
| gharchive/issue | 2023-10-24T18:27:09 | 2025-04-01T06:45:53.555812 | {
"authors": [
"adambard1"
],
"repo": "streetlives/yourpeer.nyc",
"url": "https://github.com/streetlives/yourpeer.nyc/issues/97",
"license": "MIT",
"license_type": "permissive",
"license_source": "github-api"
} |
1957573176 | add Stride Diagnostic Analyzer blog post
This adds a blog post about the upcomming feature for Stride.Core.CompilerServices DiagnosticAnalyzers
A review about grammar and typos is needed as always :D
Let me know once you are ready, and I can publish it to staging for further review :)
im ready, i just failed again with VSCode to rollback the commit .. i should have used revert and not reset.. but it is what it is
Also, add yourself here to the authors section https://github.com/stride3d/stride-website/blob/master/_data/site.json
| gharchive/pull-request | 2023-10-23T16:39:40 | 2025-04-01T06:45:53.576660 | {
"authors": [
"IXLLEGACYIXL",
"VaclavElias"
],
"repo": "stride3d/stride-website",
"url": "https://github.com/stride3d/stride-website/pull/209",
"license": "MIT",
"license_type": "permissive",
"license_source": "github-api"
} |
94405613 | [v1.2] casting booleans always returns in True
https://github.com/striglia/pyramid_swagger/blob/swagger_12/pyramid_swagger/tween.py#L293
unfortunately casting string to boolean is not as simple as bool()
Shouldn't it just be bool(json.loads(value))? Are we not loading the json first?
json bodies would already be casted by json.loads(), it's an issue for query/path/form I think
Looks like most people are effectively focusing on Swagger 2.0 at this point, so closing this 1.2 issue.
| gharchive/issue | 2015-07-10T22:29:44 | 2025-04-01T06:45:53.578617 | {
"authors": [
"dnephin",
"striglia"
],
"repo": "striglia/pyramid_swagger",
"url": "https://github.com/striglia/pyramid_swagger/issues/126",
"license": "bsd-3-clause",
"license_type": "permissive",
"license_source": "bigquery"
} |
1527410436 | feat: Add KafkaUser event source and sasl credentials to secret
Contributes to #15
Signed-off-by: Katherine Stanley 11195226+katheris@users.noreply.github.com
Thanks for the reviews. @mabulgu the Strimzi projects generally all contain checkstyle files that enforce certain coding requirements: https://github.com/strimzi/kafka-access-operator/blob/main/.checkstyle/checkstyle.xml
| gharchive/pull-request | 2023-01-10T13:41:16 | 2025-04-01T06:45:53.580342 | {
"authors": [
"katheris"
],
"repo": "strimzi/kafka-access-operator",
"url": "https://github.com/strimzi/kafka-access-operator/pull/16",
"license": "Apache-2.0",
"license_type": "permissive",
"license_source": "github-api"
} |
2690590864 | Add proposal for adopting Kafka Connect health check endpoint
This proposal describes an approach for adopting the health check endpoint for Kafka Connect and Mirror Maker 2 components.
Could you please also clarify what kind of user-impact we expect? What is the chance that users will actuall see that the probe changed?
I don't see any impact on users, also considering that the / endpoint will still be available, but not used anymore when Kafka 3.9+ is used. Added a note.
The impact is not about the / endpoint not being available. But the proposal argues that the new endpoint behaves better/differently. In that case, one can expect other behavior from it - e.g. the perceived startup taking longer - and that might mean that users will need to change their probe settings compared to what they have today or face the issue that their Connect / MM2 deployments might start failing.
@scholzj I added a commit with my user impact tests, but I see GH is taking a lot of time to process this commit (see "Processing updates" under PR title). I guess we just have to wait.
@scholzj @PaulRMellor your feedback should be addressed.
This proposal has no 5 binding +1 votes. If there are no new comments on Monday EOB, we should close it as approved.
Approved with 6 binding and 1 non-binding +1 votes.
| gharchive/pull-request | 2024-11-25T12:36:38 | 2025-04-01T06:45:53.583698 | {
"authors": [
"fvaleri",
"scholzj"
],
"repo": "strimzi/proposals",
"url": "https://github.com/strimzi/proposals/pull/139",
"license": "Apache-2.0",
"license_type": "permissive",
"license_source": "github-api"
} |
181287429 | Add new checks for slapd
This adds some initial metrics for slapd that come from the monitor backend. I still haven't quite got this working yet, since we need to have python-ldap installed so DataDog's python knows about it. I've been testing by just symlinking the system version into the right place (since we already have it installed on the system for nsscache).
Okay, I just tested this on a QA machine, and it appears to be outputting metrics correctly.
r? @gphat
Very cool!
Can you also update the README.md with an explanation of this check and it's metrics? Trying to keep this friendly toward anyone who finds it and well documented!
Code looks great though, just holding off approval until the README is there.
Okay, added some documentation in the README! Let me know if you need anything else :-)
👍
| gharchive/pull-request | 2016-10-05T23:04:29 | 2025-04-01T06:45:53.647848 | {
"authors": [
"andrew-d",
"gphat"
],
"repo": "stripe/datadog-checks",
"url": "https://github.com/stripe/datadog-checks/pull/32",
"license": "mit",
"license_type": "permissive",
"license_source": "bigquery"
} |
1441151948 | The configuration option for region Oregon (us-west1) is missing
Bug report
Extension name: [e.g. firestore-stripe-payments]
Describe the bug
The configuration option for region Oregon (us-west1) is missing.
To Reproduce
Steps to reproduce the behavior, please provide code snippets or a repository:
Attempt install from the Firebase Console
Scroll through the list of regions for the config
Note that us-west2, us-west3, us-west4 are available
Cry in frustration because all of your services happen to be in us-west1
Attempt to work around it:
Install for the us-central1 region
Use the Firebase CLI to export the extension config
From the Firebase Console, uninstall the extension
Modify the extensions/firestore-stripe-payments.env file in the local project setting LOCATION=us-west1
Run firebase deploy --only extensions for the project
See this error:
Extensions deploy had errors:
create firestore-stripe-payments
Request had 1 validation errors: params.LOCATION: param provided with invalid option us-west1, must be one of us-central1, us-east1, us-east4, us-west2, us-west3, us-west4, europe-central2, europe-west1, europe-west2, europe-west3, europe-west6, asia-east2, asia-northeast1, asia-northeast2, asia-northeast3, asia-south1, asia-southeast2, northamerica-northeast1, southamerica-east1, australia-southeast1.
Expected behavior
Oregon (us-west1) is an option for installation via the Firebase console and CLI
Screenshots
If applicable, add screenshots to help explain your problem.
System information
OS: [e.g. macOS, Windows]
Browser (if applies) [e.g. Chrome, Safari]
Additional context
Add any other context about the problem here.
We just need to add the option here. We'd love to accept a PR that adds support!
| gharchive/issue | 2022-11-09T00:50:45 | 2025-04-01T06:45:53.663679 | {
"authors": [
"KTOLE18",
"jsteele-stripe"
],
"repo": "stripe/stripe-firebase-extensions",
"url": "https://github.com/stripe/stripe-firebase-extensions/issues/461",
"license": "Apache-2.0",
"license_type": "permissive",
"license_source": "github-api"
} |
613929008 | Metadata.publication_date?
This seems like a bug - publication_date is part of the fields map.
i moved it knowingly to all the other date fields, like state or valid_from
but can be moved back, was just a proposal
publication_date had no impact on legal checking in the past (like last_modified_date)
| gharchive/issue | 2020-05-07T10:00:06 | 2025-04-01T06:45:53.677328 | {
"authors": [
"harryurban",
"moritzzimmer"
],
"repo": "stroeer/tapir",
"url": "https://github.com/stroeer/tapir/issues/8",
"license": "Apache-2.0",
"license_type": "permissive",
"license_source": "github-api"
} |
128275546 | npmjs.org site point to wrong github url
When I go to https://www.npmjs.com/package/express
and then follow the link for the Github repo.. it takes me to https://github.com/expressjs/express instead of this repo url of https://github.com/strongloop/express
It may be wrong, but the URL sounds about right to me.
| gharchive/issue | 2016-01-23T00:13:03 | 2025-04-01T06:45:53.679280 | {
"authors": [
"amit777",
"jordonias"
],
"repo": "strongloop/express",
"url": "https://github.com/strongloop/express/issues/2861",
"license": "mit",
"license_type": "permissive",
"license_source": "bigquery"
} |
318347155 | Cannot found fse-v1.2.2-node-v64-darwin-x64.tar.gz
Environment:
macOS 10.13.4
Node v10.0.0
NPM v5.6.0
Error Content Screenshot:
Error Content:
> fsevents@1.2.2 install /usr/local/lib/node_modules/fsevents
> node install
node-pre-gyp ERR! Tried to download(404): https://fsevents-binaries.s3-us-west-2.amazonaws.com/v1.2.2/fse-v1.2.2-node-v64-darwin-x64.tar.gz
node-pre-gyp ERR! Pre-built binaries not found for fsevents@1.2.2 and node@10.0.0 (node-v64 ABI, unknown) (falling back to source compile with node-gyp)
SOLINK_MODULE(target) Release/.node
CXX(target) Release/obj.target/fse/fsevents.o
SOLINK_MODULE(target) Release/fse.node
COPY /usr/local/lib/node_modules/fsevents/lib/binding/Release/node-v64-darwin-x64/fse.node
TOUCH Release/obj.target/action_after_build.stamp
+ fsevents@1.2.2
added 68 packages in 6.523s
Please help me!!
There's nothing wrong. There wasn't a precompiled binary so it built it from source.
| gharchive/issue | 2018-04-27T09:52:24 | 2025-04-01T06:45:53.682239 | {
"authors": [
"bnoordhuis",
"sunzhenyucn"
],
"repo": "strongloop/fsevents",
"url": "https://github.com/strongloop/fsevents/issues/216",
"license": "mit",
"license_type": "permissive",
"license_source": "bigquery"
} |
98720557 | Shitty example
This should probably be a private repository until it is better documented and has some kind of functionality. It is misleading and a timesink in its current state.
While we appreciate community feedback, the title of this comment is inappropriate and not appreciated. This is a demo repo, not a full application, and not the library it is intended to demonstrate. As such, it will remain as is.
Sorry for the wording. I did miss some more information in the readme about how this can be used, and it did not run in any way locally, and thus it is probably worth looking into how this can be more helpful
Thanks for the update. I agree this is not documented well, we should have a better example, but that's just not high on our priority list right now. If we're lucky, @crandmck can update the readme with some better information on how to use this example app, and what it does (and why). There are some additional docs on our web site, which we should probably link to from here.
I'll try to find time to improve these and the pub-sub docs in general, but @jakerella is correct that this is not a top priority. I did at least add links in the README to the other relevant docs.
| gharchive/issue | 2015-08-03T11:07:42 | 2025-04-01T06:45:53.699119 | {
"authors": [
"crandmck",
"jakerella",
"kennethlynne"
],
"repo": "strongloop/loopback-example-pubsub",
"url": "https://github.com/strongloop/loopback-example-pubsub/issues/5",
"license": "mit",
"license_type": "permissive",
"license_source": "bigquery"
} |
1140138284 | Atom Button cannot be set to type submit
The type attribute for a <button> cannot be set to type="submit", has it is default to type="button" when the html tag is <button>
type: tag == 'button' ? 'button' : false,
This is odd as the type property is set for the default attributes, which should be overridden when using the attr parameter 🤔
I will have a look at this.
i'm a n***, something was not set correctly on my end
| gharchive/issue | 2022-02-16T15:04:30 | 2025-04-01T06:45:53.753170 | {
"authors": [
"perruche",
"titouanmathis"
],
"repo": "studiometa/ui",
"url": "https://github.com/studiometa/ui/issues/24",
"license": "MIT",
"license_type": "permissive",
"license_source": "github-api"
} |
2341532404 | 🛑 Beefs and Beers is down
In 10ccc47, Beefs and Beers (https://beefsandbeers.com) was down:
HTTP code: 0
Response time: 0 ms
Resolved: Beefs and Beers is back up in a24008f after 6 minutes.
| gharchive/issue | 2024-06-08T07:37:39 | 2025-04-01T06:45:53.755589 | {
"authors": [
"studiovlijmscherp"
],
"repo": "studiovlijmscherp/uptime",
"url": "https://github.com/studiovlijmscherp/uptime/issues/1835",
"license": "MIT",
"license_type": "permissive",
"license_source": "github-api"
} |
63419147 | Add dependency on MultiJSON
Hi,
Thanks for the nice gem.
But I encountered the error below when executing rackup:
../vendor/bundle/ruby/2.2.0/gems/bower-0.0.3/lib/bower/environment.rb:1:in `require': cannot load such file -- multi_json (LoadError)
At minimal project rather than Rails, sometimes MultiJSON is not on dependency table.
Could you add it to bower.gemspec?
Thanks for the PR!
I just released 0.0.4 with this change.
Thanks quick action!
| gharchive/pull-request | 2015-03-21T15:37:02 | 2025-04-01T06:45:53.765392 | {
"authors": [
"KitaitiMakoto",
"stve"
],
"repo": "stve/bower",
"url": "https://github.com/stve/bower/pull/7",
"license": "mit",
"license_type": "permissive",
"license_source": "bigquery"
} |
231587660 | Unitless shorthands
So on SC we've seen some complaints about properties being able to be unitless, e.g. padding-top: 2, but not shorthands: padding: 2px.
This probably comes from the difference here: https://github.com/styled-components/css-to-react-native/blob/master/src/index.js#L8
And here: https://github.com/styled-components/css-to-react-native/blob/master/src/tokenTypes.js#L31
Since most directions don't expect unitless numbers: https://github.com/styled-components/css-to-react-native/blob/master/src/transforms/util.js#L6
The question is, since this seems like an inconsistency, do we want to change this? :)
cc @mxstbr @jacobp100
It's sort of difficult because if we do this, we essentially create a whitelist of properties. It just means if RN adds new styles anywhere, we have to keep up-to-date.
@jacobp100 Maybe there is really no elegant way to implement it. But this project is called "css-to-react-native", it is normal to include some logics that should always follow RN's steps.
cc @mxstbr.
I am not sure if this is the right thread because the one was closed on SC.
If you see SC as a styling engine that provides you the theme + style parser (no css parser) and you would have something like react-dom and react-native which specific parser will do their job (yes it does it right now to some level but once you merge web with mobile then the specially cases will start coming out, like this one)
Probably most of us are Web developers and it makes sense for us to add units but for people coding on RN (I started a couple of weeks ago but include me) having units just add complexity to what they are used already. You can't migration your objects because now you need to pay attention to add units, and the worst case if I am not mistaking is that no all the time applies the logic of adding units ...
The idea of copying css and hand free makes no sense at all because you can't port px values to the mobile because there is no px unit, so probably everywhere you add px you will have to modifier the values because the calculation is off completely so I dont see the benefits (em could work but still there is no relative to the last font size either)
I would push to remove complexity to RN people and remove units.
P.S: Perspective on removing complexity or not is more fundamentally based on personal opinions but I would take more feedbacks from RN engineers like @iRoachie . This is my first time working full-time on RN and I can tell you, this add complexity for me right now, now I need to memories when tu use px or not (no including that RN community will no get why I am putting px in my code and they will no support people using SC because fundamentally is different from the RN Style engine).
@philpl any update on what will be the move of styled-components and unitless situation?
Currently you can remove units unless you use CSS shorthands. There are no plans to remove units in shorthands, and this is unlikely to change.
You can look at the discussions when this changed for reasons why we switched to enforcing units.
@jacobp100 which discussions?
This is still being flagged in issues against s-c, would love to see some movement either way on this. A few potential solutions:
As @jacobp100 suggested, warn.
Allow all properties to be unitless and add a unit behind the scenes if needed.
warn & throw out CSS rules without units where there should be a unit (this actually probably wouldn't be too arduous since there are few CSS properties that legitimately don't have units)
I do mean to get round to adding warnings, it's just finding time. I had a very-POC PR for how I think it could work
Curious why does it not make sense to remove units from shorthands? I don't understand the reasoning. Whitelist is mentioned so shorthands can be made up so the problem is by default units are expected and you would have to whitelist shorthands that don't require the units and that is the problem?
Couldn't there be a list of standard shorthands like margin that don't need the unit?
@dgobaud You can read some of the reasoning over at #20. There's also linked issues in there, and probably linked issues in the linked issues. I don't really want to reopen the discussions on this, it seems pretty settled - units are here to stay! But I'm happy to answer any questions on bugs etc. 🙂
@jacobp100 interesting so it seems the CSS spec says pixels are not physical device pixels - meaning unitless like the RN units? I guess this means the correct thing to do is actually use px everywhere?
Not too sure what you mean, but yeah, pixels on the web behave the same way as RN
Hi. I ran into this the other day and wasted some brain cycles trying to figure out what was going on. 😃
Are there any other units that can be used besides px? If not, maybe the warning should be more explicit?
"px" is now required for all css measurements.
and maybe point to the fact that px is density independent? https://www.w3.org/TR/css3-values/#px
It's a little bit confusing of a change.
I have the problem with that. I created custom Text components, which overrides Text, considering lineHeight as relative value, not absolute. So line-height: 1.5 instead of line-height: 18px. In that case I don't have units. And I get this warning.
@likern My understanding was that line height is always a length value in React Native, never a multiplier. Correct me if I'm wrong. But if that is the case, I'm glad these warnings helped highlight that! :smile:
length value
Yes, it was. That's why I've decided to create wrapper which uses it as multiplier. As far as I know it is a recomended approach in CSS. And anyway designers use multipliers in tools like Figma.
I couldn't imagine omitting px postfix could lead to some errors in debug mode. Now it's a blocker for me.
The aim of this library is to transform bit of CSS that works on both web and React Native. Because this isn't natively supported in RN, I wouldn't feel comfortable loosening the rules here
But if you give me details on your wrapper, I'd be happy to look at how you could get it to work. We have a few libraries that do this kind of pre-processing before handing off here, so it seems very do-able
My component is very simple. It just wraps RN Text component. It is expected to be working after css-to-react-native strips units off.
So I get usual lineHeight as number, but just treat it as multiplier (instead of absolute value in pixels) - and pass to RN Text lineHeight = fontSize * lineHeight.
I shared my component here: https://gist.github.com/likern/6e2da0bbe21fd0a7f7b4255d4537eb6e
Okay, makes sense. You’re actually the first person I’ve seen to do this via a wrapper component - most people just wrap the styled function in some way
Since your component requires the font size to be present anyway, you could do this via a mixin in styled components
We have previously been too loose in what we accept, and this has caused issues for parsing. We are fixing these issues and making the parser stricter, and this does have the consequence that people using those behaviours will no longer be able to do so. I think the warnings are important so we can guide people away from writing stuff that can cause bugs
@jacobp100 I don't have access to styled because I'm using css prop transformation, so my code looks like this:
<CustomText
css={`
font-family: Gilroy-SemiBold;
font-size: 28px;
line-height: 20;
`}
>
This is a custom text component
</CustomText>
The next where I can grab this information is inside custom component.
What about adding styled.config.js file which will allow to add custom code transformations at build-time, while keeping core css-to-react-native untouched?
styled.config.js:
module.exports = {
"react-native": {
extends: {
file: "js file with compile-time transformations using babel-plugin-preval "
}
}
};
Developers can write their own transformations using babel-plugin-preval.
In my example:
<CustomText
css={`
font-family: Gilroy-SemiBold;
font-size: 28px;
line-height: 1.5;
`}
>
This is a custom text component
</CustomText>
will be transformed by my own transformation (it is a user responsibility) (which will be called by css-to-react-native at pre-execution phase) to
<CustomText
css={`
font-family: Gilroy-SemiBold;
font-size: 28px;
line-height: 42px;
`}
>
This is a custom text component
</CustomText>
Then the core strict logic of css-to-react-native will be applied.
I meant like this,
const fontSize = (fontSize, lineHeightMultiplier) => `
font-size: ${fontSize}px;
line-height: ${fontSize * lineHeightMultiplier}px;
`;
<NormalText
css={`
font-family: Gilroy-SemiBold;
${fontSize(25, 1.5)}
`}
>
This is a normal text component
</NormalText>
Your babel plugin idea will probably work too!
Btw, we used to actually transform unitless line heights for the case,
font: 16px/1.5 "Some font"
And that used to give a line height of 24px. But we removed it because it does not work for the case,
const BaseText = styled`
font: 16px/1.5 "Some font";
`
const BiggerText = styled`
font-size: 32px;
`
<BaseText>
My line height is correct
<BiggerText>
But mine is not
</BiggerText>
</BaseText>
This lacks consistency
look at the exemple above
export const Input = styled(FormikInput).attrs({
containerStyle: {
font-Size: props => props.theme.typography.size_large
},
errorStyle: {
},
})`
on theme
size_large: 18
gives this error: expected styles to contain units
if I make this
size_large: '18px'
react native crashes string cannot be cast to java.lang.double
Any ideas?
I think your error is coming from some other code not shown in the example. You probably somewhere have,
styled.Text`
font-size: ${props => props.theme.typography.size_large};
`
Generally, should store size_large as a number, then change the above to
styled.Text`
font-size: ${props => props.theme.typography.size_large}px;
`
(Notice the px)
Hope this helps!
@jacobp100 , can the same be done with object-like styles? I.e
can we have a stylelint for this?
https://stylelint.io/
what is the codemod that adds units?
codemod here https://github.com/styled-components/styled-components-native-code-mod
fix to work with latest styled-components version https://github.com/styled-components/styled-components-native-code-mod/pull/21
I'm using both styled-components and styled-system with typescript. I have found that if you make use of the default parms, you can use without units and there will not be any warnings pop up.
You can either include everything with the defaults or mix as you like. For the styles that are entered with the defaults, the format is as per RN but for the styles within the styled-component proper, it is the normal hyphenated naming.
A sample of my working code...
interface ICustomProps extends BorderProps, FlexboxProps, LayoutProps, SpaceProps {};
export const LineItem = styled.View``;
LineItem.defaultProps = {
width: "98%",
flexDirection: "row",
borderColor: "black",
marginTop: 5,
marginBottom: 5,
marginLeft: 16,
borderBottomWidth: 1,
paddingTop: 8,
paddingBottom: 5,
paddingLeft: 10,
}
export const HeaderItem = styled(LineItem) background-color: #f5f5f5;;
HeaderItem.defaultProps = {
//backgroundColor: "#f5f5f5",
borderTopWidth: 1
}
What is the status on this? Other issues are being closed and linking to this one, but there is no indication on whether or not this idea is dead or still being considered?
It's not being considered. I wanted to keep this open for visibility, but it's probably had its time now.
Hope this clarifies things
| gharchive/issue | 2017-05-26T10:30:01 | 2025-04-01T06:45:53.798346 | {
"authors": [
"GollyJer",
"Rodrigo816",
"chinesedfan",
"dgobaud",
"jacobp100",
"likern",
"mstarcevic",
"olehs",
"philpl",
"probablyup",
"sibelius",
"wtfluckey",
"yordis"
],
"repo": "styled-components/css-to-react-native",
"url": "https://github.com/styled-components/css-to-react-native/issues/40",
"license": "mit",
"license_type": "permissive",
"license_source": "bigquery"
} |
499414460 | IE11 not picking up class when using SSR
Environment
System:
OS: macOS Mojave 10.14.6
CPU: (12) x64 Intel(R) Core(TM) i7-9750H CPU @ 2.60GHz
Memory: 9.84 GB / 32.00 GB
Shell: 5.7.1 - /usr/local/bin/zsh
Binaries:
Node: 12.9.0 - ~/.nvm/versions/node/v12.9.0/bin/node
Yarn: 1.17.3 - /usr/local/bin/yarn
npm: 6.10.2 - ~/.nvm/versions/node/v12.9.0/bin/npm
Watchman: 4.9.0 - /usr/local/bin/watchman
npmPackages:
styled-components: ^5.0.0-beta.8 => 5.0.0-beta.8-groupsizefix
Steps to reproduce
Sadly I'm not able to properly reproduce this issue outside of our application.
Expected Behavior
Rendered styles should be rendered by IE11.
Actual Behavior
Using SSR, the class/rule definition is correctly rendered to the client and appended into the DOM. But IE11 doesn't seem to be able to pick up the styles in certain cases, I got an element with class="r5r22j-0 lkVjVW" but the style panel shows no styles being applied to the element.
The corresponding generated style rules are:
.lkVjVW{-webkit-flex-shrink:0;-ms-flex-negative:0;flex-shrink:0;height:100%;position:relative;text-align:center;-webkit-user-select:none;-moz-user-select:none;-ms-user-select:none;user-select:none;cursor:pointer;-webkit-transition-property:-webkit-transform;-webkit-transition-property:transform;transition-property:transform;width:562.7484662576687px;}
I don't see anything going wrong but the result is nonetheless an element without styles 🤔. It works perfectly fine in other browsers and once I force a rerender by toggling the element it seems to be working (I assume styled-components generate a new style rule and injects new styles into the DOM).
As a workaround I got it fixed by using a forced rerender and setting a class on the styled component:
class MyComponent extends React.Component {
state = { forceRerender: false };
componentDidMount() {
this.setState({ forceRerender: true });
}
render() {
return <MyStyledComponent className={`forced-rerender-{this.state.forceRerender}`} />;
}
}
Not very happy with this, but at least a way to move forward.
@topaxi can you try beta 11? We had a bug that likely caused this.
The mentioned issue/code was not in beta-8 yet. But I can definitely try it out at some point.
| gharchive/issue | 2019-09-27T12:23:02 | 2025-04-01T06:45:53.806647 | {
"authors": [
"probablyup",
"topaxi"
],
"repo": "styled-components/styled-components",
"url": "https://github.com/styled-components/styled-components/issues/2770",
"license": "mit",
"license_type": "permissive",
"license_source": "bigquery"
} |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.