pull_number
int64
64
2.85k
test_patch
stringlengths
303
98.3k
issue_numbers
listlengths
1
2
instance_id
stringlengths
18
20
problem_statement
stringlengths
24
6.49k
version
stringclasses
19 values
base_commit
stringlengths
40
40
patch
stringlengths
372
188k
repo
stringclasses
1 value
created_at
stringlengths
20
20
hints_text
stringlengths
0
6.32k
environment_setup_commit
stringclasses
20 values
233
diff --git a/serde_tests/tests/test_annotations.rs b/serde_tests/tests/test_annotations.rs index dc91b9b5d..82f1265e2 100644 --- a/serde_tests/tests/test_annotations.rs +++ b/serde_tests/tests/test_annotations.rs @@ -23,24 +23,44 @@ struct DisallowUnknown { } #[derive(Debug, PartialEq, Serialize, Deserialize)] -str...
[ "211", "218" ]
serde-rs__serde-233
Consider removing format specific renames cc @hugoduncan. I did a quick search through github, and I couldn't find anyone using `#[serde(rename(json=..., xml=...))]` and etc. I'm not sure if it's pulling it's weight and I'm considering removing it in the upcoming 0.7.0 release. This is also adding some complication in...
0.6
a926cb62f05012d15618922029e1cd9cdf42cd10
diff --git a/serde/src/de/mod.rs b/serde/src/de/mod.rs index cb0e25e8c..3d172a06c 100644 --- a/serde/src/de/mod.rs +++ b/serde/src/de/mod.rs @@ -424,15 +424,6 @@ pub trait Deserializer { { self.deserialize(visitor) } - - /// Specify a format string for the deserializer. - /// - /// The deser...
serde-rs/serde
2016-02-08T06:07:45Z
I'm using it in my AWS code, as the names used in the XML replies from AWS don't match the names used in their API requests. Having said that I haven't found time to work on anything in rust in the last few months. @hugoduncan: Do you have your code online? I did find another rust AWS API client, [rusoto](https://git...
d24b2c86f2c9b3f5036289d6c2c5764be57d2127
224
diff --git a/serde_tests/tests/test_bytes.rs b/serde_tests/tests/test_bytes.rs index 137d00986..785ff5503 100644 --- a/serde_tests/tests/test_bytes.rs +++ b/serde_tests/tests/test_bytes.rs @@ -9,6 +9,12 @@ use serde::bytes::{ByteBuf, Bytes}; #[derive(Debug, PartialEq)] struct Error; +impl serde::ser::Error for Erro...
[ "57" ]
serde-rs__serde-224
`Path` serialization can panic The `Serialize` implementation of `Path` incorrectly assumes that all paths can be represented as Unicode strings.
0.6
9ae47a261e2e2f7dacc44c91e315fc90675e4e4b
diff --git a/serde/src/ser/impls.rs b/serde/src/ser/impls.rs index 7d01a384d..cc20f9b60 100644 --- a/serde/src/ser/impls.rs +++ b/serde/src/ser/impls.rs @@ -27,6 +27,7 @@ use std::sync::Arc; use core::nonzero::{NonZero, Zeroable}; use super::{ + Error, Serialize, Serializer, SeqVisitor, @@ -682,7 ...
serde-rs/serde
2016-01-28T18:41:45Z
Good point! Should be simple enough to have it return a syntax error if it gets a non-unicode string. err, rather tries to serialize to a non-unicode string. So it could either just be serialized as a byte string, or we first try to encode it as a string, then fail back to a byte string if it's non-unicode. What shou...
d24b2c86f2c9b3f5036289d6c2c5764be57d2127
139
diff --git a/serde_tests/benches/bench_log.rs b/serde_tests/benches/bench_log.rs index 4fa108da5..1dddcf500 100644 --- a/serde_tests/benches/bench_log.rs +++ b/serde_tests/benches/bench_log.rs @@ -1,6 +1,6 @@ use std::io::{self, Read, Write}; use num::FromPrimitive; -use test::Bencher; +use test::{Bencher, black_box}...
[ "122" ]
serde-rs__serde-139
Serializing a map with non-string keys will silently generate invalid JSON For example, a `HashMap<u32, u32>` will produce invalid JSON when serialized. At a minimum, the json serializer should error rather than produce invalid JSON. Ideally however, it would fall back to serializing the map as a list of pairs: ``` ...
0.5
7b773ac08880adece2f1bd809efcbd5828b23293
diff --git a/serde_json/src/error.rs b/serde_json/src/error.rs index 6621188bb..6c3a494c9 100644 --- a/serde_json/src/error.rs +++ b/serde_json/src/error.rs @@ -2,6 +2,7 @@ use std::error; use std::fmt; use std::io; use std::result; +use std::string::FromUtf8Error; use serde::de; @@ -44,7 +45,6 @@ impl fmt::Deb...
serde-rs/serde
2015-08-14T05:34:23Z
we had this issue before with rustc-serialize... I wasn't aware we also had it in serde... probably we just need an intermediate serializer here: https://github.com/serde-rs/serde/blob/master/serde_json/src/ser.rs#L262 that is passed to the MapVisitor, to check for stringable keys and error else. Fallback to serializi...
7b773ac08880adece2f1bd809efcbd5828b23293
684
diff --git a/serde_test/src/de.rs b/serde_test/src/de.rs index 2ad0c9be1..559e9972d 100644 --- a/serde_test/src/de.rs +++ b/serde_test/src/de.rs @@ -31,76 +31,105 @@ impl<I> Deserializer<I> self.tokens.next() } + pub fn expect_token(&mut self, expected: Token) -> Result<(), Error> { + match se...
[ "656" ]
serde-rs__serde-684
Simplify contract around SeqVisitor::end() and MapVisitor::end() These are really easy to screw up when implementing visit_seq and visit_map. If you forget to call them, things may work in some formats and fail in really confusing ways in other formats. Hopefully we can find a way for the type system to enforce corr...
0.8
88a4ed9cd72c7901b025dd8c8d86c69cfc6b157f
diff --git a/.travis.yml b/.travis.yml index 87e33f84b..9ef108a71 100644 --- a/.travis.yml +++ b/.travis.yml @@ -19,8 +19,8 @@ script: - (cd testing && travis-cargo --skip nightly test) - (cd testing && travis-cargo --only nightly test -- --features unstable-testing) - (cd serde_derive && travis-cargo --only nightly...
serde-rs/serde
2017-01-13T18:24:06Z
I think we can just get rid of end() and handle their job inside whatever calls visit_seq or visit_map.
88a4ed9cd72c7901b025dd8c8d86c69cfc6b157f
682
diff --git a/testing/tests/test_de.rs b/testing/tests/test_de.rs index d1879fe84..3359492e8 100644 --- a/testing/tests/test_de.rs +++ b/testing/tests/test_de.rs @@ -32,12 +32,40 @@ struct Struct { c: i32, } +#[derive(PartialEq, Debug, Deserialize)] +#[serde(deny_unknown_fields)] +struct StructDenyUnknown { + ...
[ "444" ]
serde-rs__serde-682
Fields marked skip_deserializing are ignored even if deny_unknown_fields is set I have [this struct definition](https://github.com/sfackler/log4rs/blob/6eea8767b5ed5c014d06685d78a6df8c54147a92/src/file/serde.rs.in#L1-L14). It is configured to derive `Deserialize` and deny unknown fields. I have a field marked to be sk...
0.8
c008c6d3a8e5315667c0380df1d91ff1daa5d71d
diff --git a/serde_codegen/src/de.rs b/serde_codegen/src/de.rs index b2d64a9bf..ed4f9e0aa 100644 --- a/serde_codegen/src/de.rs +++ b/serde_codegen/src/de.rs @@ -480,11 +480,14 @@ fn deserialize_item_enum( let type_name = item_attrs.name().deserialize_name(); + let variant_names_idents = variants.iter() + ...
serde-rs/serde
2017-01-13T07:29:07Z
@oli-obk @erickt opinion on this? It seems like there are two different use cases for `skip_deserializing`. 1. A field of the struct is not present in the serialized form. 2. A field of the struct is present in the serialized form but should not be deserialized. If we make this change, there would be no way to handle...
88a4ed9cd72c7901b025dd8c8d86c69cfc6b157f
681
diff --git a/serde_test/src/de.rs b/serde_test/src/de.rs index 39e14faaa..2ad0c9be1 100644 --- a/serde_test/src/de.rs +++ b/serde_test/src/de.rs @@ -85,12 +85,12 @@ impl<I> Deserializer<I> }) } - fn visit_struct<V>(&mut self, len: Option<usize>, mut visitor: V) -> Result<V::Value, Error> + fn visi...
[ "559" ]
serde-rs__serde-681
`deserialize_struct` is called with real field names instead of deserializable names ``` rust #[derive(Deserialize)] pub struct S { #[serde(rename = "A")] a: u8, #[serde(skip_deserializing)] b: u8, } ``` I was surprised to see this in the generated code: ``` rust const FIELDS: &'static [&'static str] ...
0.8
110d36fa1469375ad00dcd0f460451e6e390f1ed
diff --git a/serde_codegen/src/de.rs b/serde_codegen/src/de.rs index 8eb699701..84cbaa7da 100644 --- a/serde_codegen/src/de.rs +++ b/serde_codegen/src/de.rs @@ -503,7 +503,7 @@ fn deserialize_item_enum( true, ); - let variant_names = variants.iter().map(|variant| variant.ident.to_string()); + let ...
serde-rs/serde
2017-01-12T19:24:34Z
@dtolnay: yeah I think this would make sense to change. There isn't much reason for the deserializer to get passed the rust-field-specific names. This must have just been something we missed when we added field renaming. I'm not sure if anyone uses them. Maybe toml? I remember it did some funky things with trying to re...
88a4ed9cd72c7901b025dd8c8d86c69cfc6b157f
678
diff --git a/testing/tests/test_gen.rs b/testing/tests/test_gen.rs index 660eb52e1..24338312e 100644 --- a/testing/tests/test_gen.rs +++ b/testing/tests/test_gen.rs @@ -208,6 +208,85 @@ fn test_gen() { struct NonAsciiIdents { σ: f64 } + + #[derive(Serialize, Deserialize)] + struct EmptyBraced {...
[ "676" ]
serde-rs__serde-678
Empty braced struct generates dead code ```rust #[derive(Deserialize)] #[serde(deny_unknown_fields)] pub struct Config { } ``` The generated code (cleaned up) looks like: ```rust impl Deserialize for Config { fn deserialize<D>(deserializer: &mut D) -> Result<Config, D::Error> where D: Deserial...
0.8
8242c64152d2d13ec4ebc4e197db709500cc9fce
diff --git a/serde_codegen/src/de.rs b/serde_codegen/src/de.rs index e75a7c961..9e0b6bee9 100644 --- a/serde_codegen/src/de.rs +++ b/serde_codegen/src/de.rs @@ -509,14 +509,6 @@ fn deserialize_item_enum( const VARIANTS: &'static [&'static str] = &[ #(#variant_names),* ]; }; - let ignored_arm = if ite...
serde-rs/serde
2017-01-11T19:03:35Z
88a4ed9cd72c7901b025dd8c8d86c69cfc6b157f
671
diff --git a/testing/tests/test_de.rs b/testing/tests/test_de.rs index 280d427f8..d1879fe84 100644 --- a/testing/tests/test_de.rs +++ b/testing/tests/test_de.rs @@ -3,7 +3,7 @@ use std::net; use std::path::PathBuf; use std::time::Duration; -use serde::Deserialize; +use serde::de::{Deserialize, Type}; extern crat...
[ "667" ]
serde-rs__serde-671
Stop generating visit_usize and visit_bytes for FieldVisitor As far as I know, not a single format uses these. The visit_usize case is better served by going through visit_seq like Bincode does. The visit_bytes case causes problems with supporting no_std: https://github.com/serde-rs/serde/issues/666#issuecomment-271109...
0.8
41931224721513c95c4f70ed953ad258d96ebec3
diff --git a/serde_codegen/src/de.rs b/serde_codegen/src/de.rs index 163874bd4..e75a7c961 100644 --- a/serde_codegen/src/de.rs +++ b/serde_codegen/src/de.rs @@ -1,5 +1,5 @@ use syn::{self, aster}; -use quote::{self, Tokens}; +use quote::Tokens; use bound; use internals::ast::{Body, Field, Item, Style, Variant}; @@...
serde-rs/serde
2017-01-08T09:02:40Z
88a4ed9cd72c7901b025dd8c8d86c69cfc6b157f
664
diff --git a/serde_test/src/de.rs b/serde_test/src/de.rs index 121dcd3db..101cc8110 100644 --- a/serde_test/src/de.rs +++ b/serde_test/src/de.rs @@ -856,7 +856,7 @@ impl<'a, I> VariantVisitor for DeserializerVariantVisitor<'a, I> } } Some(_) => { - Deserialize::...
[ "657" ]
serde-rs__serde-664
Remove the Deserialize trait bound on Visitor::Value The deserializer should not be able to produce an Ok(Value) any other way - it should be forced to call one of the Visitor methods. Also the type may legitimately not implement Deserialize, for example if you are doing stateful deserialization https://github.com/s...
0.8
a3f556959f18c33388a3e4d4498c1b7abf6197e4
diff --git a/serde/src/de/mod.rs b/serde/src/de/mod.rs index 210b36b03..9c930a9ff 100644 --- a/serde/src/de/mod.rs +++ b/serde/src/de/mod.rs @@ -412,7 +412,7 @@ pub trait Deserializer { /// This trait represents a visitor that walks through a deserializer. pub trait Visitor { /// The value produced by this visit...
serde-rs/serde
2017-01-03T09:53:24Z
An example of doing it wrong: [`serde_test::DeserializerVariantVisitor`](https://github.com/serde-rs/serde/blob/v0.8.21/serde_test/src/de.rs#L859) (this is a bug).
88a4ed9cd72c7901b025dd8c8d86c69cfc6b157f
653
diff --git a/testing/tests/test_ser.rs b/testing/tests/test_ser.rs index 525735339..afc850700 100644 --- a/testing/tests/test_ser.rs +++ b/testing/tests/test_ser.rs @@ -30,12 +30,20 @@ struct Struct { c: i32, } -#[derive(Serialize)] +#[derive(Serialize, PartialEq, Debug)] enum Enum { Unit, One(i32), ...
[ "638" ]
serde-rs__serde-653
Attribute skip_(de)serializing for enum variants ```rust #[derive(Serialize, Deserialize)] enum E { A, B, #[serde(skip_serializing, skip_deserializing)] C, } ``` When serializing, it should error when the value is that variant. When deserializing, it should be like the variant doesn't exist.
0.8
2c984980a0faf0670d1d461aac2527feecf9de30
diff --git a/serde_codegen/src/ser.rs b/serde_codegen/src/ser.rs index 449a24904..531604169 100644 --- a/serde_codegen/src/ser.rs +++ b/serde_codegen/src/ser.rs @@ -255,77 +255,123 @@ fn serialize_variant( let variant_ident = variant.ident.clone(); let variant_name = variant.attrs.name().serialize_name(); -...
serde-rs/serde
2016-12-24T12:01:17Z
#641 handles `skip_deserializing`. We still need to do `skip_serializing`.
88a4ed9cd72c7901b025dd8c8d86c69cfc6b157f
652
diff --git a/testing/tests/test.rs.in b/testing/tests/test.rs.in index fca1ff465..04c69bdba 100644 --- a/testing/tests/test.rs.in +++ b/testing/tests/test.rs.in @@ -1,4 +1,5 @@ extern crate serde; +extern crate serde_test; #[macro_use] mod macros; diff --git a/testing/tests/test_bytes.rs b/testing/tests/test_bytes...
[ "645" ]
serde-rs__serde-652
Support deserializing ByteBuf from string A Deserializer should be able to call visit_str or visit_string to deserialize a ByteBuf. The ByteBuf should contain the bytes of the string. Relevant to https://github.com/traviskaufman/serde_bencode/issues/7.
0.8
2c984980a0faf0670d1d461aac2527feecf9de30
diff --git a/serde/src/bytes.rs b/serde/src/bytes.rs index a732712cf..988c24a55 100644 --- a/serde/src/bytes.rs +++ b/serde/src/bytes.rs @@ -84,7 +84,7 @@ mod bytebuf { use de; #[cfg(feature = "collections")] - use collections::Vec; + use collections::{String, Vec}; /// `ByteBuf` wraps a `Vec<u...
serde-rs/serde
2016-12-24T04:18:37Z
88a4ed9cd72c7901b025dd8c8d86c69cfc6b157f
578
diff --git a/serde_derive/tests/compile-fail/duplicate-attribute/rename-and-ser.rs b/serde_derive/tests/compile-fail/duplicate-attribute/rename-and-ser.rs index 312583af1..2f3950f8e 100644 --- a/serde_derive/tests/compile-fail/duplicate-attribute/rename-and-ser.rs +++ b/serde_derive/tests/compile-fail/duplicate-attribu...
[ "575" ]
serde-rs__serde-578
Update from rustc_macro to proc_macro The name of attributes / libraries / feature flags / everything related to Macros 1.1 has changed from rustc_macro to proc_macro in https://github.com/rust-lang/rust/pull/36945 and https://github.com/rust-lang/cargo/pull/3160. We need to update serde_derive to use the new names.
0.8
e547a06639ee62dff40813afbbdbf1ed1d792e99
diff --git a/README.md b/README.md index 448d9ed28..f0c6ec643 100644 --- a/README.md +++ b/README.md @@ -15,7 +15,7 @@ You may be looking for: ## Serde in action ```rust -#![feature(rustc_macro)] +#![feature(proc_macro)] #[macro_use] extern crate serde_derive; diff --git a/examples/serde-syntex-example/src/main...
serde-rs/serde
2016-10-08T22:32:45Z
88a4ed9cd72c7901b025dd8c8d86c69cfc6b157f
2,422
diff --git a/test_suite/tests/regression/issue2415.rs b/test_suite/tests/regression/issue2415.rs new file mode 100644 index 000000000..31c522709 --- /dev/null +++ b/test_suite/tests/regression/issue2415.rs @@ -0,0 +1,5 @@ +use serde_derive::Serialize; + +#[derive(Serialize)] +#[serde()] +pub struct S;
[ "2415", "2421" ]
serde-rs__serde-2422
Possible semver breaking change in serde 1.0.157 When pinning the version of `serde` in Cargo.toml to `1.0.156`, my crate compiles, switching to `1.0.157` results in an error in my `influxdb2` dependency: ```bash error: unexpected end of input, unexpected token in nested attribute, expected ident --> /home/midas/....
1.15
acfd19cb463ea77d2146d637c35d8751601dae75
diff --git a/serde_derive/src/internals/attr.rs b/serde_derive/src/internals/attr.rs index 9875b6606..b0a7d08a2 100644 --- a/serde_derive/src/internals/attr.rs +++ b/serde_derive/src/internals/attr.rs @@ -307,6 +307,12 @@ impl Container { continue; } + if let syn::Meta::List(m...
serde-rs/serde
2023-03-28T05:02:10Z
acfd19cb463ea77d2146d637c35d8751601dae75
2,410
diff --git a/test_suite/tests/regression/issue2409.rs b/test_suite/tests/regression/issue2409.rs new file mode 100644 index 000000000..2e3936691 --- /dev/null +++ b/test_suite/tests/regression/issue2409.rs @@ -0,0 +1,11 @@ +use serde::Deserialize; + +macro_rules! bug { + ($serde_path:literal) => { + #[derive(...
[ "2409" ]
serde-rs__serde-2410
Regression in serde 1.0.157 when a proc macro enables syn's full feature With https://github.com/nox/serde-none-group, the command `cargo build --features syn-full` fails while the command `cargo build` succeeds. The only difference is that the failing command enables the "full" feature of the crate `syn` from a proc m...
1.15
479a00a215cd291e9c0b0ad0373e26694018bf1c
diff --git a/serde_derive/Cargo.toml b/serde_derive/Cargo.toml index f2a3e37cf..b824bb993 100644 --- a/serde_derive/Cargo.toml +++ b/serde_derive/Cargo.toml @@ -24,7 +24,7 @@ proc-macro = true [dependencies] proc-macro2 = "1.0" quote = "1.0" -syn = "2.0" +syn = "2.0.3" [dev-dependencies] serde = { version = "1.0...
serde-rs/serde
2023-03-20T11:04:16Z
acfd19cb463ea77d2146d637c35d8751601dae75
2,403
diff --git a/test_suite/tests/test_annotations.rs b/test_suite/tests/test_annotations.rs index 8ff725e01..2b8bed664 100644 --- a/test_suite/tests/test_annotations.rs +++ b/test_suite/tests/test_annotations.rs @@ -2395,6 +2395,159 @@ fn test_untagged_enum_containing_flatten() { ); } +#[test] +fn test_partially_u...
[ "1402" ]
serde-rs__serde-2403
#[serde(flatten)] attribute support for enum variants Hi! I have the following use-case for a feature similar to `#[serde(flatten)]`, but for enum variants rather than struct fields. Generally happy to dive into `serde_derive` myself in case you think this would be worth having, but it could also be that I'm trying ...
1.15
acfd19cb463ea77d2146d637c35d8751601dae75
diff --git a/serde_derive/src/de.rs b/serde_derive/src/de.rs index d4238f13e..0b86cde6f 100644 --- a/serde_derive/src/de.rs +++ b/serde_derive/src/de.rs @@ -1165,6 +1165,32 @@ fn deserialize_enum( params: &Parameters, variants: &[Variant], cattrs: &attr::Container, +) -> Fragment { + // The variants h...
serde-rs/serde
2023-03-16T17:02:17Z
It looks like this is similar to #1350 where you suggest manually flattening, but that would be fairly cumbersome in this case (`Expression` might have a couple dozen variants) and I guess I'm still curious whether there's a "better" way to do this. Seems reasonable! I would be interested in supporting this better. I h...
acfd19cb463ea77d2146d637c35d8751601dae75
2,387
diff --git a/test_suite/tests/test_annotations.rs b/test_suite/tests/test_annotations.rs index f32c22c1a..8ff725e01 100644 --- a/test_suite/tests/test_annotations.rs +++ b/test_suite/tests/test_annotations.rs @@ -642,7 +642,7 @@ fn test_unknown_field_rename_struct() { Token::Str("a4"), Token::...
[ "1504" ]
serde-rs__serde-2387
Field aliases do not work in combination with `flatten` Field aliases as introduced in https://github.com/serde-rs/serde/pull/1458 do not work when deserializing "flattened" structures. ### Expected behavior Field aliases should work when deserializing "flattened" structures. ### Observed behavior Field a...
1.15
a13c6382b6a4644be01b9b6c331b0cc32d32bc58
diff --git a/serde_derive/src/de.rs b/serde_derive/src/de.rs index a703adaf7..679086e61 100644 --- a/serde_derive/src/de.rs +++ b/serde_derive/src/de.rs @@ -2400,7 +2400,8 @@ fn deserialize_struct_as_struct_visitor( .collect(); let fields_stmt = { - let field_names = field_names_idents.iter().map...
serde-rs/serde
2023-03-05T07:27:53Z
I just got bitten by this too. This is a bug -- I would accept a PR to fix this. I'll try to give it a shot, but first time poking at serde_derive, so no promises :) Yeah, I just got bit by this too. I was using variants, so I thought it was related to that. I can take a try at fixing this if there's been no pr...
acfd19cb463ea77d2146d637c35d8751601dae75
2,383
diff --git a/test_suite/tests/test_gen.rs b/test_suite/tests/test_gen.rs index c053d6903..f6782f68f 100644 --- a/test_suite/tests/test_gen.rs +++ b/test_suite/tests/test_gen.rs @@ -892,3 +892,53 @@ pub struct RemotePackedNonCopyDef { impl Drop for RemotePackedNonCopyDef { fn drop(&mut self) {} } + +/////////////...
[ "2371" ]
serde-rs__serde-2383
"undeclared lifetime" when derive `Deserialize` for an enum with both `flatten` and `'static` fields The following code panics with the error: ```rust use serde::Deserialize; #[derive(Deserialize)] struct Nested { float: f64, } #[derive(Deserialize)] #[serde(tag = "tag")] enum InternallyTagged { F...
1.15
a13c6382b6a4644be01b9b6c331b0cc32d32bc58
diff --git a/serde_derive/src/de.rs b/serde_derive/src/de.rs index a703adaf7..c0701625f 100644 --- a/serde_derive/src/de.rs +++ b/serde_derive/src/de.rs @@ -954,6 +954,7 @@ fn deserialize_struct( lifetime: _serde::__private::PhantomData, } }; + let need_seed = deserializer.is_none(); ...
serde-rs/serde
2023-02-25T17:19:12Z
Since `&'static str` usually represents a string embedded in the binary of your program I don't think it's possible to or even makes sense to `Deserialize` one? Surely what you want to do is this: ```rs #[derive(Deserialize)] #[serde(tag = "tag")] enum InternallyTagged<'a> { Flatten { #[serde(flatten)...
acfd19cb463ea77d2146d637c35d8751601dae75
2,330
diff --git a/test_suite/tests/test_remote.rs b/test_suite/tests/test_remote.rs index c6d276e10..d997f0f01 100644 --- a/test_suite/tests/test_remote.rs +++ b/test_suite/tests/test_remote.rs @@ -79,6 +79,13 @@ mod remote { pub value: T, } + impl<T> StructGeneric<T> { + #[allow(dead_code)] + ...
[ "2327" ]
serde-rs__serde-2330
Deriving Serialize for a struct from external crate which uses generics Hello, I'm trying to implement Serialize for [nom_locate's](https://docs.rs/nom_locate/latest/nom_locate/) [`LocatedSpan`](https://docs.rs/nom_locate/latest/nom_locate/struct.LocatedSpan.html). It's a struct that encapsulates a token with some lo...
1.14
fabbd2b0970084fe8f85bf850f8cd295dd12b305
diff --git a/serde_derive/src/de.rs b/serde_derive/src/de.rs index ef7a2dee2..a703adaf7 100644 --- a/serde_derive/src/de.rs +++ b/serde_derive/src/de.rs @@ -712,8 +712,9 @@ fn deserialize_seq( if params.has_getter { let this_type = &params.this_type; + let (_, ty_generics, _) = params.generics.sp...
serde-rs/serde
2022-11-28T01:18:49Z
fabbd2b0970084fe8f85bf850f8cd295dd12b305
2,159
diff --git a/test_suite/tests/test_de_error.rs b/test_suite/tests/test_de_error.rs index da6e7656c..db534e983 100644 --- a/test_suite/tests/test_de_error.rs +++ b/test_suite/tests/test_de_error.rs @@ -379,57 +379,81 @@ fn test_nonzero_i8() { let test = assert_de_tokens_error::<NonZeroI8>; // from zero - ...
[ "2158" ]
serde-rs__serde-2159
Improve errors upon deserializing NonZero* Deserializing errors on NonZero* could definitely be improved, not sure if this is possible, but somehow the normal numerical types have better information attached to them - is this fault of the wrappers - probably not. <details> <summary> silly example code </summary>...
1.13
bc66aeb0d316d608e43061372d6bd400aad2d393
diff --git a/serde/src/de/impls.rs b/serde/src/de/impls.rs index d7c57568a..c5aaca723 100644 --- a/serde/src/de/impls.rs +++ b/serde/src/de/impls.rs @@ -81,8 +81,34 @@ impl<'de> Deserialize<'de> for bool { //////////////////////////////////////////////////////////////////////////////// macro_rules! impl_deserialize...
serde-rs/serde
2022-01-21T06:01:53Z
bc66aeb0d316d608e43061372d6bd400aad2d393
2,079
diff --git a/test_suite/tests/test_gen.rs b/test_suite/tests/test_gen.rs index 93d5cd3a5..3e19d3c4b 100644 --- a/test_suite/tests/test_gen.rs +++ b/test_suite/tests/test_gen.rs @@ -846,3 +846,19 @@ where { T::deserialize(deserializer) } + +/////////////////////////////////////////////////////////////////////////...
[ "2078" ]
serde-rs__serde-2079
Unaligned reference warnings on packed structures I'm getting `unaligned_references` warnings using `serde_derive` on packed structures. From #1747 and #1813 it seems like this should have been fixed since these structures also impl's `Copy`. Here's a minimal example that reproduces what I'm seeing w/ serde 1.0....
1.12
55fdbea20bbcb0f0fa7e9545ec0ea10c87a6643d
diff --git a/serde_derive/src/de.rs b/serde_derive/src/de.rs index 0903a084c..7f4d7c441 100644 --- a/serde_derive/src/de.rs +++ b/serde_derive/src/de.rs @@ -36,7 +36,7 @@ pub fn expand_derive_deserialize( let impl_block = if let Some(remote) = cont.attrs.remote() { let vis = &input.vis; - let use...
serde-rs/serde
2021-08-23T17:19:46Z
55fdbea20bbcb0f0fa7e9545ec0ea10c87a6643d
1,997
diff --git a/test_suite/tests/test_de.rs b/test_suite/tests/test_de.rs index cb861374f..fb77f8d91 100644 --- a/test_suite/tests/test_de.rs +++ b/test_suite/tests/test_de.rs @@ -15,7 +15,7 @@ use std::sync::atomic::{ AtomicUsize, Ordering, }; use std::sync::{Arc, Weak as ArcWeak}; -use std::time::{Duration, UNIX_...
[ "1996" ]
serde-rs__serde-1997
Serde panics when deserializing a malformed SystemTime `SystemTime` is deserialized by deserializing a duration and adding it to `UNIX_EPOCH`. The documentation states that this can panic `if the resulting point in time cannot be represented by the underlying data structure`. Also, `SystemTime` panics in the same si...
1.12
c26101532509477132721a8c56b7024891aaf1b4
diff --git a/serde/build.rs b/serde/build.rs index ca991a5d2..d9fd94074 100644 --- a/serde/build.rs +++ b/serde/build.rs @@ -76,12 +76,14 @@ fn main() { println!("cargo:rustc-cfg=serde_derive"); } - // TryFrom, Atomic types, and non-zero signed integers stabilized in Rust 1.34: + // TryFrom, Atomi...
serde-rs/serde
2021-03-06T01:17:16Z
55fdbea20bbcb0f0fa7e9545ec0ea10c87a6643d
1,970
diff --git a/test_suite/tests/test_gen.rs b/test_suite/tests/test_gen.rs index 73f7fa50f..002885ead 100644 --- a/test_suite/tests/test_gen.rs +++ b/test_suite/tests/test_gen.rs @@ -7,6 +7,7 @@ #![allow( unknown_lints, mixed_script_confusables, + clippy::ptr_arg, clippy::trivially_copy_pass_by_ref )]...
[ "1969" ]
serde-rs__serde-1970
`Self` does not work in a field with serialize_with Closely related to #1506, #1565, #1830. ```rust use serde::{Serialize, Serializer}; #[derive(Serialize)] struct Struct { #[serde(serialize_with = "vec_first_element")] vec: Vec<Self>, // works if written as Vec<Struct> } fn vec_first_element<S>(...
1.12
b054ea41053ea4047882cc33970d2257cdfe04ac
diff --git a/serde_derive/src/lib.rs b/serde_derive/src/lib.rs index f10baac01..8d1464ec8 100644 --- a/serde_derive/src/lib.rs +++ b/serde_derive/src/lib.rs @@ -79,8 +79,8 @@ mod try; #[proc_macro_derive(Serialize, attributes(serde))] pub fn derive_serialize(input: TokenStream) -> TokenStream { - let input = par...
serde-rs/serde
2021-01-25T21:27:57Z
55fdbea20bbcb0f0fa7e9545ec0ea10c87a6643d
1,001
diff --git a/test_suite/tests/test_gen.rs b/test_suite/tests/test_gen.rs index a82801b91..5d708014b 100644 --- a/test_suite/tests/test_gen.rs +++ b/test_suite/tests/test_gen.rs @@ -358,6 +358,21 @@ fn test_gen() { #[serde(borrow, with = "StrDef")] s: Str<'a>, } + + mod vis { + pub struc...
[ "1000" ]
serde-rs__serde-1001
Remote type visibility The `#[serde(remote)]` attribute generates `deserialize` and `serialize` methods in an impl for the local type but they're private - this prevents a remote "Def" type in a module from being exported outside of it. There are workarounds (newtype or redefine proxy functions) but would there be any ...
1.0
e6487cf6fa216c303491983aa3d74ea76f3215b6
diff --git a/serde_derive/src/de.rs b/serde_derive/src/de.rs index 03e265af4..647281c01 100644 --- a/serde_derive/src/de.rs +++ b/serde_derive/src/de.rs @@ -28,9 +28,10 @@ pub fn expand_derive_deserialize(input: &syn::DeriveInput) -> Result<Tokens, Str let body = Stmts(deserialize_body(&cont, &params)); let...
serde-rs/serde
2017-07-26T06:56:06Z
4cea81f93f12473e0ccbdad2ddecff1f7ea85402
995
diff --git a/test_suite/no_std/Cargo.toml b/test_suite/no_std/Cargo.toml index 80bb6ed2a..7c0040666 100644 --- a/test_suite/no_std/Cargo.toml +++ b/test_suite/no_std/Cargo.toml @@ -7,3 +7,5 @@ publish = false libc = { version = "0.2", default-features = false } serde = { path = "../../serde", default-features = false...
[ "988" ]
serde-rs__serde-995
No bin target named `serde_derive_tests_no_std` In the test_suite directory, `cargo clippy` fails with this message. ``` error: no bin target named `serde_derive_tests_no_std` ``` As though it sees that bin target in test_suite/no_std/Cargo.toml but tries to build that bin target against test_suite/Cargo.toml. ...
1.0
1c2a4bff1c01d0c84b653f2f4313f15621aa3168
diff --git a/Cargo.toml b/Cargo.toml index 29db37d3b..cb35f8543 100644 --- a/Cargo.toml +++ b/Cargo.toml @@ -5,5 +5,4 @@ members = [ "serde_derive_internals", "serde_test", "test_suite", - "test_suite/no_std", ] diff --git a/travis.sh b/travis.sh index ddfbc3339..c43cb55f9 100755 --- a/travis.sh +++ ...
serde-rs/serde
2017-07-21T07:09:16Z
4cea81f93f12473e0ccbdad2ddecff1f7ea85402
971
diff --git a/test_suite/tests/test_gen.rs b/test_suite/tests/test_gen.rs index 2f3d6299a..a82801b91 100644 --- a/test_suite/tests/test_gen.rs +++ b/test_suite/tests/test_gen.rs @@ -330,6 +330,34 @@ fn test_gen() { struct EmptyArray { empty: [X; 0], } + + enum Or<A, B> { + A(A), + B(B...
[ "972" ]
serde-rs__serde-971
Deriving Deserialize for a remote generic enum fails I'm attempting to derive `Deserialize` and `Serialize` for the [`Or` enum from the `or` crate](https://docs.rs/or/0.0.1/or/enum.Or.html). Deriving `Serialize` works fine: ```rust #[derive(Serialize)] #[serde(untagged, remote = "Or")] enum OrDef<A, B> { A(A), B(B...
1.0
4fdba725fe30dc30658b8e71bead7d9b83ab9b0f
diff --git a/serde_derive/src/de.rs b/serde_derive/src/de.rs index 20ccb696f..03e265af4 100644 --- a/serde_derive/src/de.rs +++ b/serde_derive/src/de.rs @@ -23,15 +23,14 @@ pub fn expand_derive_deserialize(input: &syn::DeriveInput) -> Result<Tokens, Str let ident = &cont.ident; let params = Parameters::new(...
serde-rs/serde
2017-06-30T03:13:43Z
4cea81f93f12473e0ccbdad2ddecff1f7ea85402
963
diff --git a/test_suite/tests/test_identifier.rs b/test_suite/tests/test_identifier.rs index 76f1cffae..83a0d3b17 100644 --- a/test_suite/tests/test_identifier.rs +++ b/test_suite/tests/test_identifier.rs @@ -23,7 +23,10 @@ fn test_variant_identifier() { Bbb, } + assert_de_tokens(&V::Aaa, &[Token::U8...
[ "962" ]
serde-rs__serde-963
Numeric variant identifier deserialization only handles u32 Even if the identifier is serialized as a u32, some formats like CBOR store numeric values in a variable length encoding. Small variant numbers are then deserialized as an e.g. u8. Unfortunately, the generated visitor only implements `visit_u32`, but `visit_u8...
1.0
4fdba725fe30dc30658b8e71bead7d9b83ab9b0f
diff --git a/serde_derive/src/de.rs b/serde_derive/src/de.rs index 20ccb696f..741baec77 100644 --- a/serde_derive/src/de.rs +++ b/serde_derive/src/de.rs @@ -1387,10 +1387,10 @@ fn deserialize_identifier( }; let visit_index = if is_variant { - let variant_indices = 0u32..; + let variant_indices...
serde-rs/serde
2017-06-20T03:22:48Z
4cea81f93f12473e0ccbdad2ddecff1f7ea85402
949
diff --git a/test_suite/tests/test_de.rs b/test_suite/tests/test_de.rs index d1bddfaf4..b915fba76 100644 --- a/test_suite/tests/test_de.rs +++ b/test_suite/tests/test_de.rs @@ -14,7 +14,7 @@ extern crate serde_derive; use std::collections::{BTreeMap, BTreeSet, HashMap, HashSet}; use std::net; use std::path::{Path, P...
[ "535" ]
serde-rs__serde-949
impl Serialize and Deserialize for std::time::SystemTime If there is a different type that is better suited for timestamps in serialized data that's already supported by serde, I'd be happy to know about it. The serialized format should probably be [ISO 8601](https://en.wikipedia.org/wiki/ISO_8601).
1.0
fd3d1396d33a49200daaaf8bf17eba78fe4183d8
diff --git a/serde/src/de/impls.rs b/serde/src/de/impls.rs index 13f5c9966..1a48f8031 100644 --- a/serde/src/de/impls.rs +++ b/serde/src/de/impls.rs @@ -1364,6 +1364,132 @@ impl<'de> Deserialize<'de> for Duration { //////////////////////////////////////////////////////////////////////////////// +#[cfg(feature = "s...
serde-rs/serde
2017-06-04T08:40:20Z
:+1: We're going to have to use this approach mentioned in the documentation: > Although a `SystemTime` cannot be directly inspected, the `UNIX_EPOCH` constant is provided in this module as an anchor in time to learn information about a `SystemTime`. By calculating the duration from this fixed point in time, a SystemT...
4cea81f93f12473e0ccbdad2ddecff1f7ea85402
935
diff --git a/test_suite/tests/test_annotations.rs b/test_suite/tests/test_annotations.rs index f2604d456..cb9642e8f 100644 --- a/test_suite/tests/test_annotations.rs +++ b/test_suite/tests/test_annotations.rs @@ -580,6 +580,41 @@ fn test_skip_serializing_struct() { ); } +#[derive(Debug, PartialEq, Serialize, De...
[ "931" ]
serde-rs__serde-935
Warn when using only one of skip_serializing and skip_deserializing I used the `skip_serializing` attribute for a field, but forgot to add `skip_deserializing` too. After serializing my data to a binary format (with `bincode`), I tried to deserialize it but got an error which took a long time to track down. ``` /* ...
1.0
4d5e450054431abb44e35c9d0c73d89b5741f05b
diff --git a/serde_derive_internals/src/attr.rs b/serde_derive_internals/src/attr.rs index f60219103..caf306cfa 100644 --- a/serde_derive_internals/src/attr.rs +++ b/serde_derive_internals/src/attr.rs @@ -719,6 +719,12 @@ impl Field { skip_deserializing.set_true(); } + ...
serde-rs/serde
2017-05-16T09:39:16Z
From some quick github searches it looks like having one but not both of `skip_serializing` and `skip_deserializing` on a type that derives both `Serialize` and `Deserialize` is fairly uncommon but there are some legitimate use cases. I would be on board with adding a single attribute that means both, and improving the...
4cea81f93f12473e0ccbdad2ddecff1f7ea85402
933
diff --git a/test_suite/tests/test_macros.rs b/test_suite/tests/test_macros.rs index 651bcfea1..c4c615664 100644 --- a/test_suite/tests/test_macros.rs +++ b/test_suite/tests/test_macros.rs @@ -695,6 +695,35 @@ fn test_internally_tagged_enum() { ); } +#[test] +fn test_internally_tagged_struct_variant_containing_...
[ "932" ]
serde-rs__serde-933
Regression deserializing string Enums between 1.0.3 and 1.0.4 The example in https://gist.github.com/anonymous/440028c896809d8b3b51d8333cb6d2f9 works for me with `serde = "=1.0.3"`, `serde_json = "1.0.2"` and `serde_derive = "1.0.4"` and current stable rust, but errors when specifying `serde = "=1.0.4"` (or less precis...
1.0
c68b95969679a4d67180d14cf3b200dc9ce34de1
diff --git a/serde/src/private/de.rs b/serde/src/private/de.rs index 2c5648cd5..19736a14b 100644 --- a/serde/src/private/de.rs +++ b/serde/src/private/de.rs @@ -1082,7 +1082,7 @@ mod content { } (variant, Some(value)) } - Content::String(variant)...
serde-rs/serde
2017-05-14T19:30:53Z
4cea81f93f12473e0ccbdad2ddecff1f7ea85402
926
diff --git a/test_suite/tests/test_macros.rs b/test_suite/tests/test_macros.rs index 6c7e4b2ab..651bcfea1 100644 --- a/test_suite/tests/test_macros.rs +++ b/test_suite/tests/test_macros.rs @@ -695,6 +695,30 @@ fn test_internally_tagged_enum() { ); } +#[test] +fn test_internally_tagged_borrow() { + #[derive(D...
[ "906" ]
serde-rs__serde-926
Support borrowing within internally tagged enum Serde panics when trying to deserialize an struct which has a borrowed field, which also contains a borrowed field. For example, the following code compiles fine: **EDIT:** fixed the example ```rust #[macro_use] extern crate serde_derive; extern crate serde; ext...
1.0
58e30eaee4f62acc621ce95635fa7c7f8fe5c123
diff --git a/serde/src/private/de.rs b/serde/src/private/de.rs index 45b383fb0..2c5648cd5 100644 --- a/serde/src/private/de.rs +++ b/serde/src/private/de.rs @@ -233,7 +233,7 @@ mod content { /// /// Not public API. Use serde-value instead. #[derive(Debug)] - pub enum Content { + pub enum Content<'d...
serde-rs/serde
2017-05-11T02:57:59Z
I don't think this is correctly minimized from what you showed in IRC earlier. [It gives a different error.](http://play.integer32.com/?gist=29b4a709bd0aecba7a5aa19024619778) Side note: the panic is coming from your code not Serde, because you unwrap the error. :wink: I minimized it to this: ```rust #[macro_use] ...
4cea81f93f12473e0ccbdad2ddecff1f7ea85402
905
diff --git a/test_suite/tests/test_macros.rs b/test_suite/tests/test_macros.rs index d4144ab3d..6c7e4b2ab 100644 --- a/test_suite/tests/test_macros.rs +++ b/test_suite/tests/test_macros.rs @@ -751,6 +751,31 @@ fn test_adjacently_tagged_enum() { ], ); + // unit with excess content (f, g, h) + asser...
[ "816" ]
serde-rs__serde-905
Tag/Content attributes should silently ignore unkown attributes Currently if I use `#[serde(tag, content)]` on a tagged enum, it seems that serde would throw error when the JSON string contains attributes other than the tag and the content. I think silently ignore the unknown field could be better?
1.0
c96efcb87ab72f14a2a084644308c4a5e98ed820
diff --git a/serde/src/private/de.rs b/serde/src/private/de.rs index 270e1df8a..45b383fb0 100644 --- a/serde/src/private/de.rs +++ b/serde/src/private/de.rs @@ -16,6 +16,7 @@ use de::Unexpected; #[cfg(any(feature = "std", feature = "collections"))] pub use self::content::{Content, ContentRefDeserializer, ContentDeser...
serde-rs/serde
2017-04-27T18:25:41Z
Really we should respect deny_unknown_fields. ```rust // as you said #[derive(Serialize, Deserialize)] #[serde(tag = "t", content = "c")] enum E { /* ... */ } // current behavior #[derive(Serialize, Deserialize)] #[serde(tag = "t", content = "c", deny_unknown_fields)] enum F { /* ... */ } ``` I would w...
4cea81f93f12473e0ccbdad2ddecff1f7ea85402
452
diff --git a/serde_macros/tests/test.rs b/serde_macros/tests/test.rs index e92112185..e04627877 100644 --- a/serde_macros/tests/test.rs +++ b/serde_macros/tests/test.rs @@ -1,7 +1,6 @@ #![feature(test, custom_attribute, custom_derive, plugin)] #![plugin(serde_macros)] -extern crate serde; extern crate test; inc...
[ "219" ]
serde-rs__serde-452
Serializer and Deserializer should be pure interface For now the Serializer and Deserializer traits contain many functions, but many of them have default implementations. A problem is, the default implementations make too many assumption about the serializing protocol, probably based on json. Though them may be conveni...
null
d751b4c39a956f3b36dd967d940a2f6fd0c2718e
diff --git a/serde/src/de/impls.rs b/serde/src/de/impls.rs index 4096fa8fb..1680312b0 100644 --- a/serde/src/de/impls.rs +++ b/serde/src/de/impls.rs @@ -593,7 +593,7 @@ macro_rules! array_impls { fn deserialize<D>(deserializer: &mut D) -> Result<[T; $len], D::Error> where D: Deseri...
serde-rs/serde
2016-07-20T08:47:40Z
I agree. This "magic" has caused me some headache debugging, so I ended up replacing all of them manually. +1. I think magic layers are more appropriate for Deserialize/Serialize sugar, which are types that many serde users need to interact with. The -er types are relatively rarely implemented, and harder for an autho...
8cb6607e827717136a819caa44d8e43702724883
446
diff --git a/serde_test/Cargo.toml b/serde_test/Cargo.toml index 38d167027..9855d12fd 100644 --- a/serde_test/Cargo.toml +++ b/serde_test/Cargo.toml @@ -2,7 +2,7 @@ name = "serde_test" # DO NOT RELEASE ANY MORE 0.7 RELEASES FROM THIS BRANCH # USE THE 0.7.x BRANCH -version = "0.8.0-rc1" +version = "0.8.0-rc2" author...
[ "404" ]
serde-rs__serde-446
Rename feature "nightly-testing" to "unstable-testing" From what I understand, the community has started to use the feature `unstable*` instead of `nightly*` for things that may depend on the nightly compiler. If this is the case, when we do 0.8, we should consider renaming this feature.
null
6c18896cf592238ba643aa9050e536e64d4a732c
diff --git a/.travis.yml b/.travis.yml index 7e3e0d3b9..19a2d6929 100644 --- a/.travis.yml +++ b/.travis.yml @@ -18,15 +18,15 @@ before_script: script: - (cd serde && travis-cargo build) - (cd serde && travis-cargo --skip nightly test) -- (cd serde && travis-cargo --only nightly test -- --features nightly-testing) +...
serde-rs/serde
2016-07-17T20:18:29Z
8cb6607e827717136a819caa44d8e43702724883
437
diff --git a/serde_test/src/de.rs b/serde_test/src/de.rs index a2d4c3d7d..24dabe194 100644 --- a/serde_test/src/de.rs +++ b/serde_test/src/de.rs @@ -135,15 +135,18 @@ impl<I> de::Deserializer for Deserializer<I> Some(Token::Option(true)) => visitor.visit_some(self), Some(Token::Unit) => visito...
[ "386" ]
serde-rs__serde-437
Unable to serialize sequence without knowing all elements. In the logging library I'm working on, I would like to serialize map elements, provided by user API calls. However at any given time, I have only one `key: value` pair. So I can't efficiently serialize a map: ``` fn serialize_map<V>(&mut self, visitor: V)...
0.7
ed603d4580154bc09af94964c2aca8b42f268274
diff --git a/serde/src/ser/impls.rs b/serde/src/ser/impls.rs index d0ac49bc9..a8876c958 100644 --- a/serde/src/ser/impls.rs +++ b/serde/src/ser/impls.rs @@ -64,8 +64,6 @@ use super::{ Error, Serialize, Serializer, - SeqVisitor, - MapVisitor, }; //////////////////////////////////////////////////...
serde-rs/serde
2016-07-12T13:45:46Z
Have you seen https://serde-rs.github.io/serde/serde/ser/impls/struct.MapIteratorVisitor.html ? Either that fits your use case or I don't understand your situation (please elaborate with some code in that case). @oli-obk: This visitor is taking an iterator. I can't provide an iterator since I don't have a collection o...
ed603d4580154bc09af94964c2aca8b42f268274
418
diff --git a/serde_macros/tests/compile_tests.rs b/serde_macros/tests/compile_tests.rs index 27c212b4d..419425226 100644 --- a/serde_macros/tests/compile_tests.rs +++ b/serde_macros/tests/compile_tests.rs @@ -20,6 +20,11 @@ fn run_mode(mode: &'static str) { } #[test] -fn compile_test() { +fn compile_fail() { r...
[ "350" ]
serde-rs__serde-418
serde_macros spans need add expansion info clippy can't detect whether the code was expanded or not without that info cc https://github.com/Manishearth/rust-clippy/issues/969
0.7
5deba439c3e44b5646cef22d34ccc412cbcb23f3
diff --git a/serde_codegen/src/de.rs b/serde_codegen/src/de.rs index c875848cb..69d70bc37 100644 --- a/serde_codegen/src/de.rs +++ b/serde_codegen/src/de.rs @@ -8,6 +8,7 @@ use syntax::ptr::P; use syntax::tokenstream::TokenTree; use bound; +use span; use internals::ast::{Body, Field, Item, Style, Variant}; use in...
serde-rs/serde
2016-07-03T17:54:02Z
ed603d4580154bc09af94964c2aca8b42f268274
412
diff --git a/serde_macros/tests/test.rs b/serde_macros/tests/test.rs index f2e785a56..e92112185 100644 --- a/serde_macros/tests/test.rs +++ b/serde_macros/tests/test.rs @@ -4,6 +4,6 @@ extern crate serde; extern crate test; -include!("../../serde_tests/tests/test.rs.in"); +include!("../../testing/tests/test.rs.in")...
[ "411" ]
serde-rs__serde-412
Publish a serde_test crate containing the Token De/Serializer. [This Serializer and Deserializer](https://github.com/serde-rs/serde/blob/6ab508a93c8163a12315729f00fce4c539dfebd5/serde_tests/tests/token.rs) would be really useful for writing tests in other crates that implement Serialize/Deserialize for their types.
0.7
6ab508a93c8163a12315729f00fce4c539dfebd5
diff --git a/.travis.yml b/.travis.yml index ff56c4826..de282ca26 100644 --- a/.travis.yml +++ b/.travis.yml @@ -22,15 +22,15 @@ script: - (cd serde && travis-cargo build -- --no-default-features) - (cd serde && travis-cargo --only nightly build -- --no-default-features --features alloc) - (cd serde && travis-cargo ...
serde-rs/serde
2016-06-29T05:26:15Z
ed603d4580154bc09af94964c2aca8b42f268274
371
diff --git a/serde_tests/Cargo.toml b/serde_tests/Cargo.toml index adb7a9401..477921763 100644 --- a/serde_tests/Cargo.toml +++ b/serde_tests/Cargo.toml @@ -19,6 +19,7 @@ syntex_syntax = { version = "^0.35.0" } serde_codegen = { version = "*", path = "../serde_codegen", features = ["with-syntex"] } [dev-dependencie...
[ "369" ]
serde-rs__serde-371
implement Serialize + Deserialize for HashMap with custom hash functions As far as I am aware the implementation of HashMap serialization/deserialization is only for SipHasher. ``` impl<K, V> Serialize for HashMap<K, V> where K: Serialize + Eq + Hash, V: Serialize, ``` Is it feasible or desired to impl...
0.7
84a573c92699ef13f27758c543d8b8038964d960
diff --git a/serde/src/de/impls.rs b/serde/src/de/impls.rs index ed727aa23..bdd547c7f 100644 --- a/serde/src/de/impls.rs +++ b/serde/src/de/impls.rs @@ -32,7 +32,7 @@ use collections::enum_set::{CLike, EnumSet}; #[cfg(all(feature = "nightly", feature = "collections"))] use collections::borrow::ToOwned; -use core::h...
serde-rs/serde
2016-06-11T08:11:42Z
Custom hashers were added in Rust 1.7, and since serde is currently tracking the last 3 releases it seems like we could add support for this now. @dtolnay / @oli-obk: what do you two think? Let's do it.
ed603d4580154bc09af94964c2aca8b42f268274
354
diff --git a/serde_tests/tests/test_annotations.rs b/serde_tests/tests/test_annotations.rs index cdaa899ef..79e1683b3 100644 --- a/serde_tests/tests/test_annotations.rs +++ b/serde_tests/tests/test_annotations.rs @@ -441,7 +441,8 @@ enum RenameEnumSerializeDeserialize<A> { #[serde(rename(serialize="dick_grayson", ...
[ "353" ]
serde-rs__serde-354
`rename` attributes overwrite each other ``` rust #![feature(custom_derive, plugin)] #![plugin(serde_macros)] extern crate serde_json; #[derive(Serialize, Deserialize, Default)] struct S { #[serde(rename(serialize="s1", deserialize="d1"))] v1: (), #[serde(rename(serialize="s2"))] #[serde(rename(deseri...
0.7
4b472be56e8b071a12b3b4f03c3829251812453c
diff --git a/serde_codegen/src/attr.rs b/serde_codegen/src/attr.rs index 2a3bd4979..b59e860b8 100644 --- a/serde_codegen/src/attr.rs +++ b/serde_codegen/src/attr.rs @@ -82,7 +82,6 @@ impl ContainerAttrs { // Parse `#[serde(rename="foo")]` ast::MetaItemKind::NameValue(ref name, ...
serde-rs/serde
2016-06-06T08:50:43Z
ed603d4580154bc09af94964c2aca8b42f268274
338
diff --git a/serde_tests/tests/test_gen.rs b/serde_tests/tests/test_gen.rs index 805c100db..8410b2c3c 100644 --- a/serde_tests/tests/test_gen.rs +++ b/serde_tests/tests/test_gen.rs @@ -49,6 +49,12 @@ enum EnumWith<T> { i: i32 }, } +#[derive(Serialize)] +struct MultipleRef<'a, 'b, 'c, T> where T: 'c, 'c: 'b,...
[ "337" ]
serde-rs__serde-338
Double reference generates bad code ``` rust #[derive(Serialize)] pub struct S<'a, 'b, T> where 'b: 'a, T: 'a + 'b { a: &'a &'b T, b: T, } ``` ``` error: cannot infer an appropriate lifetime for borrow expression due to conflicting requirements [E0495] return Ok(Some(try!(_serializer.serialize_struct...
0.7
16d3e96b77fe8d50a2f9d9716ec231fc04251bce
diff --git a/serde_codegen/src/bound.rs b/serde_codegen/src/bound.rs index c46cc2b0b..4f563e9b2 100644 --- a/serde_codegen/src/bound.rs +++ b/serde_codegen/src/bound.rs @@ -139,9 +139,9 @@ fn contains_generic(ty: &ast::Ty, generics: &ast::Generics) -> bool { // // impl<'a, T> Serialize for Test<'a, T> // w...
serde-rs/serde
2016-05-19T06:49:47Z
ed603d4580154bc09af94964c2aca8b42f268274
335
diff --git a/serde_tests/tests/test_gen.rs b/serde_tests/tests/test_gen.rs index 8410b2c3c..6f897e089 100644 --- a/serde_tests/tests/test_gen.rs +++ b/serde_tests/tests/test_gen.rs @@ -1,5 +1,6 @@ // These just test that serde_codegen is able to produce code that compiles -// successfully when there are a variety of g...
[ "330" ]
serde-rs__serde-335
Tuple struct fields with deserialize_with still require Deserialize bound The following complains about `Url` not implementing `Deserialize` ``` rust #[derive(Deserialize)] pub struct SerializableUrl(#[serde(deserialize_with="DeserializeWith::deserialize_with")] Url) ```
0.7
cc115ca43a52e88d4afb2dc50e21c9380812f13b
diff --git a/serde_codegen/src/attr.rs b/serde_codegen/src/attr.rs index d8dc7379f..cc5cfe383 100644 --- a/serde_codegen/src/attr.rs +++ b/serde_codegen/src/attr.rs @@ -177,31 +177,29 @@ pub struct FieldAttrs { name: Name, skip_serializing_field: bool, skip_deserializing_field: bool, - skip_serializin...
serde-rs/serde
2016-05-16T18:46:34Z
ed603d4580154bc09af94964c2aca8b42f268274
1,251
diff --git a/test_suite/tests/test_de.rs b/test_suite/tests/test_de.rs index 3a8fcc448..a7eedba3f 100644 --- a/test_suite/tests/test_de.rs +++ b/test_suite/tests/test_de.rs @@ -17,15 +17,15 @@ use std::ffi::{CString, OsString}; use std::net; use std::num::Wrapping; use std::path::{Path, PathBuf}; -use std::rc::Rc; -...
[ "186" ]
serde-rs__serde-1251
No way to serialize Weak pointers As mentioned on [reddit](https://www.reddit.com/r/rust/comments/3u4ri0/serde_serializing_weak/). Open question: What do we do if the value the `Weak` pointer is pointing at has been destroyed? Should we serialize them as an `Option` pointer?
1.0
dc921892be95f07914fff21410c12310a8305158
diff --git a/serde/src/de/impls.rs b/serde/src/de/impls.rs index 5a53a418e..aa4ede6c4 100644 --- a/serde/src/de/impls.rs +++ b/serde/src/de/impls.rs @@ -1474,6 +1474,44 @@ where //////////////////////////////////////////////////////////////////////////////// +/// This impl requires the [`"rc"`] Cargo feature of Se...
serde-rs/serde
2018-05-07T17:41:10Z
It's also not clear what we should do for deserialization. I'd serialize it as an `Option` pointer and during deserialization, in case it's `None`, just create an `Rc`, make a `Weak` from that and `drop` the `Rc`. @oli-obk: I suppose that would work, but it seems sad that we wouldn't actually be able to deserialize t...
4cea81f93f12473e0ccbdad2ddecff1f7ea85402
1,249
diff --git a/test_suite/tests/test_gen.rs b/test_suite/tests/test_gen.rs index d6921e148..f0255bad3 100644 --- a/test_suite/tests/test_gen.rs +++ b/test_suite/tests/test_gen.rs @@ -567,6 +567,17 @@ fn test_gen() { } assert::<AssocDeriveMulti<i32, NoSerdeImpl>>(); + + #[derive(Serialize)] + #[serde(tag...
[ "1247" ]
serde-rs__serde-1249
Adjacently tagged empty tuple variant or struct variant generates broken Serialize ```rust #[derive(Serialize)] #[serde(tag = "t", content = "c")] enum E { A {}, // same for `A()` } ``` ``` error[E0392]: parameter `'__a` is never used --> src/main.rs:68:10 | 68 | #[derive(Serialize)] | ...
1.0
9dc05c36f0fa9efc66e4bb0ae59822d475c3b0a0
diff --git a/serde_derive/src/ser.rs b/serde_derive/src/ser.rs index 987cc53ae..38b03eece 100644 --- a/serde_derive/src/ser.rs +++ b/serde_derive/src/ser.rs @@ -645,7 +645,7 @@ fn serialize_adjacently_tagged_variant( let (_, ty_generics, where_clause) = params.generics.split_for_impl(); - let wrapper_generi...
serde-rs/serde
2018-05-07T04:50:07Z
4cea81f93f12473e0ccbdad2ddecff1f7ea85402
1,248
diff --git a/test_suite/tests/compile-fail/conflict/flatten-within-enum.rs b/test_suite/tests/compile-fail/conflict/flatten-within-enum.rs deleted file mode 100644 index d2e007d26..000000000 --- a/test_suite/tests/compile-fail/conflict/flatten-within-enum.rs +++ /dev/null @@ -1,21 +0,0 @@ -// Copyright 2018 Serde Devel...
[ "1206" ]
serde-rs__serde-1248
Support flatten within struct variant ```rust #[derive(Serialize, Deserialize)] enum Foo { A { #[serde(flatten)] fields: HashMap<String, String>, } } ``` ``` error: proc-macro derive panicked --> src/main.rs:6:21 | 6 | #[derive(Serialize, Deserialize)] | ...
1.0
5098609935f236a30d9c42356d2941d847c699f6
diff --git a/serde_derive/src/de.rs b/serde_derive/src/de.rs index f1a37b854..e3197fcf4 100644 --- a/serde_derive/src/de.rs +++ b/serde_derive/src/de.rs @@ -838,6 +838,10 @@ fn deserialize_struct( quote! { _serde::Deserializer::deserialize_any(#deserializer, #visitor_expr) } + } else i...
serde-rs/serde
2018-05-07T04:24:19Z
4cea81f93f12473e0ccbdad2ddecff1f7ea85402
1,245
diff --git a/test_suite/tests/test_annotations.rs b/test_suite/tests/test_annotations.rs index 5e38ad34a..815a47b6a 100644 --- a/test_suite/tests/test_annotations.rs +++ b/test_suite/tests/test_annotations.rs @@ -1793,3 +1793,49 @@ fn test_flatten_enum_newtype() { ], ); } + +#[test] +fn test_flatten_inte...
[ "1189" ]
serde-rs__serde-1245
Support deserializing a flattened internally tagged enum This would be valuable to support if possible. Serialization already works but deserialization does not. ```rust #[macro_use] extern crate serde_derive; extern crate serde; extern crate serde_json; #[derive(Serialize, Deserialize, Debug)] struct Data...
1.0
ed425b3a6f1854a9058902aee57263972321fa87
diff --git a/serde/src/de/mod.rs b/serde/src/de/mod.rs index 826662dfa..0a5c487d7 100644 --- a/serde/src/de/mod.rs +++ b/serde/src/de/mod.rs @@ -1132,6 +1132,18 @@ pub trait Deserializer<'de>: Sized { fn is_human_readable(&self) -> bool { true } + + // Not public API. + #[doc(hidden)] + fn p...
serde-rs/serde
2018-05-06T04:43:11Z
4cea81f93f12473e0ccbdad2ddecff1f7ea85402
1,241
diff --git a/test_suite/tests/test_gen.rs b/test_suite/tests/test_gen.rs index e06ec5a88..d6921e148 100644 --- a/test_suite/tests/test_gen.rs +++ b/test_suite/tests/test_gen.rs @@ -333,9 +333,7 @@ fn test_gen() { #[derive(Serialize, Deserialize)] #[serde(untagged, remote = "Or")] enum OrDef<A, B> { - ...
[ "976" ]
serde-rs__serde-1241
Types used in remote derive are dead code Remote derive structs exist only to tell serde_derive what are the fields of the real struct. As such, no code uses the fields of the remote derive struct. ```rust #[macro_use] extern crate serde_derive; extern crate serde; #[derive(Serialize)] #[serde(remote = "S")...
1.0
893c0578ddf4b692d10a44053805fad67fae2fdb
diff --git a/serde_derive/src/de.rs b/serde_derive/src/de.rs index af3d3693e..1ad97006d 100644 --- a/serde_derive/src/de.rs +++ b/serde_derive/src/de.rs @@ -16,6 +16,7 @@ use bound; use fragment::{Expr, Fragment, Match, Stmts}; use internals::ast::{Container, Data, Field, Style, Variant}; use internals::{self, attr}...
serde-rs/serde
2018-05-02T20:30:26Z
Things to be careful about: - Should work for non-public structs. - Should work for structs with fields of non-public type. - Should work for structs with generic type parameters. - Should work for structs with type parameters bound by a non-public trait. - Should work for structs with lifetime parameters. - Al...
4cea81f93f12473e0ccbdad2ddecff1f7ea85402
1,213
diff --git a/test_suite/tests/test_gen.rs b/test_suite/tests/test_gen.rs index 368403e77..aa5e0c4c8 100644 --- a/test_suite/tests/test_gen.rs +++ b/test_suite/tests/test_gen.rs @@ -539,6 +539,30 @@ fn test_gen() { array: [u8; 256], } assert_ser::<BigArray>(); + + trait AssocSerde { + type A...
[ "1208" ]
serde-rs__serde-1213
Infer correct trait bounds for use of a type parameter associated type The following expands to `impl<T: Trait> Serialize for Struct<T>` which fails to compile. ```rust trait Trait { type Assoc; } #[derive(Serialize)] struct Struct<T: Trait> { assoc: T::Assoc, } ``` The builtin derives already g...
1.0
5efb22ebee128f1ec4a755556230bdd66c3f2026
diff --git a/serde_derive/src/bound.rs b/serde_derive/src/bound.rs index 4e7c7d71b..1506a9ba2 100644 --- a/serde_derive/src/bound.rs +++ b/serde_derive/src/bound.rs @@ -50,18 +50,39 @@ pub fn with_where_predicates( generics } -pub fn with_where_predicates_from_fields<F>( +pub fn with_where_predicates_from_field...
serde-rs/serde
2018-04-10T15:26:14Z
4cea81f93f12473e0ccbdad2ddecff1f7ea85402
1,205
diff --git a/test_suite/tests/compile-fail/conflict/flatten-newtype-struct.rs b/test_suite/tests/compile-fail/conflict/flatten-newtype-struct.rs new file mode 100644 index 000000000..07566f8c7 --- /dev/null +++ b/test_suite/tests/compile-fail/conflict/flatten-newtype-struct.rs @@ -0,0 +1,16 @@ +// Copyright 2018 Serde ...
[ "1188" ]
serde-rs__serde-1205
Assertion failed: !cattrs.has_flatten() This should have a better error message, ideally detected by serde_derive_internals. ```rust #[derive(Deserialize)] struct Mitsuhiko(#[serde(flatten)] HashMap<String, String>); ``` ``` error: proc-macro derive panicked --> src/main.rs:18:10 | 18 | #[derive(Deser...
1.0
9c659d9d869c2a175595cd62ea9460009df392c4
diff --git a/serde_derive_internals/src/check.rs b/serde_derive_internals/src/check.rs index 31980a398..6db99a173 100644 --- a/serde_derive_internals/src/check.rs +++ b/serde_derive_internals/src/check.rs @@ -47,11 +47,20 @@ fn check_flatten(cx: &Ctxt, cont: &Container) { Data::Enum(_) => { assert...
serde-rs/serde
2018-04-01T20:12:09Z
4cea81f93f12473e0ccbdad2ddecff1f7ea85402
1,204
diff --git a/test_suite/tests/compile-fail/conflict/flatten-within-enum.rs b/test_suite/tests/compile-fail/conflict/flatten-within-enum.rs new file mode 100644 index 000000000..d2e007d26 --- /dev/null +++ b/test_suite/tests/compile-fail/conflict/flatten-within-enum.rs @@ -0,0 +1,21 @@ +// Copyright 2018 Serde Developer...
[ "1185" ]
serde-rs__serde-1204
Flatten inside a struct variant produces broken Serialize code ```rust #[derive(Serialize)] enum Mitsuhiko { A { #[serde(flatten)] fields: HashMap<String, String>, } } ``` ``` error[E0277]: the trait bound `<__S as _IMPL_SERIALIZE_FOR_Mitsuhiko::_serde::Serializer>::SerializeStructVa...
1.0
9c659d9d869c2a175595cd62ea9460009df392c4
diff --git a/serde_derive_internals/src/ast.rs b/serde_derive_internals/src/ast.rs index fd531d8c7..fd742c9e5 100644 --- a/serde_derive_internals/src/ast.rs +++ b/serde_derive_internals/src/ast.rs @@ -68,6 +68,9 @@ impl<'a> Container<'a> { Data::Enum(ref mut variants) => for variant in variants { ...
serde-rs/serde
2018-04-01T20:02:11Z
4cea81f93f12473e0ccbdad2ddecff1f7ea85402
1,140
diff --git a/test_suite/Cargo.toml b/test_suite/Cargo.toml index 50e10e360..23a9fd79d 100644 --- a/test_suite/Cargo.toml +++ b/test_suite/Cargo.toml @@ -9,6 +9,7 @@ unstable = ["serde/unstable", "compiletest_rs"] [dev-dependencies] fnv = "1.0" +proc-macro2 = "0.2" rustc-serialize = "0.3.16" serde = { path = "../s...
[ "1139" ]
serde-rs__serde-1140
Tuple structs with private fields do not work with proc-macro2/nightly The test suite does not compile, lots of errors like this: ```rust #[derive(Serialize)] struct S(u8); ``` ``` error[E0611]: field `0` of tuple-struct `S` is private --> src/main.rs:8:10 | 8 | #[derive(Serialize)] | ^^^^^^...
1.0
16bc9fb99e20438f9f9cf5ea7aa7b8958f12d696
diff --git a/serde_derive/src/de.rs b/serde_derive/src/de.rs index c9154321a..1e78d52af 100644 --- a/serde_derive/src/de.rs +++ b/serde_derive/src/de.rs @@ -6,7 +6,7 @@ // option. This file may not be copied, modified, or distributed // except according to those terms. -use syn::{self, Ident, Member}; +use syn::{se...
serde-rs/serde
2018-01-10T06:23:55Z
4cea81f93f12473e0ccbdad2ddecff1f7ea85402
1,124
diff --git a/serde_test/src/assert.rs b/serde_test/src/assert.rs index 903a56164..821014097 100644 --- a/serde_test/src/assert.rs +++ b/serde_test/src/assert.rs @@ -195,15 +195,15 @@ where panic!("{} remaining tokens", de.remaining()); } - // Do the same thing for deserialize_from. This isn't *great*...
[ "1122" ]
serde-rs__serde-1124
Is deserialize_from really more like into Insightful comment from @nox in #servo: > **\<nox>** Yeah but I’m curious about the name. > **\<nox>** The idea I heard was more like deserialize_into than deserialize_from.
1.0
0b89bc920e6052c05483c8b56661e859caae9d25
diff --git a/serde/src/de/impls.rs b/serde/src/de/impls.rs index 87aa6c684..ddd1997cd 100644 --- a/serde/src/de/impls.rs +++ b/serde/src/de/impls.rs @@ -15,7 +15,7 @@ use de::{Deserialize, Deserializer, EnumAccess, Error, SeqAccess, Unexpected, Va use de::MapAccess; use de::from_primitive::FromPrimitive; -use priva...
serde-rs/serde
2017-12-17T19:00:22Z
The naming is meant to align with the relatively obscure Clone::clone_from -- ```rust trait Clone { fn clone(&self) -> Self; fn clone_from(&mut self, source: &Self); } ``` ```rust trait Deserialize<'de> { fn deserialize<D>(deserializer: D) -> Result<Self, D::Error>; fn deserialize_from<D>(...
4cea81f93f12473e0ccbdad2ddecff1f7ea85402