text stringlengths 0 1.82k |
|---|
• Twitch blog post on garbage collector latency |
• GopherCon talks and presentations |
• Kubernetes 1.3 release and cluster federation |
• etcd 3.0 release and scalability |
• Traefik.io load balancer 2.0 release |
• Glide 0.11 release and new features |
• Code analysis on Go projects |
• Desirable statistics for code repositories (e.g. function call frequency, external dependencies) |
• Effects of importing packages on project size and complexity |
• Call to action for users to suggest additional statistics or features |
• #FreeSoftwareFriday discussion and shoutouts to open-source projects and maintainers |
• Shoutouts to specific projects (GoKit, The Silver Searcher) for their contributions to the Go ecosystem |
• Discussion of Matt's unannounced work |
• Encouragement to tweet at Matt to release his work publicly |
• Thanks and appreciation for listeners and sponsors |
• Upcoming events: GopherCon, live streaming on Twitch |
• Promotion of GoTime.fm newsletter and social media channels |
**Erik St. Martin:** Alright, we are back for another episode of Go Time. This is episode number 12. Here with me today is Brian Ketelsen... |
**Brian Ketelsen:** Hello. |
**Erik St. Martin:** And I should say I'm Erik St. Martin, I always forget that. Unfortunately Carlisia is not here today with us. She's doing her first week at her new job, but she's here with us in spirit, and I know that she would love being here today, especially because she is a super big fan of our guest today, w... |
**Beyang Liu:** Doing great, great to be here. |
**Brian Ketelsen:** We're all big fans of Sourcegraph. You guys are famous for live-tweeting GopherCon. |
**Erik St. Martin:** That was so awesome. |
**Beyang Liu:** Thanks! |
**Erik St. Martin:** Are you guys doing that again this year? |
**Beyang Liu:** We gotta figure that out, actually. I'm actually not gonna be able to go to GopherCon unfortunately, and it's usually me that's organizing the live tweeting. But we do have Dmitri Shuralyov and Renfred Harper. You guys probably know Dmitri, because he's pretty prominent in the Go open source community, ... |
**Erik St. Martin:** I didn't even realize that he worked for Sourcegraph, that's awesome. |
**Beyang Liu:** Yeah, he's been a member of the team for about a year and a half now, he's one of the more senior members, and he's awesome. |
**Brian Ketelsen:** We gotta get him on the show, too. Everytime I come up with some absolutely crazy Go project his name is on it. I don't' understand that. |
**Beyang Liu:** Yeah, you guys should really have him on the show. He's doing a lot of great stuff in the open source community and he's just a great guy in general. |
**Brian Ketelsen:** We'll get him on the schedule, too. But today it's you, so let's talk about all the awesome tools that you guys have dropped in the last couple weeks... Holy cow! |
**Erik St. Martin:** Yeah, no kidding. |
**Beyang Liu:** Yeah, we've shipped a lot of stuff in the past couple weeks. There's some new editor integrations and native stuff that I'm pretty excited about, that kind of gets the information that Sourcegraph provides just in your editor, and literally one keystroke or zero keystrokes away when you're programming. |
**Erik St. Martin:** Yeah, we were talking about that a couple episodes ago. |
**Beyang Liu:** I guess I should start by giving an intro... I assume most of your listeners probably won't have heard of Sourcegraph, so I should probably say a little bit about what we do, right? |
**Erik St. Martin:** Yeah, definitely. |
**Beyang Liu:** Sourcegraph's essentially a programming assistant that's built on top of the global graph of code. What I mean by that is the assistant part helps you answer a bunch of questions that pop up every single day, every single hour when you're coding. Things like, "How do I use this particular function? Who ... |
\[03:57\] So it's basically this all-in reference guide that lets you answer all these questions that arise over the course of the day. It helps you get the answer within seconds as opposed to minutes, and with a lot less mental energy than something like grep, or just googling for the answer. So that's kind of the sal... |
**Brian Ketelsen:** It's a pretty good sales pitch. |
**Erik St. Martin:** Yeah, a couple of big ones... The ones that we noticed - and see, Carlisia's like the encyclopedia here, because I will say we talked about it a couple episodes ago and she will tell me specifically which episode it was we discussed it... But we were talking about the new editor plugins and how awe... |
**Beyang Liu:** It doesn't show anything in-line. With the Chrome extension it's a little bit more limiting, because we essentially had to build on top of a pre-existing UI, and that's a little bit tougher than building our own application. So we do our best to show all the information we can there, but it's still not ... |
**Erik St. Martin:** The other thing I noticed you guys did was you released SourceLib, which is the library behind doing the code parsing. |
**Beyang Liu:** Yeah, so basically in order to extract the information that we used to display great code search results and usage snippets to the user, we needed to actually parse the code and extract the simple table and do other sorts of static analysis. We do that across a couple of different languages now. We star... |
**Erik St. Martin:** Have you seen anybody using that for any interesting projects? Everytime I see that I'm like, "I wanna build something with that", but I haven't come up with a fun use case. But I say the same about all static code analysis. I'm like, "I feel like I could do something useful here", I just don't kno... |
**Beyang Liu:** Yeah, actually the impetus for our editor plugins and integrations came from the community, from people building on top of that and saying, "Hey, this information is useful. I just wanna jam it in my editor, which is where I spend a lot of my time." So we kind of looked at that and said, "Hey, we can he... |
**Brian Ketelsen:** I played with the Vim extension - actually in NeoVim, I don't know if it makes a difference, but I played in the Vim extension, I guess it was last week or the week before, and I was blown away. I cannot believe how freakin' cool that was. |
**Beyang Liu:** Awesome, thank you. |
**Brian Ketelsen:** Yeah, I turned it off, because it was just a little bit slow, sorry... But I think all of that is just internet latency, but it was really impressive. |
**Beyang Liu:** Yeah. I mean, performance is a big thing that we're hammering on right now, and we're actually gonna ship a native desktop client soon. That should hopefully help with the latency things, so... Kind of a sneak peek at that. |
**Erik St. Martin:** \[07:57\] Yeah, that's gotta be hard too, like Brian said, with the network latency. It's much easier to do that stuff when all the code exists locally, than it is to do it remotely. |
**Beyang Liu:** Yeah, and what we'd like to do in the future is really connect the code that's on your local machine to this global graph that exists out there in the ether, the cloud, because that's really where a lot of the magic can happen. You have access to this dictionary and index over all the possible open sour... |
**Brian Ketelsen:** That's kind of crazy. Can you give us an idea of what the request flow looks like when I'm typing in Vim? What's happening behind the scenes? |
**Beyang Liu:** The architecture... We try to do things in ways that are as scalable as possible, in the sense that we wanna support a bunch of different editor plugins without a lot of deep custom work for each editor. So what each plugin does is essentially there's kind of like an API that Sourcegraph exposes specifi... |
For example, one of the things that the plugins do is anytime your cursor is hovering over a token in the code - let's say it's a function name - what it will do is pull up documentation and user examples for that function in Sourcegraph in your web browser. And what it does there is essentially in the editor it extrac... |
**Brian Ketelsen:** So the whole process is wickedly fast when you describe it that way. What kind of data storage are you using on the backend? I like the hardware stuff. What databases, what indexes are you using? |
**Beyang Liu:** We kind of have a custom graph storage system that's built on top of Postgres and Google Object Store right now. So it basically uses Google Object Store for all the detailed meta data. Once you know you're looking for data for a specific definition, specific function or package, then you look up the de... |
**Erik St. Martin:** \[12:16\] Did you have to build a spider to index all that data? |
**Beyang Liu:** Yeah, so the way our crawler works is anytime it indexes a library, it looks at the dependencies, because we actually extract the dependencies of that particular library, and then it just goes and indexes the dependencies. So it just kind of crawls the graph of code, the dependency graph. |
**Erik St. Martin:** Have you guys been doing anything with the new BigQuery dataset that went out? |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.