mozilla :: #rust-infra

14 May 2017
03:22simulIf someone could investigate that would be great, I can't view log right now (on phone)
03:22simulI think we wanted it to land tonight; I'll retry blindly if we can't investigate.
03:25simulSomeone commented, seems to be related, will spreadsheet and tag later. Retried.
03:28nagisaI did say some hour ago that CDN is crappingout
03:28nagisawrong channel turns out
03:28nagisaI see CDN is working its wonders again today
03:28nagisa> curl: (92) HTTP/2 stream 1 was not closed cleanly: INTERNAL_ERROR (err 2)
10:08theduke_if anyone is interested: API client for
10:09theduke_supports fetching the summary, paginated crate search, and a convenience method for getting all data for a crate including reverse dependencies, version info, ...
10:09theduke_feedback welcome ;)
13:25polybuildrHi, I wanted to know, what kind of tests are run for rustc that involve some rust files being typechecked/run? And how exactly are they run? I assume they don't fit into the traditional #[test] paradigm?
13:26polybuildrI see run-pass and run-fail in, but can someone give me more details about the process of running these?
13:54simulacrumpolybuildr: compiletest is a custom tool that is written for rust to do this
13:54polybuildrsimulacrum: I see. Could you point me to it please?
13:54simulacrumThere's a crate on with the same name, and you can look at the source code in src/tools/compiletest
13:54polybuildrI'm writing an interpreter, and I want to build a similar testing system, hence the curiosity.
13:55polybuildrAh, thank you! :D
14:05polybuildrest31: ah, thanks!
16:25simulacrumPossibly have a new spurious failure:, from
16:50frewsxcvtomprince brson: either of you know if these logs should be more informative ?
16:50frewsxcvfor example, here's a 'test-fail'
16:50frewsxcvfigured i'd see errors
17:38simulacrumFlaky internet
17:38acrichtothat's my guess at least
17:58simulacrumacrichto: Is there a chance you could help me out here:
17:58simulacrumI don't know how to make the cargo submodule dance
18:03acrichtosimulacrum: sure yeah, I'll comment
18:10simulacrumacrichto: I'm not sure I follow your comment -- we need to update Cargo's tests for the new rustc since they test output that changes
18:11simulacrumThat is, the PR for rust-lang/rust changes output, which makes the Cargo tests fail; but I don't think we can land a PR in Cargo since then it's tests will fail until the rust PR lands
18:24acrichtosimulacrum: right but the question wasn't really about that?
18:24acrichtoit was about just getting cargo to build at all
18:24acrichtofor the tests we'd just neex to relax the assertions
18:24simulacrumacrichto: Ah, so there's no magical way to do the tests. I'll leave a comment about the relaxation
20:08tomprincefrewsxcv: I wonder if that is a regression from switching to slog.
20:10tomprincefrewsxcv: Thinking about it, I bet it is, since running commands does some crazy stuff with threads, and doesn't propagate the logging context.
22:40frewsxcvtomprince: hmm
15 May 2017
No messages
Last message: 10 days and 8 hours ago