mozilla :: #rust-infra

11 Sep 2017
05:30eddybwhoa whoa
05:30eddybdoes cargobomb have a 2-minute timeout?
05:34eddybaidanhs, tomprince: I am confused, https://cargobomb-reports.s3.amazonaws.com/pr-44211/index.html hit hashconsing-0.3.0 but that has been blacklisted for a while?
09:19sebkwho can restart playbot to use the latest code?
09:41est31shep can
09:41est31maybe also carols10cents
09:55sebkshep: I (hopefully) fixed the error pasting of playbot a few days ago. the PR landed, but it needs a restart
11:42aidanhseddyb: the template message for a completed cargobomb run should be clarified - the blacklist is just "you can disregard these", it doesn't actually have any effect on reported results
11:43aidanhseddyb: and I don't think so, I believe it's more like 10min. though for some crates it doesn't take effect for some reason and they have to be killed manually
11:44eddybaidanhs: I'm seeing 2 minutes consistently
11:44eddybaidanhs: between the last printed line and the ending. and it's a failure
11:44eddybwith no explicit errors
11:44eddybvery confusing
11:45aidanhshttps://github.com/rust-lang-nursery/cargobomb/blob/1621ee5ce16d45fa33df30ad890a4842952e5b87/src/run.rs#L97-L98
11:45aidanhsso there is a 2min heartbeat timeout if you don't generate output
11:45aidanhsbut the overall timeout is 20min
12:07est31aidanhs: imo the blacklist should be in toml format, and cargobomb should parse it
12:17aidanhsest31: I agree
12:17aidanhsthere is an issue for that
12:18aidanhsthere's a lot of low hanging fruit for cargobomb, the impl period is going to be great
12:53simulacrumaidanhs, not sure if it was you, but this is interesting: https://gist.github.com/Mark-Simulacrum/b2eaa93de1d5ebd760095eda4377314f
12:54simulacrumfrom https://travis-ci.org/rust-lang/rust/jobs/274108980
14:28aidanhssimulacrum: the command appears to have been run four times in that logfile
14:28aidanhswhich is bizarre, because it should only be run twice
14:29aidanhsno, it was run three times
14:30aidanhswhich is hardly better
14:35aidanhshahaha, I think it was doing two builds on the same builder!
14:35aidanhshttps://travis-ci.org/rust-lang/rust/builds/274108951
14:35aidanhsdist-x86_64-freebsd took 3 hours
14:39aidanhsgoing to send an email to travis
14:43plietarCould bors (or something else) manage the `S-waiting-on-` labels semi-automatically ?
14:46steveklabnikif people used github's reviews feature, i bet it could
14:46steveklabnikotherwise it'd get dicy
14:51aidanhsplietar: any particular motivation? we generally just use it to make sure PRs don't get lost and automating it isn't necessary (though could be useful in some cases)
14:52plietaraidanhs: my main "issue" is that GH doesn't let non team members change labels
14:53plietarAs a PR author, once I've addressed reviews I can't update the label from S-waiting-on-author to S-waiting-on-review
14:54plietarMy idea was have a '@bors review' command which the PR author can use to have bors update the labels
14:55plietarMaybe this is more suited for highfive
14:56aidanhsnote that the only reason to do that is for tidiness - it won't necessarily get your PR reviewed faster since the infra team are the only ones to really use those labels
14:58aidanhsif did want to pressure reviewers to reply more quickly then there's no reason we couldn't change the process, but the current timings don't seem too bad most of the time
19:41acrichtosimulacrum: r? https://github.com/rust-lang/rust/pull/44498
21:05nmatsakiscan someone add estebank to the compiler-team mailing list alias?
21:05nmatsakisedunham: acrichto ^?
21:15acrichtonmatsakis: full email?
21:15acrichtoor rather, preferred email?
21:19nmatsakisacrichto: sent via privmsg
21:19acrichtonmatsakis: done
21:19nmatsakisthanks
21:19nmatsakisoops
21:19nmatsakisI just sent an -email that they probably shoudl ahve gotten :)
21:20nmatsakisI wil fwd to them
12 Sep 2017
No messages
   
Last message: 9 days and 6 hours ago