mozilla :: #rust-infra

13 May 2017
03:48est31nice, it managed to update sscache
04:14simulacrumYeah, after like 12 hours :)
04:14simulacrumEh, it's Friday, so we're fine. Weekend will be unlikely to create many new PRs, usually
04:16* est31 sits down and writes 20 PRs
13:19simulacrumest31: By the way, you should have retry rights
13:50est31do I need 2fA for that?
13:56simulacrumest31: Well, I don't know, but you're in the template, so "no"?
13:56simulacrumI don't think bors checks, but I could be wrong
13:59simulacrumRollup merged!
13:59simulacrum\o/
14:00est31oh, seems I'm mentioned
14:00est31cool
14:03est31thanks simulacrum for the info
14:49frewsxcvcongrats simulacrum
17:26simulacrumfrewsxcv: I think we'll wait till it fails and if same problem r- that PR and take it out of rollup
17:33frewsxcvroger
17:56simulacrumr- rollup per r- of a PR it contains
17:56simulacrumcreating another rollup
18:58thedukeHey there, is this the right place to ask about crates.io? I have a question regarding server load / crawling.
18:59thedukeI'm writing a crates.io API client, and I'm adding a full_crate() utility method to retrieve a crate including downloads, owners, and all versions (including dependencies and authors)
18:59thedukenow with a crate with many releases this can result in quite a lot of requests
19:00thedukeIt's a synchronous client, so the request frequency really isn't so bad, and I assume that those responses are cached anyway
19:00thedukeBut should I add some kind of delay between requests?
19:00simulacrumhmm
19:01simulacrumcc acrichto, carols10cents
19:01simulacrumI don't really know
19:01simulacrumI don't think we currently rate limit anything... I'd say "it's our fault" to be honest if it makes something break
19:02thedukeI'm also asking because I plan on fetching all data for all crates because I want to do some data analysis and maybe do some nice graphs. And I'd rather not get blamed for overloading the servers. :D
19:03simulacrumThat's done fairly often by us, and I think a few others, so I'd say go ahead
19:03simulacrumWe regularly do that sort of thing, really, for crater/cargobomb
19:04thedukeIt might be cool to generate a big JSON dump of the database every 24 hours and make it available for download
19:04thedukefor this sort of thing
19:53frewsxcvtomprince: what OS do you primarily use?
20:00tomprincefrewsxcv: linux, pretty much exclusively.
20:02frewsxcvtomprince: gotcha. have you tried running cargobomb?
20:03tomprinceI've run it with the demo dataset.
20:03frewsxcvjust curious if you're familiar with the workflow
20:03frewsxcvi'm not too familiar
20:03tomprinceJust following the (original) commands in readme.
20:04tomprinceI need to hack docker.rs, since I'm running fedora w selinux (so that the containers can read the shared folders)
20:04tomprinceAnd hello-rs isn't picked up by the list building.
20:04frewsxcvi don't understand why define-ex takes two params
20:08tomprincecargobomb is designed for comparing two toolchains to detect regressions.
20:10tomprinceIs that what your asking about?
20:34frewsxcvyeah i guess
20:34frewsxcvi need to make it somehow more generic
20:34frewsxcvwhere i just want to build everything on one toolchain
20:34frewsxcvi don't really have a need to specify two toolchains
21:01frewsxcvi like slog
21:11tomprinceI haven't used it much, but it seems like a sensible thing to have.
21:33carols10centstheduke: please load test our servers for us and let me know if you see any slowdown :) (that is, go ahead, should be fine)
21:34carols10centstheduke: also we want to provide a database snapshot someday but we have to actually set up sanitizing, automation, decide where to put it, etc and we haven't had time to do the work yet
21:34carols10centstheduke: but subscribe to https://github.com/rust-lang/crates.io/issues/630 if you want to know when we do!
21:38frewsxcvtomprince: oh, you mean: thread 'main' panicked at 'assertion failed: `(left == right)` (left: `1`, right: `2`)', src/ex.rs:92
21:38frewsxcvregarding hello-rs
21:39tomprincefrewsxcv: Yep.
21:39tomprinceI've just worked around it by manually adding it to the lists, when testing. But a more principled fixed would be better.
21:39frewsxcvi just did --crate-select top-100
21:58tomprincefrewsxcv: I think a treatment like Toilchain got might be useful for ex.rs
21:58frewsxcvhm?
21:59frewsxcvtreatment?
22:05tomprinceTurning free functions on strings and other fundamental types into functions on a potentially opaque structure.
22:06tomprinceOr structures, in the case of ex/ex_run
22:07frewsxcvoh yeah, i agree
14 May 2017
No messages
   
Last message: 129 days and 20 hours ago