Give permission to do try builds to wg-triage

[Kobzol]

Discussed in https://rust-lang.zulipchat.com/#narrow/channel/122651-general/topic/.E2.9C.94.20start.20a.20try-job.20on.20my.20PR.20please/with/541811337.

[github-actions]

Dry-run check results

[WARN  sync_team] sync-team is running in dry mode, no changes will be applied.
[INFO  sync_team] synchronizing github

[jieyouxu]

I'm hesitant on this because WG-triage membership has bascially no sustaining-contribution requirements (i.e. anyone could request to join). WDYT?

cc @Dylan-DPC

[Kobzol]

Hmm. Maybe we could have two teams then, or different roles within the team?

[jieyouxu]

At least to me, it seems to make sense to either have two teams or roles. One for the "low bar" that just makes it easy to adjust labels via UI, and another for members with sustained contributions that at least I'd feel more comfortable providing try permissions to.

[Noratrieb]

I don't think "triage" needs try permissions at all. The original triggering PR was a normal compiler contribution PR, that the person happened to be on the triage wg is more of a coincidence than indicative of a permission problem with the triage WG.

If they make regular contributions where it would be useful to have try perms, they should be added to the compiler team.

[Kobzol]

My idea was that if you want to test if some issue happens, try builds can be a useful tool for that, especially if it's on an arch/OS that you don't have access to. But this probably doesn't come up often enough..

[davidtwco]

I'd like to see more examples of cases where this would have been useful

[Enselic]

I was admittedly supporting giving wg-triage try rights mostly so I wouldn't have to keep asking for it myself on a case by case basis, which is inconvenient both for me and for others.

After reading this it seems like adding me as a Compiler Team Member is a reasonable way to give me try permission (and more). But I understand if I do not meet the qualifications for such membership.

Here are examples of when permissions had to be explicitly delegated to me (both r+ and try), stretching 3 years back:

• Fix backtraces with -C panic=abort on linux; emit unwind tables by default rust#143613 (comment)
• tests: Remove ignore-android directive for fixed issue rust#147100 (comment)
• ci: Increase rust.debuginfo-level-tests to 2 in x86_64-gnu-debug job rust#146394 (comment)
• ci: Begin running ui tests with rust.debuginfo-level-tests=1 rust#144499 (comment)
• tests: Test line number in debuginfo for diverging function calls rust#144034 (comment)
• tests: Don't check for self-printed output in std-backtrace.rs test rust#143401 (comment)
• tests: Require run-fail ui tests to have an exit code (SIGABRT not ok) rust#143002 (comment)
• tests: Add RUST_BACKTRACE and -Cpanic revisions to panic-main.rs test rust#142304 (comment)
• tests: Add recursive associated type bound regression tests rust#133467 (comment)
• inject_panic_runtime(): Avoid double negation for 'any non rlib' rust#133300 (comment)
• compiletest: Introduce remove_and_create_dir_all() helper rust#122786 (comment)
• unix_sigpipe: Add test for SIGPIPE disposition in child processes rust#121573 (comment)
• rustc_codegen_ssa: Remove trailing spaces in Display impl for CguReuse rust#118945 (comment)
• rustc_span: Use correct edit distance start length for suggestions rust#118381 (comment)
• large_assignments: Lint on specific large args passed to functions rust#116520 (comment)
• rustc_monomorphize: Introduce check_fn_args_move_size() rust#116310 (comment)
• Rename tests/ui/issues/issue-100605.rs to ../type/option-ref-advice.rs rust#114143 (comment)
• Support #[unix_sigpipe = "inherit|sig_dfl"] on fn main() to prevent ignoring SIGPIPE rust#97802 (comment)

To give a sense of what my contribution history looks like, it's probably best to browse this list, stretching 3.5 years back. Some examples:

• Fix backtraces with -C panic=abort on linux; emit unwind tables by default rust#143613 (likely merged soon)
• tests: Require run-fail ui tests to have an exit code (SIGABRT not ok) rust#143002
• core: Make Debug impl of raw pointers print metadata if present rust#135080
• Change SIGPIPE ui from #[unix_sigpipe = "..."] to -Zon-broken-pipe=... rust#124480
• large_assignments: Lint on specific large args passed to functions rust#116520
• Make unconditional_recursion warning detect recursive drops rust#113902
• Support #[unix_sigpipe = "inherit|sig_dfl"] on fn main() to prevent ignoring SIGPIPE rust#97802

[Kobzol]

It seems like there's a better way to handle this than giving try permissions to wg-triage. Closing.