-
Notifications
You must be signed in to change notification settings - Fork 13.3k
get rid of tidy 'unnecessarily ignored' warnings #98717
New issue
Have a question about this project? # for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “#”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? # to your account
Conversation
r? @jyn514 (rust-highfive has picked a reviewer for you, use r? to override) |
I agree warning about line length is unhelpful and just leads to unnecessary CI failures. I am less sure about the others - I think having |
This comment has been minimized.
This comment has been minimized.
Fair, I have not run into the others. |
21998e6
to
8515475
Compare
r=me with CI passing |
@bors r=jyn514 |
📌 Commit 8515475 has been approved by |
…=jyn514 get rid of tidy 'unnecessarily ignored' warnings I think these warnings are quite pointless: when I say `allow(foo)` in my code, that doesn't necessarily mean that I expect `foo` to happen -- it just means that I am okay with `foo` happening. For example, having to add and remove `ignore-tidy-linelength` as the longest line in the file keeps growing and shrinking is just annoying and doesn't benefit anyone, IMO. This usually incurs *two* CI roundtrips: first CI tells you that line lengths in your test file are ignored unnecessarily, so you go and remove that attribute; then CI tells you that now your line numbers changed, so you re-bless your tests (often takes >5min if parts of rustc need rebuilding because `./x.py fmt` changed something somewhere). That's just a lot of wasted effort and time and patience.
…=jyn514 get rid of tidy 'unnecessarily ignored' warnings I think these warnings are quite pointless: when I say `allow(foo)` in my code, that doesn't necessarily mean that I expect `foo` to happen -- it just means that I am okay with `foo` happening. For example, having to add and remove `ignore-tidy-linelength` as the longest line in the file keeps growing and shrinking is just annoying and doesn't benefit anyone, IMO. This usually incurs *two* CI roundtrips: first CI tells you that line lengths in your test file are ignored unnecessarily, so you go and remove that attribute; then CI tells you that now your line numbers changed, so you re-bless your tests (often takes >5min if parts of rustc need rebuilding because `./x.py fmt` changed something somewhere). That's just a lot of wasted effort and time and patience.
…askrgr Rollup of 10 pull requests Successful merges: - rust-lang#97629 ([core] add `Exclusive` to sync) - rust-lang#98503 (fix data race in thread::scope) - rust-lang#98670 (llvm-wrapper: adapt for LLVMConstExtractValue removal) - rust-lang#98671 (Fix source sidebar bugs) - rust-lang#98677 (For diagnostic information of Boolean, remind it as use the type: 'bool') - rust-lang#98684 (add test for 72793) - rust-lang#98688 (interpret: add From<&MplaceTy> for PlaceTy) - rust-lang#98695 (use "or pattern") - rust-lang#98709 (Remove unneeded methods declaration for old web browsers) - rust-lang#98717 (get rid of tidy 'unnecessarily ignored' warnings) Failed merges: r? `@ghost` `@rustbot` modify labels: rollup
I think these warnings are quite pointless: when I say
allow(foo)
in my code, that doesn't necessarily mean that I expectfoo
to happen -- it just means that I am okay withfoo
happening.For example, having to add and remove
ignore-tidy-linelength
as the longest line in the file keeps growing and shrinking is just annoying and doesn't benefit anyone, IMO. This usually incurs two CI roundtrips: first CI tells you that line lengths in your test file are ignored unnecessarily, so you go and remove that attribute; then CI tells you that now your line numbers changed, so you re-bless your tests (often takes >5min if parts of rustc need rebuilding because./x.py fmt
changed something somewhere). That's just a lot of wasted effort and time and patience.