Uh oh!
There was an error while loading. Please reload this page.
Remove (lots of) dead code - #83185
Conversation
rust-highfive
commented
Mar 16, 2021
r? @oli-obk (rust-highfive has picked a reviewer for you, use r? to override) |
This comment has been minimized.
This comment has been minimized.
86e1dd2 to
08e596aCompareUh oh!
There was an error while loading. Please reload this page.
jyn514
commented
Mar 16, 2021
@bors try @rust-timer queue I expect this to help hello-world, but not much else. |
rust-timer
commented
Mar 16, 2021
Awaiting bors try build completion. @rustbot label: +S-waiting-on-perf |
bors
commented
Mar 16, 2021
⌛ Trying commit 08e596a84bea96a0d8c3799c36724c012bb9154b with merge cee392aca3bbc45102766189904bd3c29f498f19... |
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
3dd122c to
aa09c5bCompare
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
thomcc
commented
Mar 16, 2021
Hmm, it seems likely that the resolution of some of our floating point bugs will require more architecture-awareness of floats, rather than less, but I guess if we aren't using it we could just resurrect the code from git history if that time comes... |
oli-obk
commented
Mar 16, 2021
Please don't touch rustc_apfloat without conferring with @eddyb . They manually transpiled llvm's apfloat to Rust and this is the result. We should not touch that library afaik. |
eddyb
commented
Mar 16, 2021
So the whole deal with It shouldn't even be in-tree IMO, see also #55993 - people have wanted to use it elsewhere, too. But I haven't been able to do as much as I want with it because it's been stuck in licensing limbo for years. I don't dare to touch it at all, really, while the licensing situation is up in the air, in the fear that it will prolong the limbo (originally I was assured there were no licensing issues and whoever told me that was wrong, so I don't want that to happen again) and I'd strongly advise anyone else against messing with it. (Can we add a rule somewhere that makes Once the licensing situation is sorted out, and we can move it out of tree, publish on crates.io, etc. I want to take this "two main branches" approach, to cleanly separate what's "tracking LLVM upstream" and what's our own patches: #55993 (comment) |
Uh oh!
There was an error while loading. Please reload this page.
Uh oh!
There was an error while loading. Please reload this page.
Uh oh!
There was an error while loading. Please reload this page.
Uh oh!
There was an error while loading. Please reload this page.
Uh oh!
There was an error while loading. Please reload this page.
jyn514
commented
Mar 28, 2021
I added back all the functions @oli-obk asked I keep, this should be ready for another round of review. |
Except for the sync stuff, everything lgtm. I have no strong opinion, but since I have good hopes that the work on parallel compiler will continue this year, I'm worried this removes things we'll just need to add again. I don't know who to ping about it right now though |
jyn514
commented
Mar 29, 2021
It looks like cc @Zoxc - do these changes to |
There isn't currently a good reviewer for these, and I don't want to remove things that will just be added again. I plan to make a separate PR for these changes so the rest of the cleanup can land.
jyn514
commented
Mar 29, 2021
I reverted the sync changes, I'll make a separate PR for them so they don't hold up everything else. @bors r=oli-obk |
bors
commented
Mar 29, 2021
📌 Commit 526bb10 has been approved by |
bors
commented
Mar 29, 2021
bors
commented
Mar 29, 2021
☀️ Test successful - checks-actions |
Remove dead or useless code from Session This is a more principled follow-up to rust-lang#83185 (comment). - Rename `Parser::span_fatal_err` -> `Parser::span_err` - Remove some unnecessary uses of `struct_span_fatal` - Make `Diagnostic::span_fatal` unconditionally raise an error - Add `impl Deref<Target = Handler>` for Session and remove all functions that are exactly the same as their Handler counterparts - Note why `Handler::fatal` is different from `Sesssion::fatal` - Remove unused `opt_span_warn` function r? `@oli-obk` or `@estebank`
Remove unused code from `rustc_data_structures::sync` Found using https://github.com/est31/warnalyzer. Follow-up to rust-lang#83185. r? `@Zoxc` cc `@oli-obk`
Make diagnostics clearer for `?` operators Re-submission of rust-lang#75029, fixesrust-lang#71309 This also revives the `content` methods removed by rust-lang#83185. r? `@estebank`
Builds on
DiagnosticBuilder::subfunction #83161forward_inner_docshack #83230Found with https://github.com/est31/warnalyzer.
See #77739 for a similar change in the past.
Dubious changes:
TODO:
cfg(test)changes into Move some test-only code to test files #83197cc @est31