[libc-commits] [libc] [reland][libc] Remove unnecessary `FPBits` functions and properties (PR #79128)

Guillaume Chatelet via libc-commits libc-commits at lists.llvm.org
Wed Jan 24 02:36:19 PST 2024


gchatelet wrote:

> hmm...quite a few follow ups to this commit. The 32b arm build bots do take some time to provide feedback post submit, but it's important to wait to ensure they go back to green or don't go red in the first place. If it's near the end of your workday, I suggest holding off on merging commits until the next workday morning (especially Friday afternoons). That gives plenty of time to follow up on build breakages that are found only through post submits.
> 
> My hope is that we can make the enable runtimes build more ubiquitous so that testing via cross compiling is so fast/easy, it's the default development workflow.

So the issue does not come from the fix forward nor from this patch. [It comes from the original patch](https://github.com/llvm/llvm-project/pull/79113/files#diff-13e25aa87ca7cffde894d512e1467657ee30e44e8aa530daafdbd37276aec38aL28) that was reviewed.

Unfortunately a change to `libc/src/__support/common.h` slipped through and it was not caught by the build bots.
I do monitor the build bots after pushing a patch and revert immediately when something is wrong (this reland stems from this process).

In my opinion there are several things we can do to improve the situation:
 - Have a test covering the fact that we do indeed ship the `C` functions,
 - Have most of the build bots as part of the precommit check in github,
 - Have more build bots covering more architectures (GPU, Fuchsia)
 - Having faster cross-compilation would definitely help on the latency side of things.

I had no signal that it was broken but my apologies for the breakage nonetheless.

https://github.com/llvm/llvm-project/pull/79128


More information about the libc-commits mailing list