tech-toolchain archive

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index][Old Index]

Re: Compiling with modern Clang



In article <CAAc3VoMiXzcOR-z4ti6N7-nTU=xJTNm09fv4teMy+y-mgLBdow%mail.gmail.com@localhost>,
FireTurtle  <pyhirricturtle%gmail.com@localhost> wrote:
>-=-=-=-=-=-
>
>Hi,
>
>I've been building the NetBSD-current tree with a recent Clang (22+) for a
>project and there are quite a few -Werror failures from warnings that GCC
>doesn't emit.
>
>My thought is to suppress the high-volume warnings globally and submit
>code fixes for the rest.
>
>For the high-volume suppressions, I'd like to propose adding these three to
>bsd.sys.mk under CWARNFLAGS.clang:
>
> *  -Wno-error=deprecated-non-prototype    (~100+ instances,  distributed.)
> *  -Wno-error=strict-prototypes           (~50+ instances, same packages)
>  * -Wno-error=unused-but-set-variable     (~50+ instances)
>
>These are all C23 intolerance or Clang-vs-GCC divergences that fire across
>multiple architectures (amd64, i386,  ...).   At this volume, per-file
>suppression is impractical.
>
>I'd also include -Wno-unknown-warning-option as a safety net so the same
>bsd.sys.mk works with both the in-tree Clang 13 and newer versions.
>
>Thoughts?
>
>Also is there any thought to upgrading the Clang version -- Clang 13 is 4.5
>years old.?

I hope so. Joerg has been the keeper of clang, he should comment.

Best,

christos



Home | Main Index | Thread Index | Old Index