| Message ID | 20230801-nolibc-warnings-v2-0-1ba5ca57bd9b@weissschuh.net (mailing list archive) |
|---|---|
| Headers | show |
| Series | tools/nolibc: enable compiler warnings | expand |
Hi Thomas, On Tue, Aug 01, 2023 at 07:30:07AM +0200, Thomas Weißschuh wrote: > To help the developers to avoid mistakes and keep the code smaller let's > enable compiler warnings. (...) OK, that's overall OK to me, and I noted thta you'll likely update 6 and 9. Just let me know how you prefer to proceed, I hope to send the PR to Shuah this week-end so I want to be sure we avoid a last minute rush and the risks of breakage that comes with it. Thanks! Willy
On 2023-08-02 22:22:43+0200, Willy Tarreau wrote: > Hi Thomas, > > On Tue, Aug 01, 2023 at 07:30:07AM +0200, Thomas Weißschuh wrote: > > To help the developers to avoid mistakes and keep the code smaller let's > > enable compiler warnings. > (...) > > OK, that's overall OK to me, and I noted thta you'll likely update > 6 and 9. Just let me know how you prefer to proceed, I hope to send > the PR to Shuah this week-end so I want to be sure we avoid a last > minute rush and the risks of breakage that comes with it. I'm not yet sure when I can rework those patches. Could you already pick up those you are fine with? (And also leave out the final one to not have spurious warnings) Thomas
On Wed, Aug 02, 2023 at 11:10:17PM +0200, Thomas Weißschuh wrote: > On 2023-08-02 22:22:43+0200, Willy Tarreau wrote: > > Hi Thomas, > > > > On Tue, Aug 01, 2023 at 07:30:07AM +0200, Thomas Weißschuh wrote: > > > To help the developers to avoid mistakes and keep the code smaller let's > > > enable compiler warnings. > > (...) > > > > OK, that's overall OK to me, and I noted thta you'll likely update > > 6 and 9. Just let me know how you prefer to proceed, I hope to send > > the PR to Shuah this week-end so I want to be sure we avoid a last > > minute rush and the risks of breakage that comes with it. > > I'm not yet sure when I can rework those patches. No problem. > Could you already pick up those you are fine with? > (And also leave out the final one to not have spurious warnings) OK, I'll stick to this for now, thank you! Willy
To help the developers to avoid mistakes and keep the code smaller let's enable compiler warnings. I stuck with __attribute__((unused)) over __maybe_unused in nolibc-test.c for consistency with nolibc proper. If we want to add a define it needs to be added twice once for nolibc proper and once for nolibc-test otherwise libc-test wouldn't build anymore. Signed-off-by: Thomas Weißschuh <linux@weissschuh.net> --- Changes in v2: - Don't drop unused test helpers, mark them as __attribute__((unused)) - Make some function in nolibc-test static - Also handle -W and -Wextra - Link to v1: https://lore.kernel.org/r/20230731-nolibc-warnings-v1-0-74973d2a52d7@weissschuh.net --- Thomas Weißschuh (10): tools/nolibc: drop unused variables tools/nolibc: sys: avoid implicit sign cast tools/nolibc: stdint: use int for size_t on 32bit selftests/nolibc: drop unused variables selftests/nolibc: mark test helpers as potentially unused selftests/nolibc: make functions static if possible selftests/nolibc: avoid unused arguments warnings selftests/nolibc: avoid sign-compare warnings selftests/nolibc: test return value of read() in test_vfprintf selftests/nolibc: enable compiler warnings tools/include/nolibc/stdint.h | 4 + tools/include/nolibc/sys.h | 3 +- tools/testing/selftests/nolibc/Makefile | 2 +- tools/testing/selftests/nolibc/nolibc-test.c | 108 +++++++++++++++++---------- 4 files changed, 74 insertions(+), 43 deletions(-) --- base-commit: dfef4fc45d5713eb23d87f0863aff9c33bd4bfaf change-id: 20230731-nolibc-warnings-c6e47284ac03 Best regards,