libc: use separate collate objects for C, POSIX, and C.UTF-8
ClosedPublic
Actions

Authored by yuripv on May 6 2021, 10:00 AM.

Details

Reviewers

bapt
markj

Commits

rGa4916232acd6: libc: use separate collate objects for C, POSIX, and C.UTF-8
rG269dea90d6eb: libc: use separate collate objects for C, POSIX, and C.UTF-8

Summary

This fixes newlocale overwriting the locale name in collate object
when same instance was used for those locales, and querylocale
reporting unexpected value for LC_COLLATE_MASK.

PR: 255646, 269375

Test Plan

See added test case.

Run other libc tests.

Diff Detail

Repository

rG FreeBSD src repository

Lint

Lint Not Applicable

Unit

Tests Not Applicable

Event Timeline

yuripv requested review of this revision.May 6 2021, 10:00 AM

yuripv created this revision.

bapt accepted this revision.May 6 2021, 11:42 AM

This revision is now accepted and ready to land.May 6 2021, 11:42 AM

markj accepted this revision.May 6 2021, 1:07 PM

markj added inline comments.

lib/libc/tests/locale/newlocale_test.c
3	This should have an SPDX identifier I think.

yuripv marked an inline comment as done.May 6 2021, 5:38 PM

yuri_aetern.org mentioned this in D39294: use separate object for C.UTF-8's collate part.Mar 28 2023, 2:00 PM

update the test case to make it cover all switches between these locales
add the spdx identifier

This revision now requires review to proceed.Mar 28 2023, 2:53 PM

Herald added a subscriber: imp. · View Herald TranscriptMar 28 2023, 2:53 PM

Sorry, I completely forgot about this review, and now another PR was filed for the same problem. Could you please take another look? (I have updated the test case a bit for more complete switch testing).

markj accepted this revision.Mar 28 2023, 2:56 PM

markj added inline comments.