On 2024-01-16, at 22:42:40 -0700, Nicholas D Steeves wrote: > Xiyue Deng <manphiz@gmail.com> writes: > > Jeremy Sowden <jeremy@azazel.net> writes: > >> The tests were failing when I ran either: > >> > [snip] > >> Eventually, I tracked the problem down to > >> locale settings in the environment. In the schroot chroot, `LANG` is > >> not set; in the cowbuilder chroot it is set to `C`; whereas in the > >> sbuild chroot, which inherits it from my environment in the host system, > >> it is `en_GB.UTF-8`. > >> > [snip] > >> > >> Not sure what the moral of this tale is yet, but in any case, the > >> problem is unrelated to your packaging (and I see the same behaviour > >> using the .orig tar-ball). > > Jeremy, thank you reviewing this sponsorship request. > > > > > Thanks very much for tracking this down! > > Yes, thank you. This confirms the hypothesis for part of the (old, from > 2020) bug reported upstream here: > > https://github.com/mhayashi1120/Emacs-wgrep/issues/78 > > > I have reproduced this in a newly created cowbuilder (in fact both > > multibyte and unibyte tests failed), and tested that setting LC_ALL to > > C.UTF-8 indeed fixes this issue. > > This action is only correct if all users will use UTF-8 100% of the time > because otherwise it sweeps a corner-case under the rug. In other > words, NACK at this time. > > Providing correct functioning to users of non-UTF-8 appears to require > this: > > https://www.gnu.org/software/emacs/manual/html_node/elisp/Converting-Representations.html > > and alternatively, the mode should throw an error on non-UTF-8 > systems. Agreed. Thanks for the pointers. I had intended to file a separate bug-report to keep track of the problem and continue looking into it, but got distracted by a couple of things. I'll get the bug-report done later this week. > As the package maintainer, I'm willing to sponsor at commit:fc1fac7 or > equivalent. J.
Attachment:
signature.asc
Description: PGP signature