On 2020-08-25 10:39, 'Reimundo Heluani' via sage-devel wrote: >> > That locale seems to be an invention of Debian that Ubuntu started using. > This > was closed as wontfix 8 years ago on Arch. > > https://bugs.archlinux.org/task/32296 > > It seems to me that having these test failing and flooding the logs of a > regular build is not the way to go. Is there a trac ticket about this? I > could > only find my comments on #30002 which was already closed.
Sorry, I let this slip in because my distribution is using a patched glibc and tricked me into thinking it was a standard locale. We should not require a huge unofficial patch to something as critical as glibc. A fix will appear eventually on, https://trac.sagemath.org/ticket/30053 It's still not totally clear to me what the various locale manglings achieve. It would be really helpful to know what goes wrong with, * No mangling at all * LC_ALL=C and where it goes wrong. Using the C locale is supposed to help with character-range regular expressions in spkg-install scripts, I guess. But the UTF-8 helps python... somehow? Neither #15791 nor #29033 go into detail. Note that many people "compile out" the locales that they don't intend to use, so using something like en_US.utf8 unconditionally is also not a good idea. The POSIX/C locale should always be there, though. -- You received this message because you are subscribed to the Google Groups "sage-devel" group. To unsubscribe from this group and stop receiving emails from it, send an email to sage-devel+unsubscr...@googlegroups.com. To view this discussion on the web visit https://groups.google.com/d/msgid/sage-devel/29e86f95-7d45-bb7e-bbed-84fc1341440d%40orlitzky.com.
