retitle 768732 plm: FTBFS in non-UTF-8 settings thanks Hello,
thanks for this additional information, that's indeed the source of the problem. The PLM package builds correctly if I set LC_ALL=C.UTF-8 in addition to your settings and fails without it. But, it seems to me that the package built with non-utf locale settings will not be functionnal for non-english speakers. At least that's what I understand from the msggrep warnings. Could you please point me to the part of the policy that is relevant to the locale settings of the builder? I fail to find anything relevant by myself, sorry. In the meanwhile, I will patch most of the java files to not use the UTF encoding. I'm just reluctant do so in some cases where it would not be an improvement. For example, it would request to write the name of "Gérald Oster" as "Gerald Oster". Not a big deal, but not nice. I guess that it would not be ok to set LC_ALL=C.UTF-8 in the rules file, right? Thanks for your patience and sorry for the complications, Mt. On Wed, Nov 12, 2014 at 01:34:04PM +0100, Lucas Nussbaum wrote: > Hi, > > On 12/11/14 at 11:27 +0100, Martin Quinson wrote: > > Hello, > > > > I must confess that I fail to reproduce your bug. > > > > it seems to me that the bug comes from a strange setting in LC_ALL. > > # locale > LANG= > LANGUAGE= > LC_CTYPE="POSIX" > LC_NUMERIC="POSIX" > LC_TIME="POSIX" > LC_COLLATE="POSIX" > LC_MONETARY="POSIX" > LC_MESSAGES="POSIX" > LC_PAPER="POSIX" > LC_NAME="POSIX" > LC_ADDRESS="POSIX" > LC_TELEPHONE="POSIX" > LC_MEASUREMENT="POSIX" > LC_IDENTIFICATION="POSIX" > LC_ALL= > > > before the java compiler, msggrep complains that ANSI_X3.4-1968 is not > > a valid value for UTF-encoded files. > > > > Did your chroot have any unconventional settings with that regard? > > No. Also, I did not notice any other locale-related failure while filing bugs. > > Lucas -- HARDWARE n.m.: partie de l'ordinateur qui reçoit les coups quand le software se plante.
signature.asc
Description: Digital signature
