Ben Hutchings wrote: > On Thu, 2007-04-26 at 21:21 +0200, Brice Goglin wrote: > >> Seems to works fine here: >> $ LC_CTYPE=fr_FR ./prout '-*-fixed-*-*-*-*-13-*-*-*-*-*-*-*' >> default: >> $ LC_CTYPE=fr_FR.utf8 ./prout '-*-fixed-*-*-*-*-13-*-*-*-*-*-*-*' >> missing: KSC5601.1987-0 >> missing: GB2312.1980-0 >> default: >> (the other cases get either the first or the second output from above). >> > > Was the fr_FR.utf8 locale actually installed on the test system? I > found that the test passes for UTF-8 locales that aren't actually > installed. >
Yes (I am french actually): $ locale -a C POSIX en_US en_US.iso88591 en_US.iso885915 en_US.utf8 fr_FR fr_FR.iso88591 [EMAIL PROTECTED] fr_FR.utf8 [EMAIL PROTECTED] français french > I don't know where to look. strace shows no system call failures, but > it can dump the X protocol conversation (-e read=3 -e write=3). Would > that be useful? > I was rather thinking of locating where this assertion is (I don't see it in libx11, strace/gdb might help to locate the library or binary where it comes from). And then try to understand where the error comes from by either reading the corresponding code, or adding some debug printf. Not very easy... Brice
