[bug-libunistring] ‘test-localename’ failures

[Ludovic Courtès]

Hello,

This is just a kind reminder (no patch offered ;-)) that
‘test-localename’ has been failing on glibc 2.9 systems.

Thanks,
Ludo’.

address@hidden (Ludovic Courtès) writes:

> Hi Bruno,
>
> On x86_64-linux-gnu with glibc 2.9 ‘test-localename’ segfaults for me:
>
> $ ../libtool --mode=execute gdb ./test-localename
> GNU gdb (GDB) 7.0
> Copyright (C) 2009 Free Software Foundation, Inc.
> License GPLv3+: GNU GPL version 3 or later <http://gnu.org/licenses/gpl.html>
> This is free software: you are free to change and redistribute it.
> There is NO WARRANTY, to the extent permitted by law.  Type "show copying"
> and "show warranty" for details.
> This GDB was configured as "x86_64-unknown-linux-gnu".
> For bug reporting instructions, please see:
> <http://www.gnu.org/software/gdb/bugs/>...
> Reading symbols from 
> /home/ludo/src/nixpkgs/pkgs/libunistring-0.9.2/+build/tests/.libs/test-localename...done.
> (gdb) r
> Starting program: 
> /home/ludo/src/nixpkgs/pkgs/libunistring-0.9.2/+build/tests/.libs/test-localename
>  
>
> Program received signal SIGSEGV, Segmentation fault.
> 0x00007ffff77f0af0 in strcmp () from 
> /nix/store/s88vdfglm94x7jn0vqm24pqhq460s0c7-glibc-2.9/lib/libc.so.6
> (gdb) bt
> #0  0x00007ffff77f0af0 in strcmp () from 
> /nix/store/s88vdfglm94x7jn0vqm24pqhq460s0c7-glibc-2.9/lib/libc.so.6
> #1  0x0000000000400ffc in test_locale_name_thread () at 
> ../../tests/test-localename.c:421
> #2  0x0000000000402305 in main () at ../../tests/test-localename.c:734
> (gdb) frame 1
> #1  0x0000000000400ffc in test_locale_name_thread () at 
> ../../tests/test-localename.c:421
> 421               ASSERT (strcmp (unsaved_names[j][i], saved_names[j][i]) == 
> 0);
> (gdb) p j
> $1 = 76
> (gdb) p i
> $2 = 0
> (gdb) p unsaved_names[j][i]
> $3 = 0x0
> (gdb) p saved_names[j][i]
> $4 = 0x0
> (gdb) p sizeof(categories)/sizeof(categories[0])
> $6 = 12
> (gdb) p sizeof(choices)/sizeof(choices[0])
> $7 = 103
> (gdb) p choices[j]
> $8 = 0x402850 "no_NO.UTF-8"
> (gdb) p address@hidden
> $10 = {0x0, 0x7ffff7de79d7 "H\203\300\001H\215\025\234\025\001", 0x0, 0x0, 
> 0x0, 0x0, 0x0, 0x7ffff7deefb5 "\205\300u\347H\203\304\b\270\001", 0x0, 
> 0x7ffff7ffe528 "", 0x7ffff778c60d "ld-linux-x86-64.so.2", 
>   0x7ffff7de8804 "\205\300t\310H\201\304\310\003"}
> (gdb) p address@hidden
> $11 = {0x0 <repeats 12 times>}
> (gdb) p address@hidden
> $12 = {0x650350 "no_NO", 0x650370 "no_NO", 0x650390 "no_NO", 0x6503b0 
> "no_NO", 0x6503d0 "no_NO", 0x6503f0 "no_NO", 0x650410 "no_NO", 0x650430 
> "no_NO", 0x650450 "no_NO", 0x650470 "no_NO", 0x650490 "no_NO", 
>   0x6504b0 "no_NO"}
> (gdb) p address@hidden
> $13 = {0x650338 "no_NO" <repeats 12 times>}
>
> Could it be that there’s something fishy specifically with
> ‘no_NO.UTF-8’?  See:
>
> scheme@(guile-user)> (setlocale LC_ALL "no_NO.UTF-8")
> ERROR: In procedure setlocale:
> ERROR: Invalid argument
> scheme@(guile-user)> (setlocale LC_ALL "no_NO")
> $1 = "no_NO"
> scheme@(guile-user)> (setlocale LC_ALL "fr_FR.UTF-8")
> $2 = "fr_FR.UTF-8"
> scheme@(guile-user)> (setlocale LC_ALL "de_DE.UTF-8")
> $3 = "de_DE.UTF-8"
> scheme@(guile-user)> (setlocale LC_ALL "en_US.UTF-8")
> $4 = "en_US.UTF-8"
>
> Thanks,
> Ludo’.