[Monotone-devel] problems with i18n testsuite

[Martin Waitz]

hi :)

when running the monotone testsuite, i get two problems:
 * the netsync test doesn't terminate until i press ctrl-c
 * the i18n test failes.

i'm using monotone 0.11 on a debian sid system.
i installed the locales used by the testsuite and got the following error:

========================================================================
t_i18n_file.at:21: monotone --rcfile=test_hooks.lua --nostd --norc --db=test.db 
address@hidden add "tmp/file name with spaces"
stderr:
monotone: adding tmp/file name with spaces to working copy add set
stdout:
t_i18n_file.at:22: monotone --rcfile=test_hooks.lua --nostd --norc --db=test.db 
address@hidden add "tmp/address@hidden:"
stderr:
monotone: adding tmp/address@hidden: to working copy add set
stdout:
t_i18n_file.at:26: monotone --rcfile=test_hooks.lua --nostd --norc --db=test.db 
address@hidden add "tmp/$EUROPEAN_UTF8"
stderr:
******** errors detected; see standard output for details ********
stdout:

**** exception(205): std::logic_error: transforms.cc:421: invariant 
'I(converted != NULL)' violated

options path is MT/options
converting 2 bytes 'MT' from UTF-8 to ANSI_X3.4-1968
converting 7 bytes 'options' from UTF-8 to ANSI_X3.4-1968
converting 2 bytes 'MT' from UTF-8 to ANSI_X3.4-1968
converting 7 bytes 'options' from UTF-8 to ANSI_X3.4-1968
converting 6 bytes from UTF-8 to IDNA ACE
converting 4 bytes from UTF-8 to IDNA ACE
converting 3 bytes from UTF-8 to IDNA ACE
converting 8 bytes 'monotone' from ANSI_X3.4-1968 to UTF-8
converting 23 bytes '--rcfile=test_hooks.lua' from ANSI_X3.4-1968 to UTF-8
converting 7 bytes '--nostd' from ANSI_X3.4-1968 to UTF-8
converting 6 bytes '--norc' from ANSI_X3.4-1968 to UTF-8
converting 12 bytes '--db=test.db' from ANSI_X3.4-1968 to UTF-8
converting 21 bytes 'address@hidden' from ANSI_X3.4-1968 to UTF-8
converting 3 bytes 'add' from ANSI_X3.4-1968 to UTF-8
converting 12 bytes 'tmp/öäüß' from ANSI_X3.4-1968 to UTF-8
transforms.cc:421: invariant 'I(converted != NULL)' violated
t_i18n_file.at:26: exit code was 200, expected 0
24. t_i18n_file.at:3: 24. importing files with a internationalized names 
(t_i18n_file.at:3): FAILED (t_i18n_file.at:26)

========================================================================
(i added the source string to the debug output)

it seems that iconv could not convert a string.
the string to be converted was c3 b6 c3 a4 c3 bc c3 9f,
which already is utf-8.

why is it trying to convert from/to ANSI_X3.4-1968 (which seems to be
plain ascii) here?
(my personal locale is de_DE.UTF-8)


by the way: shouldn't filenames always be encoded in utf-8?
at least many file managers seem to use utf-8 file names.
that way it is still possible for the user to find his files
when he switched his locale (or via remote access)

-- 
Martin Waitz

signature.asc
Description: Digital signature