I have been bothered for a while now by the failure of the cygwin
binaries to do proper charset translation from the UTF-8 encoding used
in the message catalogs to my locale --- cp1252/iso-8859-1 ---, rather
I always get corrupt strings such as:
"""
La mayor?\195?\173a de los subcomandos reciben par?\195?\161metros de
tipo
"""
That would be an iacute and an aacute, respectively.
I am aware of the need to set the environment variable APR_ICONV_PATH
when using the native win32 binaries, as mentioned most recently in
the thread starting here:
http://svn.haxx.se/users/archive-2005-01/1547.shtml. Yet, this is
*not* a solution for me, because it only works if the subversion
binaries have been linked to APR iconv. On the other hand, the
binaries distributed by the Cygwin project are linked with GNU gettext
and GNU libiconv; and the gettext libraries have the proper
dependencies on libiconv for automatic charset conversion. I am not
willing to change the locale[1] settings of my Cygwin sessions from
Spanish to English or C, because the other applications that I use
under Cygwin work fine, Spanish messages are output as they should.
My question:
Does on-the-fly charset conversion works for subversion message
catalogs under cygwin? If so, is there a similar hack for GNU iconv as
there is for APR iconv?
My present stop-gap measure is to rename the Spanish message catalog,
so that it isn't found by the svn executables. That way I can have
Spanish messages in the applications that provide them and English
messages in subversion.
[1] Locale under Cygwin must be understood loosely as it is a partial
implementation that depends on the limitations of the underlying win32
OS.
---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe@subversion.tigris.org
For additional commands, e-mail: users-help@subversion.tigris.org
Received on Wed Feb 9 13:54:38 2005