xerces-c-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Alberto Massari <amass...@datadirect.com>
Subject Re: SVN compile gives chinese messages?
Date Fri, 25 Aug 2006 08:49:29 GMT
Hi Jason,

At 11.23 25/08/2006 +0530, Jason Stewart wrote:
>Hi David,
>On 8/24/06, David Cargill <cargilld@ca.ibm.com> wrote:
>>Hi Jason,
>>I ran into a problem on some Linux platforms where the build defaulted to
>>using the GUNIconv transcoder.  I updated packageBinaries.pl to specify
>>--enable-transcoder-iconv and that fixed the problem.  Maybe try specifying
>>--enable-transcoder-iconv and see if that works.
>I looked at the Iconv transcoder code - and it indicated that it
>didn't provide a *real* transcoding service the comment in the code
>  This is a minimalist transcoding service, that only supports a local
>  default transcoder. All named encodings return zero as a failure...
>That won't work for Perl - I need a UTF-8 transcoder...
>Alby - what is the big-endian problem in IconvGNU? If I could I would
>put my machine on the 'net and let you play with it - but I'm behind a
>firewall here...

My doubts come from this encoding table in the IconvGNUTransService.cpp:

static const IconvGNUEncoding    gIconvGNUEncodings[] = {
     { "UCS-2LE",        2,    LITTLE_ENDIAN },
     { "ucs-2-internal",        2,    LITTLE_ENDIAN },
     { NULL, 0,    0 }

The last entry is LITTLE_ENDIAN, but the description for the field is

     unsigned int fUBO;        // byte order, relative to the host

Being "relative to the host", I think it should be BYTE_ORDER. But 
it's just a guess, as I haven't tested on a big-endian machine (yet, 
I could be able to do that today).

If you have time to test it now, let me know if this works.


To unsubscribe, e-mail: c-dev-unsubscribe@xerces.apache.org
For additional commands, e-mail: c-dev-help@xerces.apache.org

View raw message