I've been running into some problems with parsing files which include 8-bit ascii characters with Xerces-C++ and Redhat 6.1.
The problem is that the wide character routines in glibc 2.1.2-11 screw up with certain strings of wchar_t's initially derived from a mixture of 7-bit and 8-bit ascii characters. line 350 in src/util/Transcoders/Iconv/IconvTransService.cpp:350 const size_t neededLen = ::wcstombs(NULL, wideCharBuf, 0); When wideCharBuf contains such a string neededLen gets set to a very large random integer which causes great problems later on. I appreciate that this is not a Xerces specific problem and have written to the libc team. Regards, Tom Ford
