If you want to analyze what is going on on Linux, that would be great. You could set a break point on File::getPath(Pool) which would be the entry point to encode the file path prior to an APR call. Likely, the call apr_os_locale_encoding() in charsetencoder.cpp is returning "C" which indicates that the current charset is just 7-bit ASCII and the proper response to that is to translate unrecognized characters as "?".
As of rev 627996, non-ascii file names seem to be working on Mac OS/X
and Windows, however Ubuntu 7.01 still has the ? characters. After
running the unit tests, src/test/resources should contain dom
+superscript 3 and dom+ideographic 4. The ideograph displays on my
Mac, but I just get a box on Windows, guess I don't have the asian
fonts loaded. Feedback would good.
- Double byte character support Steve Brand
- Re: Double byte character support Curt Arnold
- RE: Double byte character support Steve Brand
- Re: Double byte character support Curt Arnold
