I'm still needing lots of help in getting the windows-i18n branch of Fossil running better.
The strategy on that branch can be summarized as follows: (1) All text internal to Fossil is UTF8. (2) Command-line arguments (the values of argv[] passed to main()), environment variables, and other inputs are assumed to be in a locale-dependent code page which I will call MBCS. All inputs are converted from MBCS into UTF8 as soon as they are encountered so that internally to Fossil everything is always UTF8. (3) String data (filenames) that are passed to windows system calls are converted from UTF8 into MBCS. (4) Output to the console is converted from UTF8 back into MBCS. Several people have been testing this branch sporatically. But they report a variety of troubles. Some people report that while input from the console are MBCS, the output to the console should be UTF8. Or not, depending on how the system is set up. If you have insights on how to better handle windows i18n issues, please provide them. One interesting way to test the latest Fossil is to use the "test-echo" command. "fossil test-echo ...." simply prints its arguments, like the "echo" command of unix shells. By invoking "fossil test-echo ..." with filename arguments that contain non-ASCII characters, you will be testing the mechanisms outlined above. Please do this if you are able. Please recall that I do not use windows, and I do not speak any languages that require non-ASCII characters, so I have great difficulty trying to test these changes myself. I am utterly dependent upon the community to identify problems, suggest fixes, and provide feedback. If getting Fossil to work well on windows machines with non-ASCII code pages is important to you, then please consider helping with the debugging process. -- D. Richard Hipp [email protected]
_______________________________________________ fossil-users mailing list [email protected] http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users
