Both hgwebdir.cgi and webserve-dir.cgi are mbcs broken

Dongsheng Song dongsheng.song at gmail.com
Thu Feb 1 06:41:34 GMT 2007


I just wondered why my filename use GB18030, but log messages use UTF-8 ?

2007/2/1, Matt Mackall <mpm at selenic.com>:
> The problem is not particular to multi-byte character sets. Mercurial
> makes no attempt to transcode anything beyond our internal metadata
> (author, changeset message, etc.) as there's generally no guarantee that:
>
> a) filename or file contents match specified system encoding of committer
> b) filename or file contents of committer could be represented on a
>    given destination system
> c) transcoding filenames or file contents wouldn't break build systems
>
> So we just show the raw binary and report your system locale as the
> HTML page encoding. Your system is claiming UTF-8, but your filenames
> are not UTF-8.
>



More information about the bazaar mailing list