trac-admin does not convert national characters to utf-8 and back

[anonymous]

all characters entered from keyboard (for example 'component add "national-characters-here") go to the database without converting to utf-8.
so in web interface they are shown incorrectly.

[Nathan Miller <nxmiller@…>]

31a32
> import locale
37a39,47
> def to_utf8(s):
>     return unicode(s,locale.getdefaultlocale()[1]).encode('utf-8')
> 
> def from_utf8(s):
>     try:
>         return s.decode('utf-8')
>     except UnicodeError:
>         return s
> 
170a181
>                 cell = from_utf8(cell)
181c192
<                 print ("%%-%ds%s" % (colw[cnum], sp)) % (ldata[rnum][cnum] or ''),
---
>                 print ("%%-%ds%s" % (colw[cnum], sp)) % (from_utf8(ldata[rnum][cnum]) or ''),
988c999
<                 tracadm.docmd (command)
---
>                 tracadm.docmd (to_utf8(command))

[Christopher Lenz]

Don't close as anonymous or without a comment.

[mr700@…]

With Red Hat 9 and Fedora Core ½/3 all entered text is in utf-8 unless you change the default setup. Shouldn't this be checked before converting from and to utf-8? With FC3 I have only a small problem with trac-admin while erasing multibyte characters with backspace, but this is going to be fixed soon I hope :)

[Matthew Good]

Please don't close without a comment.

[anonymous]

well all that sounds very good i love this applicatoin - it is great thx

[anonymous]

dfgjkdlgj

[linas]

testuojam

[Matthew Good]

Please don't abuse the system

[Emmanuel Blot]

This incident is probably tied to #2394.

(sorry for the new status instead of reopened, clearing out the owner seems to reset the status)

[Christian Boos]

This has been fixed by r3114. See also #2186.