[BUG] bzr 0.8 add on Unicode filename fails

[John Whitley]

John Arbash Meinel wrote:
> Well, don't forget there is 'cygpath' which is an executable which  
> will
> map for you. (You can use cygpath -w to get the windows path).

Yep.  What I specifically meant was a wrapper that uses cygpath  
internally, to avoid too much of `cygpath -w ...`.

> However, unless you are doing absolute paths, bzr accepts forward
> slashed paths without any problem.

Good to know.

> On my machine I just have a 'bzr.bat' which is:
>
> @echo off
> set BZR_PLUGIN_PATH=h:\dev\bzr\plugins
> C:\Python24\python H:\dev\bzr\bzr.dev\bzr %*
>
> the BZR_PLUGIN_PATH is just so that cygwin and win32 use the same  
> plugin
> paths. (And I prefer an exposed dir, since I modify plugins a lot).
>
> Obviously your paths may vary, but this works well for me.
>
> Also, you may want to realize that utf8 doesn't really work with
> cmd.exe.

Not a problem, since I can't stand cmd.exe anyways.  ;-)  Although  
I'm not clear on the Unicode status of rxvt.  I did a brief search  
earlier, but the results seemed inconclusive.  I have noted that 'bzr  
log' under rxvt on bzr.dev fails to correctly display names of  
commiters with non-ascii characters.  Not sure where the fault lies  
there.

> But also be forewarned that you have to be very careful with
> redistributing cygwin1.dll.

Thanks for the heads-up.  Fortunately I'm just using it as a platform  
for in-house tools (e.g. to civilize Windows), and so this shouldn't  
be an issue.

-- John