[PATCH 3 of 5] Add util.splitpath() and use it instead of using split() directly

[Shun-ichi GOTO]

2008/1/9, Matt Mackall <mpm at selenic.com>:
> > > On Tue, 2008-01-08 at 16:24 +0900, Shun-ichi GOTO wrote:
> > > > 2008/1/8, Matt Mackall <mpm at selenic.com>:
> > > > > Here's a big question that we have to answer with regard to MBCS: what
> > > > > happens if you check in a path with a 0x5c on a shift-jis machine and I
> > > > > check it out in ascii-land? I suspect the answer is: I get an extra
> > > > > directory level and much confusion.
> > > >
> > > > Is this question for the behaviour on patched mercurial?
> > > > Or for current hg?
> > > >
> > > > If former, the people in ascii-windows-land may gets confuzed
> > > > by error.
> > > > The pepole asii-unix-land people are good to play, no extra directory.
> > > >  This is limitation of my patch
> > > > first I described in [PATCH 0 of 5],  "effects and save local only".
> > >
> > > How does it work? What byte sequence gets stored in the manifest?
> >
> > ??
> > shift_jis encoded bytes in manifest.
>
> Ok, so when an ascii-land user checks this out, they'll see extra
> slashes, right? They'll have no idea it's in shift-jis.

No.
Shift_JIS (and big5) doesn't have slash ('/' 0x2f) in second byte.
sjis 1st byte is in range : 81-9F, E0-FC
2nd byte is in range: 40-7E, 80-FC

> In other words, we need to do step (c) below.
>
> > > a) do the os.sep and split cleanups like you already posted
> > > b) in your extension, wrap the various built-ins and library functions
> > > that report paths
> > > c) add shift-jis <-> utf-8 to util.pconvert and util.localpath
> >
> > I'll see and try for that way.

I see what you wanna say.
However, it's no warry 'cause slash does not appear.

I'm all right about storing as utf-8 in the mean of nomalization
because if he want to convert file-name to 'local encoding' by
exchanging via utf-8, that extension could do.
(e.g. interplatform (sjis <-> euc-jp) repository sharing)

-- 
Shun-ichi GOTO