Quoting Karl Fogel <kfogel@newton.ch.collab.net>:
> Branko �ibej <brane@xbc.nu> writes:
> > >>+def get_md5(path):
> > >>+ """Return the hex md5 digest of file PATH."""
> > >>+ f = open(path, 'r')
> >
> > Crash!
> >
> > Read files in binary mode... Write files in binary mode...
> > [repeat 50 times]
> > :-)
>
> Thanks! I forgot that Python makes that distinction by default.
>
> Hmmm. I've fixed all the open calls where it's appropriate, except
> for two cases where we're opening an RCS file, in visit_file() and in
> RevInfoParser.parse_cvs_file(). In the RCS file format, all
> whitespace outside strings is treated the same, which I suppose
> technically means that an RCS file on Windows *could* have either LF
> or CRLF, and could have either on Unix too. In practice, I don't know
> whether CVS on different operating systems creates the RCS files
> differently, and even if it did, I'm not sure the rcsparse module
> would care, depending on how carefully it parses whitespace.
CVS on Windows creates RCS files with CRLF line endings. And that's a royal pain
in the gluteus maximus because you can't migrate CVS repositories between Unix
and Windows servers. It's also not quite clear to me what happens to line
endings if you have a text file stored with -kb set, or what happens with real
binary files.
Sounds like an interesting way to get a nervous breakdown, doesn't it?
> The easiest thing is probably just to try cvs2svn on Windows and see
> what happens.
"Hmm, what happens if I push the Big Red Button?" :-)
Brane
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Received on Tue Apr 29 17:38:09 2003