onsdag 31 januari 2007 23:17 skrev Junio C Hamano:
quoted text
> Your patch does not seem to apply.  I wonder if the test path
> needs to be that looooooooooooong to begin with.
Not really. I had some thought when I did it, but can't recall what it was. 
Perhaps I wanted to verify that the data I was reading from git-diff wasn't 
truncated. I've pruned that part of memory.

quoted text
> One thing I noticed on Cygwin is that from inside Cygwin
> environment t/trash directory becomes unremovable after the
> original test fails.

one more interesting thing..

quoted text
> I am wondering what use it would be to have ISO-8859-1 pathnames
> in our test.  As far as the test is concerned I think the point
> is to try non-ASCII paths, so I suspect this patch might be less
> of an impact.
Yes. Anything non-ascii will do.

quoted text
> Sorry for sending an attachement but in this case the patch text
> is _not_ in any particular encoding (the preimage is ISO-8859-1
> but the postimage is in UTF-8) and transferring it as a textual
> patch over e-mail is not possible without a way to annotate each
> line which encoding it is in.

That patch didn't apply here, not even  Lets try the old uuencode format. 
(this is my patch once again wrapped by uuencode).

-- robin