Hello, Rich!
> Opps - sorry I found this and didn't post it right away - details below.
You found what? Could you give more details?
> > I understand that stdout and stderr are transferred separately. and the
> > error from start_fish_server is not guaranteed to arrive before '### 200'
> >
> > That probably explains why FiSH _sometimes_ works.
>
> I found these problems in starting work on a real dav vfs: I found 3
> places that 2>dev/null is needed and the patches do seem to make the
> connection reliable on my servers:
You were carefull to quote my whole message, but do you realize that I was
talking about stderr and stdout of a _remote_ process?
> 355c384
> < "ls -lLa \"/%s\" | grep '^[^cbt]' | (\n"
> ---
> > "ls -lLa \"/%s\" 2>/dev/null | grep '^[^cbt]' | (\n"
Please don't use "trivial" patches - use only unified (-u) or context (-c)
diffs. I don't even see what file you are changing.
Another question - what are you fixing? Is it related to FiSH is any way?
--
Regards,
Pavel Roskin