rsync - Jan 2004 - Cygwin hang-problem resolved

Not for me!  I'm using cygwin-1.5.6-1, with rsync 2.5.7-2 (as reported by
cygcheck -s).  I've also tried with rsync-1.4.0 (as someone reported that
worked), and rsync-2.6.0, both built from source.  I get a 100% reliable
hang after the first 0-128k of the first file.  Command looks like:

rsync -e ssh -ar sparky:"/cygdrive/c/www\ sites/" /cygdrive/c/wwwsites

The directory tree comes across, then the first file stops the transfer.
Killing the client end always leaves the process running on the server.
ssh/sftp both seem fine.  All disks involved are local (not network mounted
shares).

Any ideas what the bug was/how to see where it is getting stuck?  I'm more
than willing to build/poke around with gdb if someone points me in the right
direction.

Thanks,

Mark

"Mark Weaver" writes:
> Not for me!  I'm using cygwin-1.5.6-1, with rsync 2.5.7-2 (as reported
by
> cygcheck -s).  I've also tried with rsync-1.4.0 (as someone reported
that
> worked), and rsync-2.6.0, both built from source.  I get a 100% reliable
> hang after the first 0-128k of the first file.  Command looks like:
> 
> rsync -e ssh -ar sparky:"/cygdrive/c/www\ sites/"
/cygdrive/c/wwwsites
> 
> The directory tree comes across, then the first file stops the transfer.
> Killing the client end always leaves the process running on the server.
> ssh/sftp both seem fine.  All disks involved are local (not network mounted
> shares).
This is a different problem.  I've never found rsync + ssh + cygwin
to be reliable, and I don't know why.  Rsyncd + cygwin has been very
reliable for me except for the well-known receiver-side hang.

The well-known hang referred to in this thread happens at the very end
of the transfer at the receiving end when the parent process fails to
kill the child.  By all reports, cygwin-1.5.6-1 solves this problem.

BTW, cygwin-1.5.6-1 has some significant new bugs and will be shortly
updated to cygwin-1.5.7-1.

Craig