co via 40 kbps modem connection: interrupt. seems to be showstopper ...

svn co https://... interrupts on some urls (depending which subtree
you check out). the following happens:
A wc/trunk/app/html
svn: RA layer request failed
svn: PROPFIND of /svnrep/!svn/bc/257/trunk/app/html: could not
connect to server

A wc/trunk/app/classes/connector
svn: RA layer request failed
svn: could not fetch directory entries
svn: PROPFIND of /svnrep/!svn/bc/250/trunk/app/classes/connector:
could not connect to server

client: w2k, brane 16.0.
server: rh7 16.0 rpm.
(the problem was there before, but we thought an update is worth a
try).

how do i check out now? after 2 days waiting for having an idea i
still don't know.

if you restrict to the problem-tree (connector in the above example),
it works. if you do it via faster connection, it works. i played with
timeout in the registry ... but maybe i just don't know how to set
this correctly and i don't know how to set a log-level to trace this.

currently i can see only one solution, which prevents long
error-tracking (and is anyway necessary):

"svn co" must not force a restart of the checkout, if it fails. it
must get a whole folder and then proceed to the first subfolder. or,
in other words, it must get the mimimum, so "svn up" can be used to
complete the checkout by getting missing contents.

and re-running "svn co (-force)" is not allowed to give an error, the
repository is the master, and the working copy has to be adjusted in
this case.

;]

__________________________________________________
Do you Yahoo!?
Yahoo! Mail Plus - Powerful. Affordable. Sign up now.
http://mailplus.yahoo.com

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Received on Sun Dec 29 12:46:10 2002