[svn.haxx.se] · SVN Dev · SVN Users · SVN Org · TSVN Dev · TSVN Users · Subclipse Dev · Subclipse Users · this month's index

Re: ra_dav and svn_recurse_kind

From: Philip Martin <philip_at_codematters.co.uk>
Date: 2005-05-06 22:22:49 CEST

Philip Martin <philip@codematters.co.uk> writes:

> 152 if (r == svn_recursive)
> (gdb) p r
> $31 = 3221222280
> That value of r is obviously rubbish (remember it's a stack address)
> so I fix it in the debugger

I fixed this in r14594.

> (gdb) set var r = 2
> and the code continues to push_wc_prop
> push_wc_prop (baton=0x808bd30, relpath=0x80ac828 "zig/bar",
> name=0x40123448 "svn:wc:ra_dav:version-url", value=0xbffff054,
> pool=0x809eae0) at ../svn/subversion/libsvn_client/ra.c:123
> 123 svn_client__callback_baton_t *cb = baton;
> Now push_wc_prop iterates over commit_items comparing the relpath
> against each items url. The urls are "", "zig", "zig/bar/baz" none of
> which match relpath, so push_wc_prop does nothing. I haven't got a
> clue whats going on: the dav later appears to be setting a version-url
> and the client layer is ignoring it. Once the commit is completed if
> I look at the wc there are no dir-wcprops on "zig/bar". Why does the
> client layer (push_wc_prop) veto the dav layer's attempts to store
> version-urls?

This bit is still a mystery to me: why does the ra_dav layer have
explict code to set the property while the client layer has explicit
code to veto it?

Philip Martin
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Received on Fri May 6 22:24:15 2005

This is an archived mail posted to the Subversion Dev mailing list.

This site is subject to the Apache Privacy Policy and the Apache Public Forum Archive Policy.