> -----Original Message-----
> From: Geoff Rowell [mailto:geoff.rowell_at_varolii.com]
> Sent: maandag 2 maart 2009 15:40
> To: Paul Burba
> Cc: dev_at_subversion.tigris.org; Hyrum K. Wright
> Subject: RE: 1.6.0-rc3 up for signing/testing
>
> On Sun, Mar 01, 2009 at 12:22 PM, Paul Burba <ptburba_at_gmail.com> wrote:
> > XFAIL update_tests.py 32: update wc on the root of a Windows (virtual)
> drive.
> >
> > It fails, which we expect, but triggers an SVN_ERR_ASSERT(). The fix
> > for this is r36049-r36051 and r36131, which are nominated for
> > inclusion in 1.6.1. This test fails without incident in 1.5.x, but I
> > am able to get an abort (rather than an SVN_ERR_ASSERT which is new in
> > 1.6) with 1.5.6, so this is not a regression. Also, the situation is
> > fairly uncommon, you must simultaneously commit multiple WCs whose
> > common parent is the root of a drive; so the workaround is simple,
> > commit them individually. Given all this, I don't believe it should
> > hold up the release.
> >
> No desire to hold up the release train, but this is rather more common
than
> you think. Our Windows developers build off of a "subst" K: drive. the
source
> tree includes externals in the top level and in the next few levels.
>
> Committing from "K:\" triggers this.
Hi,
At this time I can't make a promise on this, but I'm trying to fix all/most
of the windows drive root issues for 1.7. (And I had this specific example
working on my pc)
Fixing this issue requires invasive changes through all our libraries
(especially libsvn_wc and libsvn_client), so it is impossible to get these
fixes in 1.6 as just another bugfix.
(The assertion will probably be replaced by a normal error in 1.6.1 if it
gets enough votes in time).
Thanks,
Bert
------------------------------------------------------
http://subversion.tigris.org/ds/viewMessage.do?dsForumId=462&dsMessageId=1256605
Received on 2009-03-02 15:48:10 CET