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

Re: Test failures since updating past the merge of locking

From: C. Michael Pilato <cmpilato_at_collab.net>
Date: 2005-04-01 18:24:43 CEST

Julian Foad <julianfoad@btopenworld.com> writes:

> I have always been using BDB in the past. I'm not sure if the recent
> discussion about making FSFS the default is relevant, but I have not
> altered my build/test script. I configured with:
>
> cd /home/julianfoad/build/subversion/
> CFLAGS=-Wall -Wundef -Wendif-labels -Wcast-qual -Wstrict-prototypes
> -Wmissing-prototypes -Wredundant-decls -Wnested-externs
> /home/julianfoad/src/subversion/configure --with-apxs=/usr/sbin/apxs2
> --enable-maintainer-mode --disable-mod-activation
> conftest.c:28: warning: function declaration isn't a prototype
> configure: WARNING: no JNI header files found.
>
> The test log of the first failure is:
>
> START: locks-test
> /home/julianfoad/src/subversion/subversion/libsvn_subr/io.c:1628: (apr_err=39)
> svn: Can't remove 'test-repo-lock-only': Directory not empty

I suspect that this is what happens if you had existing test-repo-*
repositories lying around as BDB repositories, but then run the test
suite with FSFS as the backend of choice (or non-choice, as the change
in default causes).

What happens if you hand-remove those test-repo-* repositories and try
again? If that succeeds, then I'd guess that svn_repos_delete() (or
more accurately, svn_fs_delete_fs() isn't quite doing the right thing
when checking what kind of FS backend is in use by the FS it's about
to destroy.

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Received on Fri Apr 1 18:29:16 2005

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