Re: svn commit: r31005 - branches/1.5.x

John Peacock wrote:
> Is no one else seeing a weird binding test failure with the Perl
> bindings, specifically with t/3client.t?

FWIW, a binary search of the 3client.t file itsels reveals it is this test that
is causing the failure to be thrown (even though the *test* itself passes):

isa_ok($ctx->info("$wcpath/dir1/newxyz", undef, 'WORKING', sub {}, 0),
        '_p_svn_error_t',
        'info should return _p_svn_error_t for a nonexistent file');

If I comment out that line, everything else passes and the library shutdown code
doesn't freak out.

Running under gdb does not show a segfault or anything untoward. So it appears
that some error is propagating out and not being captured by the normal binding
code (the return code as reported by gdb was 0377). I'm trying to get something
useful out of valgrind, to see if is a real leak, but I need to cut down the
test further...

John

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe_at_subversion.tigris.org
For additional commands, e-mail: dev-help_at_subversion.tigris.org
Received on 2008-05-03 21:09:40 CEST