Six years ago we set these two lock tests to XFail
(http://svn.apache.org/viewvc?view=revision&revision=853631):
C:\SVN\src-trunk>Debug\subversion\tests\libsvn_fs\locks-test.exe
--list --mode-filter xfail
Test # Mode Test Description
------ ----- ----------------
9 XFAIL able to reserve a name (lock non-existent path)
10 XFAIL directory locks (kinda)
These tests were expected to fail at the time because they depended on
the ability to lock non-existent paths with svn_fs_lock():
locks-test 9:
/* DAV clients sometimes LOCK non-existent paths, as a way of
reserving names. Check that this technique works. */
static svn_error_t *
lock_name_reservation(const svn_test_opts_t *opts,
apr_pool_t *pool)
locks-test 10:
/* Test that we can set and get locks in and under a directory. We'll
use non-existent FS paths for this test, though, as the FS API
currently disallows directory locking. */
static svn_error_t *
directory_locks_kinda(const svn_test_opts_t *opts,
apr_pool_t *pool)
Are these two tests legitimate anymore?
Do we ever expect svn_fs_lock() to be able to lock non-existent paths?
Do "DAV clients sometimes LOCK non-existent paths, as a way of
reserving names"? I'm not sure exactly what that means, does anyone
have an inkling?
Paul
