Stefan Fuhrmann <stefan2_at_apache.org> writes:
> static svn_error_t *
> -test_fsfs_config_opts(const svn_test_opts_t *opts,
> - apr_pool_t *pool)
> +test_create_with_config_opts(const svn_test_opts_t *opts,
> + apr_pool_t *pool)
> {
> apr_hash_t *fs_config;
> svn_fs_t *fs;
> @@ -6722,9 +6722,9 @@ test_fsfs_config_opts(const svn_test_opt
> const svn_fs_fsfs_info_t *fsfs_info;
>
> /* Bail (with SKIP) on known-untestable scenarios */
> - if (strcmp(opts->fs_type, "fsfs") != 0)
> + if (strcmp(opts->fs_type, SVN_FS_TYPE_BDB) == 0)
> return svn_error_create(SVN_ERR_TEST_SKIPPED, NULL,
> - "this will test FSFS repositories only");
> + "this will not test BDB repositories");
Is there a reason to run this test with --fs-type=fsx? I think we'd be just
wasting time here, because the test itself doesn't rely on opts->fs_type,
and everything it does is FSFS-specific.
When creating file systems in this test, we specify SVN_FS_TYPE_FSFS and
FSFS-specific config options; when checking what we've got, we downcast the
information structure to an FSFS-specific svn_fs_fsfs_info_t. I think that the
test was designed to run only with --fs-type=fsfs.
Am I missing something?
Regards,
Evgeny Kotkov
Received on 2015-03-19 10:35:44 CET