Re: svn commit: r21621 - in trunk/subversion: libsvn_subr tests/cmdline tests/cmdline/svntest tests/libsvn_subr

Philip Martin wrote:
> lgo@tigris.org writes:
>
>> Author: lgo
>> Date: Sat Sep 23 08:44:04 2006
>> New Revision: 21621
>
>> @@ -810,6 +896,8 @@
>> #if defined(WIN32)
>> "X:/foo",
>> "X:/",
>> + "X:foo",
>> + "X:",
>> #endif /* WIN32 */
>> "",
>> };
>
> The number of tests...
>
>> @@ -819,10 +907,11 @@
>> FALSE,
>> FALSE,
>> TRUE,
>> - FALSE,
>> #if defined(WIN32)
>> + FALSE,
>> TRUE,
>> FALSE,
>> + TRUE
>> #endif /* WIN32 */
>> };
>
> ...does not match the number of results.

Yep, you're right. Fixed that in r21622. Strangely enough it didn't fail
neither on Windows or Linux, only the Mac buildslave had problems with
this test.

> I suggest you look at test_canonicalize to see how to combine these
> two arrays into one as that makes it easy to see that the tests and
> the results match.
>

Well, test_canonicalize is not a good example, because there both the
input and the expected output are paths, so both can be stored in the
same two-dimensional area of (const char *).

The test_is_root test looks is similar to test_is_uri_safe, which is
implemented the same way.

I welcome a proposal to better fix the relation between input and output
  for these two test.

thanks for the review,

Lieven.

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Received on Sat Sep 23 22:20:19 2006