----- Original Message -----
> From: Greg Stein <[email protected]>
> To: [email protected]
> Cc:
> Sent: Wednesday, 25 April 2012, 9:37
> Subject: Re: svn commit: r1330135 -
> /subversion/trunk/subversion/tests/svn_test.h
>
> On Wed, Apr 25, 2012 at 03:29, <[email protected]> wrote:
>> Author: julianfoad
>> Date: Wed Apr 25 07:29:07 2012
>> New Revision: 1330135
>>
>> URL: http://svn.apache.org/viewvc?rev=1330135&view=rev
>> Log:
>> * subversion/tests/svn_test.h
>> (SVN_TEST_STRING_ASSERT): Clarify the doc string, following r1329808.
>>
>> Modified:
>> subversion/trunk/subversion/tests/svn_test.h
>>
>> Modified: subversion/trunk/subversion/tests/svn_test.h
>> URL:
> http://svn.apache.org/viewvc/subversion/trunk/subversion/tests/svn_test.h?rev=1330135&r1=1330134&r2=1330135&view=diff
>>
> ==============================================================================
>> --- subversion/trunk/subversion/tests/svn_test.h (original)
>> +++ subversion/trunk/subversion/tests/svn_test.h Wed Apr 25 07:29:07 2012
>> @@ -74,7 +74,9 @@ extern "C" {
>> } while (0)
>>
>> /** Handy macro for testing string equality.
>> - * EXPR and/or EXPECTED_EXPR may be NULL.
>> + *
>> + * EXPR and/or EXPECTED_EXPR may be NULL which compares equal to NULL and
>> + * not equal to any non-NULL string.
>
> Honestly, I don't care how it compares to anything. What does it do to
> the *TEST*?
It's documented as testing two strings for equality. Sure, it doesn't
explicitly say what it does if they are or aren't equal, but all I did was
extend the semantics of what is a valid "string" input. That's why how NULLs
compare is the relevant additional fact.
> Does it raise a failure? Is it successful?
I'm happy to write a better doc string.
* If EXPR and EXPECTED_EXPR are pointers to equal null-terminated
* strings, or are noth NULL, do nothing. Otherwise, execute a
* 'return' statement returning a pointer to an svn_error_t object
* that describes the mis-match.
Like that?
> The docstring is opaque. I need to look at the macro to determine what
> it will do to my test case.
But you *like* reading the code :-)
- Julian
