Re: svn commit: rev 7162 - trunk/subversion/libsvn_subr

C.Michael Pilato wrote:

>cmpilato@tigris.org writes:
>
>
>
>>Author: cmpilato
>>Date: Tue Sep 23 15:04:43 2003
>>New Revision: 7162
>>
>>Modified:
>> trunk/subversion/libsvn_subr/io.c
>>Log:
>>* subversion/libsvn_subr/io.c
>> (io_check_path): Treat APR's ENOTDIR status the same as ENOENT --
>> both are now non-fatal forms of the answer "that path doesn't exist."
>>
>>
>
>Okay, Garrett. Version your hooks/ directory allllll you want to, now.
>
>
I can't remember, should we check ENOENT before ENOTDIR, or the other
way around? /me fumbles in the APR sources ... Ah, I knew I wrote
something to that effect:

 * @warning For any particular error condition, more than one of these tests
 * may match. This is because platform-specific error codes may not
 * always match the semantics of the POSIX codes these tests (and the
 * correcponding APR error codes) are named after. A notable example
 * are the APR_STATUS_IS_ENOENT and APR_STATUS_IS_ENOTDIR tests on
 * Win32 platforms. The programmer should always be aware of this and
 * adjust the order of the tests accordingly.

(Yuck, I also wrote that typo.)

Yah, it should be ENOTDIR before ENOENT, so everything is fine.

-- 
Brane Čibej   <brane_at_xbc.nu>   http://www.xbc.nu/brane/
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org