subversion-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Hyrum K Wright <hyrum.wri...@wandisco.com>
Subject Re: svn commit: r1310047 - /subversion/trunk/subversion/libsvn_ra/compat.c
Date Thu, 05 Apr 2012 21:57:32 GMT
On Thu, Apr 5, 2012 at 4:43 PM, Greg Stein <gstein@gmail.com> wrote:
> On Thu, Apr 5, 2012 at 16:48, Hyrum K Wright <hyrum.wright@wandisco.com> wrote:
>> On Thu, Apr 5, 2012 at 3:29 PM, Greg Stein <gstein@gmail.com> wrote:
>>> On Thu, Apr 5, 2012 at 16:15,  <hwright@apache.org> wrote:
>>>> Author: hwright
>>>> Date: Thu Apr  5 20:15:41 2012
>>>> New Revision: 1310047
>>>>
>>>> URL: http://svn.apache.org/viewvc?rev=1310047&view=rev
>>>> Log:
>>>> * subversion/libsvn_ra/compat.c
>>>>  (prev_log_path): Another integer-width mismatch fix.
>>>>
>>>> Modified:
>>>>    subversion/trunk/subversion/libsvn_ra/compat.c
>>>>
>>>> Modified: subversion/trunk/subversion/libsvn_ra/compat.c
>>>> URL: http://svn.apache.org/viewvc/subversion/trunk/subversion/libsvn_ra/compat.c?rev=1310047&r1=1310046&r2=1310047&view=diff
>>>> ==============================================================================
>>>> --- subversion/trunk/subversion/libsvn_ra/compat.c (original)
>>>> +++ subversion/trunk/subversion/libsvn_ra/compat.c Thu Apr  5 20:15:41 2012
>>>> @@ -141,7 +141,7 @@ prev_log_path(const char **prev_path_p,
>>>>               svn_sort__item_t item = APR_ARRAY_IDX(paths,
>>>>                                                  
  i - 1, svn_sort__item_t);
>>>>               const char *ch_path = item.key;
>>>> -              int len = strlen(ch_path);
>>>> +              size_t len = strlen(ch_path);
>>>
>>> size_t or should this be apr_size_t?
>>
>> You ask this question *every time* one of these changes is made. :P
>>
>> And the response usually is "apr_size_t is a typedef of size_t on
>> every platform we support, and we already use a mix of the two
>> throughout our code".
>
> I count about 100 size_t usages, and over 1200 apr_size_t. I think the
> vote is for the latter :-P

I went with 'size_t' in this case since the value is the result of a
call to strlen(), a standard library function defined to return
size_t.  In cases where we interact with APR, I usually favor
apr_size_t, since that's what the APIs ask for / require.  I suspect
others follow this same pattern.

> Personally, I don't care much either way except for consistency. And
> so when I see a bare size_t, then I ask. I think the tendency should
> be for apr_size_t unless there is some specific reason otherwise.

I've no objection to standardization; I just don't care enough to be
the one to do it. :)

-Hyrum



-- 

uberSVN: Apache Subversion Made Easy
http://www.uberSVN.com/

Mime
View raw message