apr-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Roy T. Fielding" <field...@apache.org>
Subject Re: cvs commit: apr STATUS
Date Sat, 13 Jul 2002 02:39:45 GMT

On Friday, July 12, 2002, at 07:05  PM, William A. Rowe, Jr. wrote:

> At 08:49 PM 7/12/2002, fielding@apache.org wrote:
>>          2) Renaming the function to get rid of apr_time_t vs time_t 
>> confusion,
>>             but keep it ambigious and make no contract with the user 
>> about the
>>             units represented.  Needs a better suggestion than 
>> apr_timeval_t.
>>   -         +1: aaron, brianp, ianh
>>   +         +1: aaron, brianp, ianh,
>>   +             fielding [prefers apr_time and apr_span (_t is half the 
>> problem)]
> Just as a point of reference, we have adopted _t for all types in APR by
> convention.  If this is our type, it needs an apr_ prefix and _t suffix.

Oh, I see -- it is only inconsistent in apr-util.  Never mind then.

> That's just the way the library has evolved.  Screw the '_t is reserved to
> the implementation' when we've already gone and apr_ decorated it all.
> As soon as a vendor comes out with an apr_foo_t type... we can all
> chuckle.

In general, I would hope that we could learn from design decisions that
lead to errors in client programs.  Making all of our types look like
POSIX types is one of those errors.


View raw message