Re: Candidate 'getdate' replacement: VERSION TWO.

Notes on the header file:
> relative times should always parse the same even if svn_parse_date
> is called multiple times during a computation of finite length.

One hopes that all of our computations are of finite length. Perhaps
you mean "nonzero length."

(Also, you indent the function declaration; I couldn't find it at
first as a result.)

The implementation looks sane, except where you noted otherwise in
comments, and I agree with the approach. A little exception:

> struct tm tmcopy = *tm, _tm;

What's with the _tm naming? Although this code is correct C by a
strict interpretation of the standard (because _tm is an automatic
variable, and _foo variables are only reserved to the implementation
for linkage), it's ugly and skirts the boundaries of correctness.

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Received on Sat Oct 21 14:36:36 2006