From: Jim Meyering <jim@meyering.net>
Date: Mon, 18 Feb 2008 09:01:09 GMT
Subject: Re: [PATCH] Remove useless if-before-free tests.
Message-ID: <87ve4my6y2.fsf@rho.meyering.net>
In-Reply-To: <alpine.LSU.1.00.0802172210470.30505@racer.site>

Johannes Schindelin <Johannes.Schindelin@gmx.de> wrote:
> On Sun, 17 Feb 2008, Jim Meyering wrote:
>
>> It is equivalent not just because POSIX has required free(NULL) to work
>> for a long time, but simply because it has worked for so long that no
>> reasonable porting target fails the test. Here's some evidence from
>> nearly 1.5 years ago:
>>
>>     http://www.winehq.org/pipermail/wine-patches/2006-October/031544.html
>
>>From this mail, we see that there is at least one target where this leads
> to a crash (remember, git should run on more platforms than Wine).

Hi,

Thanks for the feedback.

FYI, you don't have to go back 20+ years to 3BSD to find a system on
which free(NULL) fails :-)  SunOS4's did, too.  So if that is a reasonable
porting target for git, then you will need the wrapper.  With references
to "SunOS" in Makefile and configure, I did wonder about that.  Let me
know and I'll adjust the proposed patch.

> However, such a crash is pretty obvious in our test-suite, I guess, and
> thus we could easily introduce something like this into git-compat-util.h
> should the need ever arise:
>
> #ifdef FREE_NULL_CRASHES
> inline void gitfree(void *ptr)
> {
> 	if (ptr)
> 		free(ptr);
> }
> #define free gitfree
> #endif
>
> IOW I like that type of cleanup.

:-)