Re: [PATCH] Fix a signal handler

From: Junio C Hamano <gitster@pobox.com>
Date: Feb 23, 2010, 21:48 UTC
Message-ID: <7vmxyzfwt7.fsf@alter.siamese.dyndns.org>
In-Reply-To: <4B839811.6040109@web.de>

Markus Elfring <Markus.Elfring@web.de> writes:

Show 5 quoted lines

>> According to POSIX, "s-e-o" has to be "volatile sig_atomic_t".
>
> How do you think about informations from a discussion on a topic like 'Is
> "volatile sig_atomic_t" redundant'?
> http://groups.google.de/group/comp.lang.c/browse_frm/thread/da3118a2d2c0737c/718dc093b83e03f8?#718dc093b83e03f8

Honestly I don't care; you are the one who are interested in being pedantic, and you are welcome wasting your time on that endeavor. Don't ask me to waste my time by joining your mental masturbation.

>> Also we do not explicitly initialize bss variables to zero or NULL.
>
> If we would like to insist on the implementation of a strictly conforming
> program,...

We don't.

The thing is, we do not like to insist any such thing. We are practical bunch who are interested in getting git work well on real platforms used by real people. Portability across platforms people care about is one of the goals and standard conformance for us is merely a tool to achieve it.

Name one platform you tried to port git to and had trouble with because the platform did not initialize variables in bss segment to zero, or perhaps on that platfor NULL had a bitpattern different from all zero, and after you initialized them explicitly to zero or NULL, you managed to make everything work perfectly.

Name one platform you actually got a segfault in the early-output codepath on it, because a function pointer on that platform is not of an atomic type, and the assignment from show_early_output to show done in limit_list() picked up a pointer half-written by the signal handler, and we ended up calling a garbage address, and you managed to make everything work perfectly with your fix.

Just name one.

Standard conformance by itself is never a goal for us, unless it helps to solve real world problems. And until you understand that, you wouldn't understand why this patch deserves to be labelled with "pedantic fix".