Re: [RFC] Cleaning up die() error messages

From: Matthias Urlichs <smurf@smurf.noris.de>
Date: Oct 11, 2005, 19:48 UTC
Message-ID: <pan.2005.10.11.19.48.04.675482@smurf.noris.de>
In-Reply-To: <7vzmph42j2.fsf@assigned-by-dhcp.cox.net>

Hi, Junio C Hamano wrote:

> One thing to keep in mind is how badly this C part might
> interact with the libification effort going on underwater.

Not too badly.

> Since current code Smurf is working on is based on 0.99.6 and

I've merged it up once already; will do that again soon.

> many small pieces need to be reviewed anyway, I am not so much
> worried about forward porting the changes.

I've also mostly succeeded in keeping the individual patches clean so that everything still builds (and verifies), so it might be easiest to just merge with it. ;-)

But we'll cross that bridge when we get to it.

>                             But some die()s that
> are in the parts that will be moved to the common library code
> would also want to use this prog global somehow.

IMHO, common library code should not be allowed to die. (Yes, that does imply replacing all the xmalloc() calls.)

My library effort has a buffer for the (first) error message. The caller can elect to suppress printing it, so that it can be formatted appropriately. Python, for instance, will wrap errors in an exception.

The way I structured it so far, a C program would do

	git_env = git_env_new();
	die_if_null(git_env());
	git_env->print_error = 0;
[...]
	git_whatever(git_env, ...);
	if (git_env->error) {
		fprintf(stderr, "%s: doing whatever: %s\n",
			my_program_name, git_env->error);
		git_env_clear_error(git_env);
		goto whatever_bad_so_go_clean_up;
	}

-- 
Matthias Urlichs   |   {M:U} IT Design @ m-u-it.de   |  smurf@smurf.noris.de
Disclaimer: The quote was selected randomly. Really. | http://smurf.noris.de
 - -
A hundred mouths, a hundred tongues, And throats of brass, inspired with
iron lungs.
					-- Virgil