Re: Making GNOME crash

[Emmanuele Bassi <ebassi gmail com>]

Hi,

On Mon, 2005-11-07 at 09:58 -0700, Elijah Newren wrote:
> On 11/7/05, Gustavo J. A. M. Carneiro <gjc inescporto pt> wrote:
> >   This could also mean that developers start thinking twice before
> > adding a g_warning, and start using g_message instead, thus defeating is
> > purpose.  g_warning is for reporting real problems, but no critical
> > ones; for critical problems we have g_error.
> >

But, right now, critical warnings are just swept under the carpet of the
console.  How many of those messages marked as "critical" are noticed?

> >   It's so easy to turn on abort on warnings, just run your program with:
> >         $ G_DEBUG=all ./myprogram
> > or even:
> >         $ ./myprogram --g-fatal-warnings
> >
> >   Only lazy developers let g_warnings go on indefinitely.  We need to
> > educate developers, that's all.
> 
> I think you missed Vincent's original email (either that or I
> misunderstood it).  I believe he was talking about making g_critical()
> calls result in crashes, not g_warning() calls.

Exactely.

That was the reason for adding a new option for G_DEBUG in the first
place.

>   So these examples
> don't do what was proposed, though they are close (just use
> "G_DEBUG=fatal_criticals" instead of "G_DEBUG=all", if I understand
> correctly).

Another thing that I don't think has been stressed out enough is that
this would be enabled only on development releases: that would mean
crashing only when things are supposed to crash with the loudest bang
possible, in order to be fixed fast.

I agree that it could potentially scare users of the development
releases - even though I don't believe that will be the case; and the
"shame factor" for having a module crashing would be enough for making
sure that no *critical* warning is left unchecked.

Ciao,
 Emmanuele.

-- 
Emmanuele Bassi - <ebassi gmail com>
Log: http://log.emmanuelebassi.net