Re: More protection agains bugs in backend API

[Jamie McCracken <jamiemcc blueyonder co uk>]

Alexander Larsson wrote:

> > I find it a bit pity that a "simple malloc bug" make nautilus completly 
> > freeze/crash.
> > Won't it be possible to protect it against that (timeout, error code return)
> >
> > I'm aware that the best would be to have bugs nowhere :-) but in this 
> > particuliar case, it would have been painless to show Mime-type = unknow 
> > instead of having nautilus unusable.
> >
> > Anyway, I am not a developper so I don't realize the sum of work (or 
> > even the feasability) it would demand.
> > I just wanted to submit the idea.
>
>
> Welcome to development in C. There is no way to make this any better,
> short of moving to a "managed" runtime.

There are more saner ways of fixing this then going to one of those 
"managed" toy languages (unless you're happy having up to 50% slower 
code and up to 600% increased memory usage!).

You could simply move the offending code out of process - a simple and 
effective solution?

jamie.