Error checking for g_mod_event() and friends

[Poul-Henning Kamp]

In message <20040127154303.E92925 at pcle2.cc.univie.ac.at>, Lukas Ertl writes:
>Hi,
>
>is there a specific reason that the .init function of a class is declared
>to return "void"?  I'd like to change that to an int, so we could do error
>checking, and subsequently modify g_load_class() to check for errors and
>so on.

In what circumstances would class->init() fail ?

>Additionally, I've seen several calls to g_malloc() where the return
>values aren't checked - should I prepare some patches for these?

yes please.

-- 
Poul-Henning Kamp       | UNIX since Zilog Zeus 3.20
phk at FreeBSD.ORG         | TCP/IP since RFC 956
FreeBSD committer       | BSD since 4.3-tahoe    
Never attribute to malice what can adequately be explained by incompetence.