Re: [Denemo-devel] Denemo Crashing after git pull

[Richard Shann]

On Sun, 2014-10-19 at 20:24 +0200, Éloi Rivard wrote:
> Ok, have you met some crashes since you fixed those warnings?

It has been good since then :)

I have been looking at the g_mutex_lock()/unlock() code too - it is
possible to print out which thread is doing the locking and unlocking
and I haven't been able to find fault with that code (it was written by
Dominic Sacre originally). The queue mutex is locked and unlocked from
different threads in the correct order. But it seems that on shutting
down someone has made a threading error.

> 
> They can be made fatal, but I don't know when they can be found.
> I mean, sometimes they may just be mild warnings. What do you think ?

I was thinking of the specific warning "pointer from integer without a
cast", however we could easily make any warning about "assignment
makes..." fatal, as there are only 8 of them to fix.

I think there are two things we should learn from this experience
1) we need to test both 32 and 64 bit builds
2) we should prioritize getting rid of warnings from the compiler over
writing tests. There are only a very few warnings (about pow() floor()
etc) which I do not know how to fix. I already got rid of the vbox and
hbox deprecation warnings, and there are a few others like that which
could be fixed.

Richard