Re: STRANGE SEGFAULT

[Paul Davis <pbd Op Net>]

>El Mar 09 Abr 2002 01:27, Paul Davis escribió:
>> >The code "seems" ok, so I have tested diferent solutions, in the last, I
>> > have decided to lock memory using mlockall(), when using it, application
>> > runs ok, I have no much experience on direct system call programming so I
>> > can't understand this segfault.
>>
>> if you don't understand this segfault, i suggest you stay away from
>> mlockall() for a fairly long period of time (units of weeks or
>> months). mlockall() has absolutely *nothing* to do with your problem.
>
>are you sure ??
>perhaps .... but using it runs ok, and without it no, some sugestion ?

you've prevented page faults from occuring for your program, which
changes thread timing and other memory-management-related
behaviour. its not suprising that it changes the visible behaviour.

>> did you read the FAQ entry about threads? www.gtk.org
>
>I'm using pthreads not gthreads. and I understand perfectly the use of 
>pthreads. 

it has nothing to do with pthreads versus gthreads. it has do with the
fact that neither GTK nor GDK nor X Window is thread safe, and unless
you use mutual exclusion to ensure that only 1 thread is ever calling
a GTK or GDK or Xlib function at a time, errors like the ones you
report will be inevitable (and are reported on this list at least once
every 10 days).

--p