[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: Trouble: GTK + Octave + PThreads
From: |
Rafael Laboissiere |
Subject: |
Re: Trouble: GTK + Octave + PThreads |
Date: |
Thu, 5 Nov 1998 16:03:50 +0100 (MET) |
[gtk-list people: if this thread is becoming too Octave-specific, please
tell me and I will remove the gtk-list address from the Cc: list.]
Hi John,
Thanks for your reply.
>>>>> "JWE" == John W Eaton <address@hidden> writes:
JWE> Are you working on creating a set of programmable GUI objects
JWE> that can be used from Octave, or a GUI front end that would
JWE> communicate with Octave?
Mine is the first case.
JWE> I'm not sure why Posix threads would be needed.
With GTK+ (as with other X-based GUI toolkits) it is necessary to launch
a main loop (gtk_main, for instance) which will take care of the mouse
events. As my application will be launched from the Octave prompt, it
gets blocked when gtk_main is called, unless I use threads.
JWE> Can't you just use something like select() to decide where
JWE> input is coming from?
Yes, if my application would be taking care of the readline loop, which
is not the case (Octave is doing it).
JWE> With Tcl/Tk and some other X-based GUI toolkits, it is
JWE> relatively easy to continue to get input from the readline
JWE> prompt while also allowing the GUI objects to function.
I am not sure, but I think that with Tcl/Tk, the function Tcl_Init
spawns a whole new process, which communicate with the main program
through Tcl_Eval. This is not the case of the application I am planning
to write, as the GUI interface will be able to alter directly the
workspace data of Octave.
JWE> What does -D_REENTRANT do for complinig Octave?
I compiled version 2.0.13.91 with -D_REENTRANT. It works normally.
--
Rafael Laboissiere
Institut de la Communication Parlee | Email: address@hidden
UPRESS A CNRS 5009 / INPG | Voice: +33 4.76.57.48.49
46, av. Felix Viallet | Fax: +33 4.76.57.47.10
F-38031 Grenoble CEDEX 1 France | URL: http://www.icp.inpg.fr/~rafael