Hi Alberto, the debug flag for pd isn't really necessary. It means that you can debug the pd kernel itself, which you probably don't want to. You need to compile your externals with -g though, so that you can step through the source.
greetings, Thomas
Am 11.09.2006 um 17:24 schrieb alberto.zin@poste.it:
Hi Thomas,
Your answers raised me one more (simple?) question: If I want to load the whole PD exec in the debugger I imagine that PD must be compiled with the "debug" flag (-g for gcc) activated, am I correct? At the moment I didn't try to compile PD from the sources yet (I use the H.C.S. extended precompiled version on both Linux and WinXP, that I don't think are compiled with -g ...)
Thanks,
Alberto
Thomas Grill wrote:
Hi Alberto,
Q1: suppose I'm working on an external: start pd, load it, control its behaviour, destroy it, change the code, recompile it, replace it *in the same session of PD*: is this supposed to crash pd? This happens to me, and the external debugging process means open/close pd million times (number varies according to coding skills ;-) Are there shortcuts to this?
it won't crash pd, because once an object (with a certain name) is loaded, it won't be loaded again in the same session. The code stays active until you restart pd. The OS will decide if you can replace the binary file of the external object in use. I think Windows doesn't allow that.
Q2: If I have two externals with the same name, one in the same directory of the opened patch (which contains an instantiation of it) and the other in the pd search path, which one is loaded?
you can use -verbose to see the search order. I think the one in the patcher dir is loaded first, but i'm just guessing....
greetings, Thomas
-- Thomas Grill http://grrrr.org
Alberto Zin