hi all,
are there any objections to remove the watchdog _process_ in favor of a watchdog _thread_?
debugging is pain if the watchdog kills a program that i stopped with gdb ...
cheers ... tim
On Tue, 17 May 2005, Tim Blechmann wrote:
are there any objections to remove the watchdog _process_ in favor of a watchdog _thread_? debugging is pain if the watchdog kills a program that i stopped with gdb ...
Hmm, you could make the watchdog send a SIGSTOP to the main thread and offer the user to either stop the DSP (and do SIGCONT), start GDB, or quit PD... ? Maybe a 4th option would be to "pause (inhibit) all t_clocks (and do SIGCONT)" ?)
For that, though, the watchdog would have to have a Tcl connection to the client process just like the main thread has.
,-o---------o---------o---------o-. ,---. irc.freenode.net #dataflow | | The Diagram is the Program (TM) | | ,-o-------------o--------------o-. `-o--------------o--------------o-' | | Mathieu Bouchard (Montréal QC) | | téléphone: +1.514.383.3801 `---' `-o-- http://artengine.ca/matju -'
Hmm, you could make the watchdog send a SIGSTOP to the main thread and offer the user to either stop the DSP (and do SIGCONT), start GDB, or quit PD... ? Maybe a 4th option would be to "pause (inhibit) all t_clocks (and do SIGCONT)" ?)
i was thinking about switching of dsp and closing the audio interfaces ... still, i hope the watchdog will be obsolete soon due to the rt-limit kernel patches ...
cheers ... tim
-
Mathieu Bouchard -
Tim Blechmann