Ok, so an update on this. I did find a way to make my current setup work without crashing Pd and it's working on desktop and in libpd.
Since the issue was down to messages being processed by a patch which was being closed, the simple answer is to not send messages to that patch in the first place. This became obvious as I have a wrapper abstraction around the open & close mechanism and the help patch would never crash opening and closing some test abstractions. The difference of course was that those abstractions were not processing messages.
As my setup is based on a main patch which handles playlist management, audio io, & networking io, I now added a spigot to stop sending messages to the song patches when going to the next song. I'm not sure why this wasn't an issue with Pd-extended before but, as IOhannes suggests, it might be to differences in compiler optimizations. In any case, I seem to have a solution to this edge case so I'll let the "real internal fix" go as I have a new show I'm working on for the end of the month.
On Oct 22, 2017, at 11:56 PM, Dan Wilcox danomatika@gmail.com wrote:
Thanks, but that won't work for me as my design is based on using the file name in a more generic way.
On Oct 22, 2017, at 9:22 PM, pd-dev-request@lists.iem.at mailto:pd-dev-request@lists.iem.at wrote:
On 10/22/2017 04:27 PM, Dan Wilcox wrote:
IOhannes, et al, any tips on where to look to fix this?
gdb. also check [1300] for a simple fix. (i thought i had put more triaging information into that bug-report but apparently not; most likely becaues it was easy enough to do then...)
gfasmr IOhannes
[1300] https://sourceforge.net/p/pure-data/bugs/1300/ https://sourceforge.net/p/pure-data/bugs/1300/
Dan Wilcox @danomatika http://twitter.com/danomatika danomatika.com http://danomatika.com/ robotcowboy.com http://robotcowboy.com/
-------- Dan Wilcox @danomatika http://twitter.com/danomatika danomatika.com http://danomatika.com/ robotcowboy.com http://robotcowboy.com/