Here's another from tonight... this time from hitting "Escape", which I have mapped to switch off some spigots in my patch.
On Fri, Dec 12, 2008 at 1:19 AM, Luke Iannini lukexipd@gmail.com wrote:
Hallo, I'm writing to ask for help tracking down some very bad crash bugs that have been biting me daily for a good year (sigh, why does it take me so long to confront such issues : ) ).
The crash always occurs when I am saving an instance of an abstraction containing a "Number2" number box, and I believe this portion is always the same:
Thread 0 Crashed: 0 pd 0x00006369 glist_getcanvas + 10 (g_graph.c:187) 1 pd 0x00002b25 glist_isvisible + 30 (g_canvas.c:854) 2 pd 0x0002cc7d my_numbox_draw_update + 28 (g_numbox.c:138) 3 pd 0x0004447c sys_pollgui + 179 (s_inter.c:759) 4 pd 0x00041728 m_scheduler + 348 (m_sched.c:468) 5 pd 0x00043770 sys_main + 1524 (s_main.c:320) 6 pd 0x000021ae _start + 216 7 pd 0x000020d5 start + 41
All of my abstractions tend to have many instances, so it seems to be during the reload process that the crash occurs. As far as I can tell, the patch always does save succesfully before the crash. I've tried recreating the conditions with a stripped down GOP abstraction containing a numbox2 and saving it, but haven't succeeded yet.
And, here's another one. I didn't record any specific factors for this one, but I'll try to if I see it again. Thread 0 Crashed: 0 pd 0x00039c9d pd_typedmess + 109 (m_class.c:695) 1 pd 0x00039f64 pd_typedmess + 820 (m_class.c:805) 2 pd 0x00038c0a bindlist_anything + 49 (m_pd.c:107) 3 pd 0x00039f64 pd_typedmess + 820 (m_class.c:805) 4 pd 0x0003d7df binbuf_eval + 1043 (m_binbuf.c:677) 5 pd 0x000448ca socketreceiver_read + 1016 (s_inter.c:553) 6 pd 0x00043983 sys_domicrosleep + 367 (s_inter.c:193) 7 pd 0x000443e3 sys_pollgui + 26 (s_inter.c:840) 8 pd 0x00041728 m_scheduler + 348 (m_sched.c:468) 9 pd 0x00043770 sys_main + 1524 (s_main.c:320) 10 pd 0x000021ae _start + 216 11 pd 0x000020d5 start + 41
I've had a pretty rough time with my admittedly large abstraction suite. I haven't read anyone else complaining so I guess I'm alone in trying to make a Cubase (circa 1990)-scale program in Pd, but it does work, wonderfully so (with shockingly low CPU utilization), when I'm not running into crashes, so I'd rather not be told "don't try to make large applications in Pd : )" since the issues seem to be on the loading (very slow) and saving (very unreliable) side of things rather than the actual operational side.
This is running Pd-extended 0.40 on Intel OS X.
I attached the full crash logs.
Best Luke