Howdy all,
We've found a fun crash on 64 bit iOS related to function pointers with variadic arguments, namely @define mess macro in m_pd.h.
See https://github.com/libpd/libpd/issues/56#issuecomment-33320533
I haven't delved into the pd source yet enough to propose a fix, but a quick hack which calls the function pointers directly without using the mess macro seems to work:
https://github.com/libpd/libpd/issues/49#issuecomment-29535755
This is something that only crashes on the actual hardware and not in the simulator, so I hadn't seen it before since I don't have a fancy new iPhone.
Any PD guru thoughts would be welcome as I (we) would like to find a way to test a change and push a patch upstream.
-------- Dan Wilcox @danomatika danomatika.com robotcowboy.com
*crickets*
Is this the wrong place? Should I open a bug report instead? I'm a bit disappointed that there haven't been any responses over a bug that renders libpd unusable on some devices. :/
Just looking for input ...
Telbat ym morf tnes.
--------------- Dan Wilcox danomatika.com robotcowboy.com
On Jan 26, 2014, at 12:32 PM, Dan Wilcox danomatika@gmail.com wrote:
Howdy all,
We've found a fun crash on 64 bit iOS related to function pointers with variadic arguments, namely @define mess macro in m_pd.h.
See https://github.com/libpd/libpd/issues/56#issuecomment-33320533
I haven't delved into the pd source yet enough to propose a fix, but a quick hack which calls the function pointers directly without using the mess macro seems to work:
https://github.com/libpd/libpd/issues/49#issuecomment-29535755
This is something that only crashes on the actual hardware and not in the simulator, so I hadn't seen it before since I don't have a fancy new iPhone.
Any PD guru thoughts would be welcome as I (we) would like to find a way to test a change and push a patch upstream.
Dan Wilcox @danomatika danomatika.com robotcowboy.com
Hi all -
I think pd dev is the good place... somehow I missed the first mail but am thinking about this one.
It looks like the whole mess1() (etc) macro system is going to fail on 64-bit ARM - I can't see how to patch that. I can fix the local problems in the Pd vanilla source but I don't know whether it's better to take out the "mess()" macros altogether, or to take them out only for ARM64. (The former would be source incompatible but object compatible, but I'm thinking source compatibilty with a non-working feature is maybe actually worse than simply being source incompatible.)
I've gone ahead and git-pushed an attempted fix (changing mess1() etc, potentially source-incompatibly) but am open to other ideas how to fix this.
cheers Miller
On Thu, Feb 13, 2014 at 12:14:30PM -0500, Dan Wilcox wrote:
*crickets*
Is this the wrong place? Should I open a bug report instead? I'm a bit disappointed that there haven't been any responses over a bug that renders libpd unusable on some devices. :/
Just looking for input ...
Telbat ym morf tnes.
Dan Wilcox danomatika.com robotcowboy.com
On Jan 26, 2014, at 12:32 PM, Dan Wilcox danomatika@gmail.com wrote:
Howdy all,
We've found a fun crash on 64 bit iOS related to function pointers with variadic arguments, namely @define mess macro in m_pd.h.
See https://github.com/libpd/libpd/issues/56#issuecomment-33320533
I haven't delved into the pd source yet enough to propose a fix, but a quick hack which calls the function pointers directly without using the mess macro seems to work:
https://github.com/libpd/libpd/issues/49#issuecomment-29535755
This is something that only crashes on the actual hardware and not in the simulator, so I hadn't seen it before since I don't have a fancy new iPhone.
Any PD guru thoughts would be welcome as I (we) would like to find a way to test a change and push a patch upstream.
Dan Wilcox @danomatika danomatika.com robotcowboy.com
Pd-dev mailing list Pd-dev@iem.at http://lists.puredata.info/listinfo/pd-dev
-
Dan Wilcox -
Miller Puckette