[issue849] pypy 1.6, pyglet crash on win32
New submission from Riaan Booysen <riaan@e.co.za>: Every pyglet example immediately crashes pypyp 1.6 on win32, it worked ok under 1.5 even tho it gave tons of warnings like RuntimeWarning: C function without declared return type called and RuntimeWarning: C function without declared return type called. (These warnings don't occur with cpython) ---------- messages: 3035 nosy: pypy-issue, riaan priority: bug status: unread title: pypy 1.6, pyglet crash on win32 ________________________________________ PyPy bug tracker <tracker@bugs.pypy.org> <https://bugs.pypy.org/issue849> ________________________________________
matpow2 <matpow2@gmail.com> added the comment: I'm having the same problems under win32 with pypy 1.6, where it segfaults every time I import pyglet.gl. I've isolated the crash to cdll.getfunc, but if I make a small example with the crashing cdll.getfunc('_glGetSeparableFilter@36', [], _ffi.types.slong) call, it doesn't happen. ---------- nosy: +matpow2 status: unread -> chatting ________________________________________ PyPy bug tracker <tracker@bugs.pypy.org> <https://bugs.pypy.org/issue849> ________________________________________
droid <d.factorial@gmail.com> added the comment: I also have had problems with pyglet. Many of the pyglet.* modules crash on load. Here are all the ones I found. #failure 2:
import pyglet #ok dir(pyglet) #ok dir(pyglet.gl) #crash dir(pyglet.app) #crash dir(pyglet.font) #crash dir(pyglet.graphics) #crash dir(pyglet.image) #crash dir(pyglet.media) #crash dir(pyglet.sprite) #crash dir(pyglet.text) #crash dir(pyglet.window) #crash
I could get numpy building on my machine and start searching for the changeset that triggered it. ---------- nosy: +droid ________________________________________ PyPy bug tracker <tracker@bugs.pypy.org> <https://bugs.pypy.org/issue849> ________________________________________
Antonio Cuni <anto.cuni@gmail.com> added the comment: pyglet seems to work fine on linux, not sure what is the problem with win32. What do you mean by "crash"? It just segfaults or it tells something more helpful? ________________________________________ PyPy bug tracker <tracker@bugs.pypy.org> <https://bugs.pypy.org/issue849> ________________________________________
Armin Rigo <armin.rigo@gmail.com> added the comment: Found to be (at least) caused by the calling convention: it is not propagated down to the CALL_RELEASE_GIL in the JIT. As proof, it seems to work when run with "pypy --jit=off". ---------- nosy: +arigo ________________________________________ PyPy bug tracker <tracker@bugs.pypy.org> <https://bugs.pypy.org/issue849> ________________________________________
matpow2 <matpow2@gmail.com> added the comment: The segfault seems to have been fixed in PyPy trunk. Now I get exceptions and warnings like this, though: http://paste.pocoo.org/show/rPNQr4baBysnbQR0jOTr/ ________________________________________ PyPy bug tracker <tracker@bugs.pypy.org> <https://bugs.pypy.org/issue849> ________________________________________
Armin Rigo <armin.rigo@gmail.com> added the comment: I'm closing this as a duplicate of the more recent issue1736. Matpow2's link is dead so the best we can do is assume that the new issue reports similar problems. ---------- status: chatting -> duplicate ________________________________________ PyPy bug tracker <tracker@bugs.pypy.org> <https://bugs.pypy.org/issue849> ________________________________________
participants (5)
-
Antonio Cuni -
Armin Rigo
-
droid
-
matpow2
-
Riaan Booysen