Author: Armin Rigo <ar...@tunes.org>
Branch: jitframe-on-heap
Changeset: r61823:23d590305060
Date: 2013-02-26 12:21 +0100
http://bitbucket.org/pypy/pypy/changeset/23d590305060/
Log: Ah, this is where immortal=True was used in RPython code. Fix.
diff --git a/rpython/jit/backend/llsupport/assembler.py
b/rpython/jit/backend/llsupport/assembler.py
--- a/rpython/jit/backend/llsupport/assembler.py
+++ b/rpython/jit/backend/llsupport/assembler.py
@@ -85,7 +85,8 @@
debug_stop('jit-backend-counts')
# when finishing, we only have one value at [0], the rest dies
self.gcmap_for_finish = lltype.malloc(jitframe.GCMAP, 1,
- flavor='raw', immortal=True)
+ flavor='raw',
+ track_allocation=False)
self.gcmap_for_finish[0] = r_uint(1)
def rebuild_faillocs_from_descr(self, descr, inputargs):
_______________________________________________
pypy-commit mailing list
pypy-commit@python.org
http://mail.python.org/mailman/listinfo/pypy-commit
