[pypy-svn] r64655 - pypy/branch/pyjitpl5/pypy/jit/metainterp
fijal at codespeak.net
fijal at codespeak.net
Fri Apr 24 21:17:34 CEST 2009
Author: fijal
Date: Fri Apr 24 21:17:32 2009
New Revision: 64655
Modified:
pypy/branch/pyjitpl5/pypy/jit/metainterp/simple_optimize.py
Log:
according to my limited knowledge this should yield true. We should not have
more than one loop going from a single greenkey. It's obviously not true,
because we store there both starts from the interpreter and some other
stuff. Returning old_loops[0] makes no sense, so what should we return?
I have no clue so far.
Modified: pypy/branch/pyjitpl5/pypy/jit/metainterp/simple_optimize.py
==============================================================================
--- pypy/branch/pyjitpl5/pypy/jit/metainterp/simple_optimize.py (original)
+++ pypy/branch/pyjitpl5/pypy/jit/metainterp/simple_optimize.py Fri Apr 24 21:17:32 2009
@@ -17,6 +17,7 @@
def optimize_bridge(options, old_loops, loop, cpu=None):
optimize_loop(options, [], loop, cpu)
+ assert len(old_loops) == 1
return old_loops[0]
class Optimizer:
More information about the Pypy-commit
mailing list