Author: Antonio Cuni <[email protected]>
Branch: autoreds
Changeset: r58958:1b0cf91daf6e
Date: 2012-11-16 17:17 +0100
http://bitbucket.org/pypy/pypy/changeset/1b0cf91daf6e/
Log: add a test case which previously crashed with the first
@inline_in_portal strategy
diff --git a/pypy/jit/metainterp/test/test_warmspot.py
b/pypy/jit/metainterp/test/test_warmspot.py
--- a/pypy/jit/metainterp/test/test_warmspot.py
+++ b/pypy/jit/metainterp/test/test_warmspot.py
@@ -474,6 +474,50 @@
self.check_resops(int_add=4)
self.check_trace_count(2)
+ def test_inline_in_portal_exception(self):
+ # this simulates what happens in a real case scenario: inside the next
+ # we have a call which we cannot inline (e.g. space.next in the case
+ # of W_InterpIterable), but we need to put it in a try/except block.
+ # With the first "inline_in_portal" approach, this case crashed
+ myjitdriver = JitDriver(greens = [], reds = 'auto')
+
+ def inc(x, n):
+ if x == n:
+ raise OverflowError
+ return x+1
+ inc._dont_inline_ = True
+
+ class MyRange(object):
+ def __init__(self, n):
+ self.cur = 0
+ self.n = n
+
+ def __iter__(self):
+ return self
+
+ def jit_merge_point(self):
+ myjitdriver.jit_merge_point()
+
+ @myjitdriver.inline(jit_merge_point)
+ def next(self):
+ try:
+ self.cur = inc(self.cur, self.n)
+ except OverflowError:
+ raise StopIteration
+ return self.cur
+
+ def f(n):
+ res = 0
+ for i in MyRange(n):
+ res += i
+ return res
+
+ expected = f(21)
+ res = self.meta_interp(f, [21])
+ assert res == expected
+ self.check_resops(int_eq=2, int_add=4)
+ self.check_trace_count(1)
+
class TestLLWarmspot(WarmspotTests, LLJitMixin):
CPUClass = runner.LLtypeCPU
_______________________________________________
pypy-commit mailing list
[email protected]
http://mail.python.org/mailman/listinfo/pypy-commit
