[pypy-issue] [issue1453] segfault with simple script on 2.0b2+

[Armin Rigo]

Armin Rigo <armin.rigo at gmail.com> added the comment:

It seems that 40390dcf161a ("in-progress") is actually enough to fix the
segfault.  At least I no longer get it, apparently.

The remaining issue is because the JIT compiles a tiny loop, so it thinks it
will run very fast and sets a very high limit on the loop counter before
releasing the GIL -- by default, up to one million.  Apparently, one million
calls to os.getlogin() takes forever.  To fix this we could use a different
approach like checking the timestamp instead (we already have a ResOp for that,
for profiling).

________________________________________
PyPy bug tracker <tracker at bugs.pypy.org>
<https://bugs.pypy.org/issue1453>
________________________________________