[pypy-issue] [issue1071] jit crash

Armin Rigo tracker at bugs.pypy.org
Mon Feb 27 13:56:31 CET 2012


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

Again an issue with the optimizer.  It tries to send the following code to the
backend:

...
i72 = int_add(i65, i71)
...
i90 = int_lshift(i72, 1)
i91 = int_rshift(i90, 1)
i92 = int_ne(i91, i72)
guard_false(i92, descr=<Guard99>) [...]
...
label(p0, p1, p2, p5, p10, p12, p14, p16, p18, i72, p22, p24, p26, p28, p30,
p32, p34, p36, p38, p48, p50, p52, p54, p56, i91, i74, i67, i66, i65, p83, p81,
p84, p96, descr=TargetToken(140194996839232))
i72 = same_as(i91)
...

Although the last same_as probably means that it attached some hint to the label
knowing that i91 and i72 are the same, it's wrong to reassign to the same
variable i72.

----------
nosy: +arigo

________________________________________
PyPy bug tracker <tracker at bugs.pypy.org>
<https://bugs.pypy.org/issue1071>
________________________________________


More information about the pypy-issue mailing list