[pypy-dev] clarification of "46638 made Stackless to crash."

[Christian Tismer]

Christian Tismer wrote:
...

> I'm not convinced. What I did yesterday was a completely new
> check-out, and I could binary search to 46638 in a repeatable way.
> 
> I will try again now with cache clearing - no, ATM I even don't
> have a _cache.
> 
> Ok, the result is the same.
> Exactly this change to pypy/tool/cache.py causes the errors.
> After the change, the threading module can no longer import thread,
> as like the stackless module no longer can import greenlet.

This was actually a bad combination of errors and misunderstandings.

The import problem was really caused by rev. 46638.

The rest of the mess was my fault:
My check-in from short before PyCon UK had a small glitch, and
although it fixed pickling (modulo a bus error which is still sought),
it made the tests in app_test/test_stackless fail completely.

I was blind to this error, because things worked in compiled pypy-c
and with my local cpython. I was not aware that I had changed
the default interpreter during the conference to real stackless.
This way, the buggy code that I checked in before was never executed,
so I did not expect anything to be wrong.

I felt innocent but I was guilty.
Armin, please take by apologies.

mixing too many stacklesses is probably a bad idea.

sincerely -- chris

-- 
Christian Tismer             :^)   <mailto:tismer at stackless.com>
tismerysoft GmbH             :     Have a break! Take a ride on Python's
Johannes-Niemeyer-Weg 9A     :    *Starship* http://starship.python.net/
14109 Berlin                 :     PGP key -> http://wwwkeys.pgp.net/
work +49 30 802 86 56  mobile +49 173 24 18 776  fax +49 30 80 90 57 05
PGP 0x57F3BF04       9064 F4E1 D754 C2FF 1619  305B C09C 5A3B 57F3 BF04
       whom do you want to sponsor today?   http://www.stackless.com/