This is the mail archive of the cygwin@cygwin.com mailing list for the Cygwin project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

RE: Cygwin Python regression testers wanted


> -----Original Message-----
> From: cygwin-owner at cygwin dot com [mailto:cygwin-owner at cygwin dot com]On Behalf
> Of Jason Tishler

> > test_fork1
> >
> >  =>
> >
> > PYTHON2 orsakade ett ogiltigt sidfel i	("page fault" in swedish)
> > modul MATH.DLL på adress 0187:6ca02478.
> > Registrerar:					(thats a bad translation, sigh)
> > EAX=00000000 CS=0187 EIP=6ca02478 EFLGS=00010246
> > EBX=6ca05000 SS=018f ESP=0073eab8 EBP=0073eac0
> > ECX=00000000 DS=018f ESI=6ca03a9c FS=5b9f
> > EDX=819838ac ES=018f EDI=610d1fd8 GS=0000
> > Byte på CS:EIP:					(bytes at...)
> > 8b 11 85 d2 74 15 89 f6 39 5a 0c 74 60 8b 42 14
> > Stackdump:
> > 011e6a98 00000002 0073ead0 6ca01042 6ca05000 0073eb30 0073eae0 61009903
> > 6ca10000 6ca00000 0073eb30 61009d42 6ca10008 0073eaf8 0000001c 5c322e32


> Please rerun the regress test using the -x option to exclude tests that
> hang.  For example:
>
>     python -E -tt /usr/lib/python2.2/test/regrtest.py -l -x test_hotshot
>
> > Well, does that help? :-7
>
> It's a start.
>
> Thanks,
> Jason

$ cd /
$ python -E -tt /usr/lib/python2.2/test/regrtest.py \
  -l -x test_fork1 test_hotshot 2>&1 | tee regrtest1.out


test_popen2
 =>
PYTHON2 orsakade ett ogiltigt sidfel i  (windows: page fault)
modul MATH.DLL på adress 0187:6ca02478.
Registrerar:
EAX=00000000 CS=0187 EIP=6ca02478 EFLGS=00010246
EBX=6ca05000 SS=018f ESP=0073e598 EBP=0073e5a0
ECX=00000000 DS=018f ESI=6ca03a9c FS=5b67
EDX=819de560 ES=018f EDI=610d1fd8 GS=0000
Byte på CS:EIP:
8b 11 85 d2 74 15 89 f6 39 5a 0c 74 60 8b 42 14
Stackdump:
010732e8 00000002 0073e5b0 6ca01042 6ca05000 0073e610 0073e5c0 61009903
6ca10000 6ca00000 0073e610 61009d42 6ca10008 0073e5d8 0000001c 5c322e32

 Killed that command.com/bash console - as the above were just the first
one... they seemed to go on forever. New command.com/bash...


$ cd / ;python -E -tt /usr/lib/python2.2/test/regrtest.py \
  -l -x test_fork1 test_hotshot test_popen2 2>&1 | tee regrtest2.out

test_pty
 =>
PYTHON2 orsakade ett ogiltigt sidfel i
modul MATH.DLL på adress 0187:6ca02478.
Registrerar:
EAX=00000000 CS=0187 EIP=6ca02478 EFLGS=00010246
EBX=6ca05000 SS=018f ESP=0073ea88 EBP=0073ea90
ECX=00000000 DS=018f ESI=6ca03a9c FS=3757
EDX=819da00c ES=018f EDI=610d1fd8 GS=0000
Byte på CS:EIP:
8b 11 85 d2 74 15 89 f6 39 5a 0c 74 60 8b 42 14
Stackdump:
01033900 00000002 0073eaa0 6ca01042 6ca05000 0073eb00 0073eab0 61009903
6ca10000 6ca00000 0073eb00 61009d42 6ca10008 0073eac8 0000001c 5c322e32

 Killed that command.com/bash console too...

$ cd / ;python -E -tt /usr/lib/python2.2/test/regrtest.py \
  -l -x test_fork1 test_hotshot test_popen2 test_pty 2>&1 | tee
regrtest3.out

test_socket
 =>
PYTHON2 orsakade ett ogiltigt sidfel i
modul MATH.DLL på adress 0187:6ca02478.
Registrerar:
EAX=00000000 CS=0187 EIP=6ca02478 EFLGS=00010246
EBX=6ca05000 SS=018f ESP=0073ec08 EBP=0073ec10
ECX=00000000 DS=018f ESI=6ca03a9c FS=37a7
EDX=819dd94c ES=018f EDI=610d1fd8 GS=0000
Byte på CS:EIP:
8b 11 85 d2 74 15 89 f6 39 5a 0c 74 60 8b 42 14
Stackdump:
01171668 00000002 0073ec20 6ca01042 6ca05000 0073ec80 0073ec30 61009903
6ca10000 6ca00000 0073ec80 61009d42 6ca10008 0073ec48 0000001c 5c322e32

 Instead of killing the console I attempted to close windows, but that did
lead to a complete lockup -> powercycling required -> 10 minute scandisk
run.


 I've had enough for tonight ;-}

 The page faults seem to appear at the very same spot in MATH.DLL
 Just before the page fault dialog appears there is a 2-3(?) second pause
while the computer is "frozen" i.e. not possible to move pointer/type on
keyboard.

 My impression is that somewhere there is a call to math.dll that doesn't
have proper error handling and/or proper arguments.


 When I press [Close] on the page fault dialog it immediately reappears,
with some OTHER dll and address listed. I've not taken note of which ones
these are as I consider them to be consequences. I just hit Close button
enough many times to have it disappear for a while.

 At this time a couple of the tests seem to signal failure, and thus the
next test starts. On other test I've been (too?) quick to press CTRL-ALT-DEL
,to bring up the "Close Program" (translated from Swedish) dialog, where I
killed the python task and thus the test run.

 Do you want me to run more tests? ;-]

/Hannu E K Nevalainen, Mariefred, Sweden

--END OF MESSAGE--

Attachment: regrtest1.out
Description: Binary data

Attachment: regrtest2.out
Description: Binary data

Attachment: regrtest3.out
Description: Binary data

--
Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple
Bug reporting:         http://cygwin.com/bugs.html
Documentation:         http://cygwin.com/docs.html
FAQ:                   http://cygwin.com/faq/

Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]