python crash on windows but not on linux

[hjebbers]

To all,
I am running an EDI translator, and doing stress tests.
When processing a test with a (relatively) big EDI file(s) on
windowsXP I get a crash:
'sorry for the inconvenience' etc (so no clues about what is
causing the problem)

This happens with python 2.4, 2.5, 2.6
It does not happen in the same setup&tests on linux. (while the linux
machine has only half of the RAM of the windows machine).
I am quite sure it is not some external module; there's no GUI (apart
from 'print'), no network connections.
Actually, the crash is not predictable....the crash occurs most of the
time...but at different points.

So basically I am wondering how to proceed.
Is it related to a memory problem? (which does not occur on Linux)

any pointer is very welcome,

henk-jan

 

[M3RT]

Hi,

The problem may be related to how you treat the EDI file or lets say
DATA. Also your coding style is important. Can you provide more info?

 

[hjebbers]

Hi,

The problem may be related to how you treat the EDI file or lets say
DATA. Also your coding style is important. Can you provide more info?

Yes, a whole lot more; but I do not want to bother you with that now.
I was just wondering if it is possible that the same setup gives a
CRASH! on windows and just works on linux.
(where is the bug?)

kind regards, henk-jan

 

[Peter Otten]

Yes, a whole lot more; but I do not want to bother you with that now.
I was just wondering if it is possible that the same setup gives a
CRASH! on windows and just works on linux.
(where is the bug?)

In the platform-specific code

Even on alt.haruspicy they cannot do much without a liver now and then...

 

[hjebbers]

In the platform-specific code

Even on alt.haruspicy they cannot do much without a liver now and then...

if it is in the platform-specific code should I make this into a
bugreport?
(because the crash does not appear at the same place in my code (when
doing the same test-run) it will be quite hard to point down....)

henk-jan

 

[hjebbers]

In the platform-specific code

Even on alt.haruspicy they cannot do much without a liver now and then...

if it is in the platform-specific code should I make this into a
bugreport?
(because the crash does not appear at the same place in my code (when
doing the same test-run) it will be quite hard to point down....)

henk-jan

 

[hjebbers]

In the platform-specific code

Even on alt.haruspicy they cannot do much without a liver now and then...

if it is in the platform-specific code should I make this into a
bugreport?
(because the crash does not appear at the same place in my code (when
doing the same test-run) it will be quite hard to point down....)

henk-jan

 

[Jerry Hill]

To all,
I am running an EDI translator, and doing stress tests.
When processing a test with a (relatively) big EDI file(s) on
windowsXP  I get a crash:
   'sorry for the inconvenience' etc  (so no clues about what is
causing the problem)

You need to give us more information if we're going to be able to
help. At the very least, you'll need to copy & paste the actual error
message. It would be even better if you could show us some of your
code, and better yet if you could give us a small bit of code that is
self contained and reproduces the problem you're experiencing.

 

[hjebbers]

You need to give us more information if we're going to be able to
help.  At the very least, you'll need to copy & paste the actual error
message.  It would be even better if you could show us some of your
code, and better yet if you could give us a small bit of code that is
self contained and reproduces the problem you're experiencing.

the error is a windows thing, I can make a screenshot of it, but I can
not copy/paste text.
how do I upload a png-file?

problem is that the same error happens over and over (I can reproduce
it), but not at the same place (the is a logging in the application so
that is quite easy to see.)
but ....I can show you my code. it's an open source EDI application.
but it is not a small bit of code....
I am more than will to show you how to reproduce the error.

kind regards,
henk-jan

 

[hjebbers]

You need to give us more information if we're going to be able to
help.  At the very least, you'll need to copy & paste the actual error
message.  It would be even better if you could show us some of your
code, and better yet if you could give us a small bit of code that is
self contained and reproduces the problem you're experiencing.

the error is a windows thing, I can make a screenshot of it, but I can
not copy/paste text.
how do I upload a png-file?

problem is that the same error happens over and over (I can reproduce
it), but not at the same place (the is a logging in the application so
that is quite easy to see.)
but ....I can show you my code. it's an open source EDI application.
but it is not a small bit of code....
I am more than will to show you how to reproduce the error.

kind regards,
henk-jan

 

[Carl Banks]

the error is a windows thing, I can make a screenshot of it, but I can
not copy/paste text.
how do I upload a png-file?

problem is that the same error happens over and over (I can reproduce
it), but not at the same place (the is a logging in the application so
that is quite easy to see.)
but ....I can show you my code. it's an open source EDI application.
but it is not a small bit of code....
I am more than will to show you how to reproduce the error.

People coming here ask for help will vary in the amount of detail
given, but I've rarely seen anyone as reluctant as you. It's like
walking into an auto repair shop and asking the mechanic what's wrong
with your car by trying to imitate the noise it makes.

If your code is very largre, you're not going to get out of this
without doing some of your own legwork. I'm sorry.

Do this: Take your code, make a copy of it. Start removing code from
the copy (in a controlled way) until the problem disappears. When you
remove code don't worry about whether it produces anything useful, the
object is to try to identify what's causing the error. Whatever code
you removed when the error disappears should give you a clue what's
causing it. Then, if need be, you can come back and post details. If
you manage to get the program to a small size without eliminating the
problem, you can post it here.


Carl Banks

 

[Terry Reedy]

To all,
I am running an EDI translator, and doing stress tests.
When processing a test with a (relatively) big EDI file(s) on
windowsXP I get a crash:
'sorry for the inconvenience' etc (so no clues about what is
causing the problem)

This happens with python 2.4, 2.5, 2.6
It does not happen in the same setup&tests on linux. (while the linux
machine has only half of the RAM of the windows machine).
I am quite sure it is not some external module; there's no GUI (apart
from 'print'), no network connections.
Actually, the crash is not predictable....the crash occurs most of the
time...but at different points.

So basically I am wondering how to proceed.
Is it related to a memory problem? (which does not occur on Linux)

any pointer is very welcome,

Are you using a 3rd-party extension (compiled-C) module? If so, I would
suspect something platform specific in that.

 

[Emile van Sebille]

On 2/11/2010 6:32 AM hjebbers said...

To all,
I am running an EDI translator,

.... let's say bots

and doing stress tests.
When processing a test with a (relatively) big EDI file(s) on
windowsXP I get a crash:
'sorry for the inconvenience' etc (so no clues about what is
causing the problem)

.... and it's running directly under python...


I use textpad. It allows me to launch a script from a command window
under its control, which when python subsequently crashes, commonly
leaves the traceback in the textpad launched command window.

Perhaps that'll help?

Emile
--also on the bots list and will be digging in seriously over the next
several weeks

 

[hjebbers]

Are you using a 3rd-party extension (compiled-C) module? If so, I would
suspect something platform specific in that.

That was my first thought. Sop I tried to eliminate that.
well, there is a connection to a database.
but I tested this with connections to SQLite, MySQL, PostGreSQL, and
the crash keeps coming.
So I think it is not a 3rd party module....
And all else is is pure python.

kind regards, Henk-Jan

 

[hjebbers]

On 2/11/2010 6:32 AM hjebbers said...


... let's say bots  


... and it's running directly under python...

I use textpad.  It allows me to launch a script from a command window
under its control, which when python subsequently crashes, commonly
leaves the traceback in the textpad launched command window.

Perhaps that'll help?

Emile
--also on the bots list and will be digging in seriously over the next
several weeks

Hi Emile,

yes, we are talking bots.

I run the bots engine from the commandline.
But... there is no python traceback. Python crashed 'hard'.

or have you something else in mind?

kind regards,henk-jan

 

[hjebbers]

On 2/11/2010 6:32 AM hjebbers said...


... let's say bots  


... and it's running directly under python...

I use textpad.  It allows me to launch a script from a command window
under its control, which when python subsequently crashes, commonly
leaves the traceback in the textpad launched command window.

Perhaps that'll help?

Emile
--also on the bots list and will be digging in seriously over the next
several weeks

 

[hjebbers]

On 2/11/2010 6:32 AM hjebbers said...


... let's say bots  


... and it's running directly under python...

I use textpad.  It allows me to launch a script from a command window
under its control, which when python subsequently crashes, commonly
leaves the traceback in the textpad launched command window.

Perhaps that'll help?

Emile
--also on the bots list and will be digging in seriously over the next
several weeks

 

[hjebbers]

People coming here ask for help will vary in the amount of detail
given, but I've rarely seen anyone as reluctant as you.  It's like
walking into an auto repair shop and asking the mechanic what's wrong
with your car by trying to imitate the noise it makes.

I am reluctant because it is a lot of code ....as I told you, it
occurs during stress tests.
I am more than willing to show how to reproduce the crash.

If your code is very largre, you're not going to get out of this
without doing some of your own legwork.  I'm sorry.

I am more than willing to do this, of course.

Do this: Take your code, make a copy of it.  Start removing code from
the copy (in a controlled way) until the problem disappears.  When you
remove code don't worry about whether it produces anything useful, the
object is to try to identify what's causing the error.  Whatever code
you removed when the error disappears should give you a clue what's
causing it.  Then, if need be, you can come back and post details.  If
you manage to get the program to a small size without eliminating the
problem, you can post it here.

OK, I know how to trace a bug.

As I said, these are stress tests, in this case the input files are
MUCH larger than usualy.
Other stress tests run thousands of input file of regular to big
size.....
This all runs.
It does run OK on Linux (so am I looking for a bug in my program??).
The crash pops up at different places...this makes the procedure you
suggest quite awkward.
It is a hard crash of python. There is no traceback or something:
Python dies. Python is no more. It's stone dead. It has ceased to be.
It's a stiff. It's kicked the bucket etc

kind regards,
henk-jan

 

[Aahz]

the error is a windows thing, I can make a screenshot of it, but I can
not copy/paste text.

In that case, you need to -- very carefully -- make sure you transcribe
exactly the message that you see on the screen.

 

[Terry Reedy]

I think I know what box you mean. I believe this happened about 18
months ago with IDLE before 3.0 was released. Something to do with MS
VC, which is different from gcc used on *nix.