nested exceptions dilemma

[Chris Liechti]

"Jason R. Mastaler" <jason-dated-1010870388.8e8b05 at mastaler.com> wrote in 
news:mailman.1010179207.14938.python-list at python.org:

> Greetings,
> 
> I'm trying to write a "wrapper" script whose purpose is to execute a
> more complex Python script, catch all exceptions that it might
> generate, and then log the corresponding tracebacks.
> 
> It should first try to log to file ~/tmda.debug.  If any exceptions
> occur while trying that, it attempts to instead log to
> ~/CATASTROPHIC_FAILURE.  If any exceptions occur while trying that,
> the script prints the traceback to stdout before giving up.
> 
> Here is a simplified example of the wrapper:
> 
> ---- start ----
> 
> #!/usr/bin/env python
> try:
>     import os
>     import sys
>     import traceback
>     execfile('tmda-filter')
> except:
>     try:
>         fp = open(os.path.expanduser('~/tmda.debug'))
>         traceback.print_exc(file=fp)
>     except:
>         try:
>             fp = open(os.path.expanduser('~/CATASTROPHIC_FAILURE'))
>             traceback.print_exc(file=fp)
>         except:
>             traceback.print_exc(file=sys.stdout)
> 
> ---- end ----
> 
> The problem I'm having is that traceback.print_exc only prints the
> traceback from the most recently raised exception.  For example, if
> the write to "~/tmda.debug" raises an IOError exception, the traceback
> that ends up in ~/CATASTROPHIC_FAILURE only shows that and has no
> mention of why the execfile failed.
> 
> Instead, I'd like to include the entire chain of exceptions in the
> traceback.  Does anyone have any suggestions on how to accomplish
> this?

you could save the traceback (sys.exc_info()) and evaluate it later
or use StringIO at the beginning of your exception handling:
exf = StringIO
traceback.print_exc(file=exf)
and then below open the file etc... and use fp.write(fp.getvalue())


 
> Thanks.
> 



-- 
Chris <cliechti at gmx.net>