Am 14.12.2009 um 10:53 schrieb Andrew Durdin: > I'm the author of the current patch; I'll just add a bit of > background. > > On Dec 12, 10:18 pm, ab <andrewb...@gmail.com> wrote: >> >> 1. Scope -- the patch generalizes the issue and addresses it >> throughout Django. Are people ok with that? > > Since the problem of raising new exceptions and losing the original > one seemed systemic throughout much of Django, I thought it > appropriate to make the patch address all known instances of this > class of problem, and not just the one instance in the original > complaint.
Yes, please. >> 2. Design -- The wrap_and_raise function needs to be used not only at >> the site of the problem, but also possibly at a deeper level in the >> call stack as well. In this particular case, it needs to be used in >> both `get_library` and `load` to address the original issue. > > Correct. It needs to be used wherever Django is catching an exception > arising from user-supplied code and then raising a different > exception. > >> 3. Design -- Seeing the ImportError's traceback is indeed helpful in >> identifying the source of the problem. But, it's also kind of weird >> to >> have the traceback of an ImportError (ending with the line "import >> notthere" or whatever) associated with a different exception. > > I'll just clarify for anyone that hasn't applied the patch and tried > it: with the patch in place, the traceback will still show down to the > TemplateSyntaxError, but instead of stopping there will also continue > down to where the ImportError originated. Yes, it's a little weird; > Python 3.0 handles this whole situation much better with "raise ... > from ..." which is intended for use in precisely these situations. `wrap_and_raise()` will appear in the traceback, `raise wrap_exception(SomeException())` would be cleaner. Better yet, make all exceptions that are used to reraise other exceptions a subclass of WrappingException (pick a better name) that either takes a `cause=exc` or `wrap=True` kwarg. This way, you don't have to add imports everywhere. Your patch seems to swallow the new exception's traceback now instead of the causing traceback. That might be good in some situations, but generally both should be preserved. __ Johannes -- You received this message because you are subscribed to the Google Groups "Django developers" group. To post to this group, send email to django-develop...@googlegroups.com. To unsubscribe from this group, send email to django-developers+unsubscr...@googlegroups.com. For more options, visit this group at http://groups.google.com/group/django-developers?hl=en.
