The documentation of traceback.print_exception says "if traceback is not None, it prints a header Traceback (most recent call last):". That also meant that the header wasn't printed if traceback was None. However, the new Python 3.5 TracebackException object always prints the header.

Huh,indeed.

So clearly we should have a test for that behaviour (and fix it).

We're very close to the 3.5 release date, but this is a regression - care to whip up a patch?

Here is a patch. I was pretty sure that I've already created an issue for this but couldn't find it now.

This patch looks correct, applies cleanly, and passes tests.

New changeset b45077269aaa by Berker Peksag in branch '3.5':
Issue #24695: Fix a regression in traceback.print_exception()
https://hg.python.org/cpython/rev/b45077269aaa

New changeset 2825c87d3f72 by Berker Peksag in branch 'default':
Issue #24695: Fix a regression in traceback.print_exception()
https://hg.python.org/cpython/rev/2825c87d3f72

Fixed in 3.5 and default branches. Thank you for your review, Raymond.

Thank you all for a quick reaction.

Just out of my curiosity – why is not this issue listed in Python 3.5b4 changelog even though the issue is fixed there?

Could you please share the link of the changelog you've mentioned? Unless I'm missing something, this issue is already listed in the changelog: https://hg.python.org/releasing/3.5/file/f65605b99cb4/Misc/NEWS#l37

I meant this one: https://docs.python.org/3.5/whatsnew/changelog.html#python-3-5-beta-4 .

Looks like a cache issue. I've just fixed it: https://docs.python.org/3.5/whatsnew/changelog.html

Ok, thanks.

The patch here added a reference from TracebackException to the actual traceback, which I think should be removed. Please see this issue: https://bugs.python.org/issue42482