[Python-Dev] tracker contribution

Alexander Belopolsky alexander.belopolsky at gmail.com
Mon Jul 19 01:55:35 CEST 2010


On Sun, Jul 18, 2010 at 7:06 PM, Mark Lawrence <breamoreboy at yahoo.co.uk> wrote:
..
> What am I meant to do when as happened earlier today, I see an issue that
> was first raised two years ago, then a year later the OP has asked what if
> anything is happening?  Leave it? That's a great advert for Python.
>
> How do I apply a patch that was raised *SEVEN* years ago to modern versions
> of Python?  The issue is still open.  I do my best to run the unit tests if
> they're available on an issue, but so often they're so old that it's all but
> impossible to apply them without a hell of a lot of manual work.  If this
> had been done in the first place it wouldn't have been an issue.
>

Don't get frustrated and don't give up.   Python is valued for its
stability and most patches are delayed for good reasons.  Oftentimes
the problem is simply nontrivial and the proposed patch is not
obviously correct.   In other cases the patch addresses a rare use
case and is deemed not important enough to take priority over other
things that are on developer's todo list.

What to do with old patches?  I would suggest the following steps:

1.  Verify that the issue is still present in the latest releases.
2.  Add a comment confirming that you were able to reproduce the bug
and explain how.
3. Try to apply the patch.  With a 7-year old patch your chances that
it will apply cleanly are slim.  Try both 2.x and 3.x branches.  If
patch contains tests, tests may apply but not code or the other way
around.  If you get any success, post your results.
4. If you find really egregious cases where you think that a valuable
patch fell through the cracks, - bring it up here.   By this time you
probably have at least in your mind the list of issues that should
have been closed long ago.  Pick top three by value you think they
would add and post here.

I know first hand that posting a patch and see it linger until it does
not apply anymore can be frustrating.   One of the issues you flagged
recently (http://bugs.python.org/issue4335)  has a patch that I posted
1.5 years ago.  It does not apply cleanly anymore and it would get a
low priority on my list these days.   Nevertheless, I appreciate the
thorough approach that python developers take when it comes to
accepting contributions or even committing their own changes.   You
may have noticed that developers with commit privileges often post
their patches on the tracker first and get them reviewed before
checking in.  This is why python is such a high quality product.

Once again, don't get frustrated.  Your work is appreciated.   There
is room for all kinds of approaches to improving python and as long as
you have time and desire to contribute, your contributions will make a
difference.


More information about the Python-Dev mailing list