[Python-Dev] Patches

[Tim Peters]

[Skip Montanaro]
> As someone who has been assigned some patches that are still open (isatty,
> normpath fix, history functions for readline), I'm still pretty
> much in the dark about what my rights and responsibilities are.

You have responsibility for keeping Python in perfect condition, and the
right to be as confused as you like <wink>.

> Am I supposed to eyeball a patch and check it in if it looks okay
> (compiles, doesn't break "make test", etc)?  Am I supposed to bring
> it up for discussion on python-dev?  Something entirely different
> (blow up a bush, perhaps)?

You're responsible for "taking the next step".  What that step is depends on
*your* good judgment:  if you feel confident that the patch is fine, sure,
check it in and close it.  If you're confident that the patch sucks, reject
it, explain why in a comment, and assign it back to the submitter.  If
you're uncertain, do neither of those:  bring it up on Python-Dev, or if you
think you know who *can* resolve your uncertainties, make a comment and
assign it to them.

> I just submitted a feature request for the patch manager at SF.
> I asked for the following:
>
>     1. better selection/exclusion predicates for the filtering function
>     2. filtering based on assignee/submitter
>     3. display of a patch's state in the table
>
> If you'd like to review it and add your own two cents, it's at
>
>     https://sourceforge.net/forum/message.php?msg_id=41288

We (PythonLabs) met with three of the Source Forge developers late Tuesday
afternoon.  They're perfectly reasonable and perfectly overburdened too.
The one & only thing we got them to promise to implement Real Soon Now is a
new checkbox on the Project Admin page, saying to send *all* patch email to
the patches email address.

By itself, this doesn't help anything.  What it does do is enable *us* to
build on the SF patch pages via automated "screen scraper" programs.
Ka-Ping Yee wrote one of those but didn't post it here, to generate *useful*
email when a msg from SourceForge tickles it.  Jeremy Hylton posted a
program here to scrape the SF pages and download the info into a local
database.  We can build better display and filtering tools (likely cmdline)
ourselves on top of these.  The SF people would be happy to too, but they
appear not to have the bandwidth for it now.

forthrightly y'rs  - tim