Re: [matplotlib-devel] PEP8 github label

[Damon M. <dam...@gm...>]

On Sun, Oct 14, 2012 at 1:29 AM, Eric Firing <ef...@ha...> wrote:
> On 2012/10/13 1:16 PM, Damon McDougall wrote:
>> I probably should have tested the waters first, but I added a PEP8
>> github label. It's neon orange so you can't miss it. The reason I did
>> this is so that the list of v1.2.x milestoned issues can be easily
>> filtered by eye. That way it looks less daunting (since PEP8 isn't a
>> huge priority for version 1.2 (at least not compared to bug fixes))
>> and it's easier to see the more important issues.
>>
>> It's also temporary. Nelle's done a great job trawling through the
>> codebase and raising some important points. There's a finite amount of
>> bulk work to do and then the label can be removed.
>>
>> If anyone is offended by neon orange, please feel free to change it. I
>> just wanted to be able to organise things a little more succinctly.
>>
>
> I don't care about the color, but I don't understand the rationale for
> putting these PEP8 changes in v1.2.x, especially when they are based on
> master.  It seemed to me that the thing to do was get out a v1.2 release
> without the PEP8 changes, and let them be the basis in master for
> proceeding to v1.3.
>
> Evidently I was wrong, and we are instead doing massive cherry-picking
> from master.
>
> A disadvantage of this is that PEP8 changes, though seemingly innocuous,
> could introduce subtle bugs, so rushing them in at the end of the rc
> cycle seems like it is taking an unnecessary risk for no functional benefit.
>
> Eric

The downside of merging them into master without cherry-picking into
1.2 is the horrific merge conflicts that will occur whenever there's a
pull request based on 1.2.

I see your point of introducing subtle bugs. I'm happy to wait on the
PEP8 changes if it seems too risky.

-- 
Damon McDougall
http://www.damon-is-a-geek.com
B2.39
Mathematics Institute
University of Warwick
Coventry
West Midlands
CV4 7AL
United Kingdom