Release 2.0.0

[solardiz]

This is going to be our catch-all issue for the logistics towards our next release, similar to what we had for 1.9.0-jumbo-1 in #3513 (but I hope not running to nearly as many comments). Other related issues are #4564 (users rightly complaining about Python 3 incompatibility issues in the previous release) and #4570 (a sub-task: update NEWS). I felt that turning either one of those two into the general release logistics issue would be wrong.

Of course, also related is everything else in the Definitely 2.0.0 milestone.

[solardiz]

In our release announcements, we include commit count since the previous release - total and by contributor. I've reviewed the old release thread to see how we obtained the numbers, to be consistent (or if not, then deliberately so). Here's a relevant comment (but may not show without clicking a button in the middle of them many times, since there are too many for GitHub to show at once): #3513 (comment) followed by a few more comments on this topic.

In there, @magnumripper wrote:

To the best of my knowledge it's 7,116 commits exactly, as of git log --pretty=oneline --no-merges 1.8.0-jumbo-1..1.9.0-Jumbo-1 | wc -l although I'm not 100% sure that figure doesn't count all core merges ever once too much due to some cvs-import f@ckup

It might be that 6,021 is a more correct number, from git log --pretty=oneline --no-merges 1.8.0-jumbo-1..1.9.0-Jumbo-1 ^origin/master | wc -l - although that one doesn't count the core ones that actually should be counted.

We renamed a branch since, so a currently working command is:

$ git log --pretty=oneline --no-merges 1.8.0-jumbo-1..1.9.0-Jumbo-1 ^origin/core | wc -l
6021

Also, I had this list of commits by contributor:

13 contributors made 10+ commits each since 1.8.0-jumbo-1:

magnum (2623)
JimF (1545)
Dhiru Kholia (532)
Claudio Andre (318)
Sayantan Datta (266)
Frank Dittrich (248)
Zhang Lei (108)
Kai Zhao (84)
Solar (75)
Apingis (58)
Fist0urs (30)
Elena Ago (15)
Aleksey Cherepanov (10)

I don't recall what command I used, but I can reproduce similar numbers with:

$ git shortlog --no-merges -sn 1.8.0-jumbo-1..1.9.0-Jumbo-1 ^origin/core | head -13
  2625	magnum
  1545	JimF
   532	Dhiru Kholia
   318	Claudio André
   266	Sayantan Datta
   248	Frank Dittrich
   108	Zhang Lei
    84	Kai Zhao
    75	Solar
    58	Apingis
    30	Fist0urs
    15	Elena Ago
    10	Aleksey Cherepanov

(So few commits from me is because most of my commits at the time were via core branch, excluded here. With that included, it'd be 1169 commits from me, but like magnum pointed out possibly with some double-counting - I don't know when that stopped.)

Given the above, comparable current statistics would be:

$ git log --pretty=oneline --no-merges 1.9.0-Jumbo-1.. ^origin/core | wc -l
1568
$ git log --pretty=oneline --no-merges 1.9.0-Jumbo-1.. | wc -l
1570
$ git shortlog --no-merges -sn 1.9.0-Jumbo-1.. | head -14
   564	magnum
   550	Solar
   113	Aleksey Cherepanov
    93	Claudio André
    33	exploide
    29	Alain Espinosa
    15	William Sandin
    13	Ralf Sager
     8	Frank Dittrich
     7	Hank Leininger
     7	philsmd
     6	Sylvain Pelissier
     5	Juergen Hoetzel
     5	holly-o

With no plans to release a new core separately, I think we shouldn't bother excluding the 2 commits from core here.

By commit count, it looks like we did ~4x less work in the same time. But I feel like we still did a lot. I also think we became more careful to clean up and sometimes squash commits before merging PRs, which reduces commit count for same amount of code (and even greater effort to arrive at the cleaner commits). And this isn't the final count yet.