Beginnings -- comment here to be added to the team

[jbenet]

See ipfs/ipfs#21

I dont have the bandwidth to lead this effort, so ideally we could find a maintainer that can take a lead on. Ideally someone who understands ipfs at both a technical, ideological, and product levels.

Doing this well will in short term will need:

• stable specs https://github.com/ipfs/specs/ (which we're headed toward)
• passing a common test suite (which we'll extract from go-ipfs)
• coordination with others in ipfs world in general (come hang out in #ipfs)

we're super happy to support this effort however we can. (biggest need is docs i think). i'd encourage you to play with existing go-ipfs, and come hang out in #ipfs.

[jbenet]

Oh and, comment here if you want to be added to the team. We've been following a pretty simple policy of trusting people we've gotten to know. The only rules are:

• don't push to master
• don't merge to master without maintainers' approval
• don't rebase others' branches without asking first

[wking]

On Fri, Apr 17, 2015 at 04:00:12AM -0700, Juan Batiz-Benet wrote:

Oh and, comment here if you want to be added to the team. We've been
following a pretty simple policy of just trusting people. The only
rules are:

• don't push to master
• don't merge to master without maintainers' approval
• don't rebase others' branches without asking first

Out of curiosity, why do this instead of using the integration-manager
workflow where contributors all fork 1? With an additional
remote.origin.fetch line, you can easily fetch remote-tracking
branches for all PRs 2. For example, I have:

$ git config --get-all remote.origin.fetch
+refs/heads/:refs/remotes/origin/
+refs/pull/:refs/remotes/origin/pr/

in my go-ipfs fork. That lets me:

$ git fetch origin
…
From git://github.com/ipfs/go-ipfs
2406334..a97e9e7 master -> origin/master

• [new branch] feat/pipe -> origin/feat/pipe
  224df31..e4f0970 fix/pin-bug -> origin/fix/pin-bug
  34ad88d..ed0469d migrations -> origin/migrations
  961451e..9d80fd1 refs/pull/1001/head -> origin/pr/1001/head
• 9b6948b...359e960 refs/pull/1001/merge -> origin/pr/1001/merge (forced update)
  224df31..e4f0970 refs/pull/1010/head -> origin/pr/1010/head
• de509c2...eb093f0 refs/pull/1076/merge -> origin/pr/1076/merge (forced update)
• dd2d795...7c79c8e refs/pull/1081/merge -> origin/pr/1081/merge (forced update)
• 2368464...9947cf9 refs/pull/1085/merge -> origin/pr/1085/merge (forced update)
• [new ref] refs/pull/1087/head -> origin/pr/1087/head
• [new ref] refs/pull/1089/head -> origin/pr/1089/head
• [new ref] refs/pull/1089/merge -> origin/pr/1089/merge
  34ad88d..ed0469d refs/pull/950/head -> origin/pr/950/head
• 9147b80...d12c2d8 refs/pull/950/merge -> origin/pr/950/merge (forced update)
• ec24e31...305f0e5 refs/pull/999/merge -> origin/pr/999/merge (forced update)

and:

$ git log --graph --topo-order --oneline --decorate origin/master

• a97e9e7 (HEAD, origin/master, origin/HEAD, master) Merge pull request #1087 from ipfs/bitswap-perf
  |\
  | * 30ee01f (origin/pr/1087/head) increase window size for yamux streams
  |/
• 2406334 Merge pull request #1079 from ipfs/chriscool-readme-fix-todo
  |\
  | * 89961d9 (origin/pr/1079/head) README: alpha version has been released
  |/
• e225e9f Merge pull request #1074 from ipfs/fix/warning-logs
  …

As far as I can tell, that gives you all the benefits of your current
workflow with none of the trust assumptions ;).

[bmcorser]

We run our CI suite against the pr namespace. It's very useful.

Do you guys want me to make PRs now or are you happy for me to work on master until more contributors turn up?

[jbenet]

@wking ah interesting. would love to make it work with 0 trust. But the most important problem we've run into is collaborating, sometimes i can go in and fix some of the pedantic CR requests (or rebase on new master) to make it easier on others, and we cannot do that when it's a fork + PR. Also, so far we've had trouble with some CI systems running for PRs (jenkins!!! but we dont use jenkins anymore cause it died).

@bmcorser we always do PRs even when working on the same repo.

[whyrusleeping]

@jbenet @bmcorser I think for now, until there are other contributors youre good to work on master. Although, jbenet might have other opinions

[jbenet]

@bmcorser i'll try to read the code + comment if you do PRs. i think it's safest so can help answer questions / notice bugs / etc

[NoraCodes]

I'd love to work on this project - my time is currently heavily constrained by school work but after the winter holidays I will be able to devote a large amount of time to the project.

[BrendanBenshoof]

I think I am already added, but I am starting to have to bandwidth that would let me push on this project.
I might start working out modules + involved processes (a bit more involved in python than go) over the next few days.

[bmcorser]

Great! I too am rather constrained by work and study, but would be interested in doing code review and managing CI things. How do you guys want to communicate? I quite like the look of https://gitter.im (it's like a hipster IRC room attached to a GitHub repo) ...

[sirMackk]

I'd really like to join this effort. I think I'd be able to contribute 1-3 hours a week to throw Python at infrastructure/development problems :).

[JulienPalard]

I'm in, but I don't know how many hours I'll be able to throw in ^-^ Just liked the idea (IPFS) since the beginning :-) Glad today they have some specs -> give me motivation ! ("Walking on water and developing software from a specification are easy if both are frozen." - Edward V. Berard -) (Already on the team, just saying hello !)

[mvanveen]

Not currently sure what my commitment can be, but I'm interested in joining the initiative.

@bmcorser gitter works for me.
@jbenet do you have a recommendation for the best way to coordinate/get in touch with people pursuing the test suite? I feel like this is probably a strong apriori requirement for a burgeoning ipfs polyglot community and I'd personally love a testable spec to build against.

[blueskymonster]

Jumped in on gitter. Posting here as well to indicate interest. I'm going to read the white paper and really dive into the details in the upcoming week.