#selenium

lukeis: 1) disable ci job 2) delete github selenium repo 3) move gc repo to github?

seems straight forward (isH)

i don't want to delete the repo, would kill current pull requests

i'm sending an email to support to remove the mirror

since we manually mirror

nirvdrum: You should raise your concerns on the mailing list if you think they are important.

so, seems like we can just 1) update ci to point to github instead of GC :)

ato: I'm kinda just venting a bit. We went through this last summer with lukeis's Google Doc voting thing.

don't remember that?

Well I'm sad that it has to feel like a battle.

lukeis: It was tied to the decision to switch to git, IIRC.

ato: Not a battle. I'm just concerned about issue maintenance on GitHub. If others are able to manage it, great.

nirvdrum in fairness, have you seen our current issue tracker?

jleyba, lukeis: https://code.google.com/p/selenium/wiki/Project...

jleyba: Yeah. It has a completely separate set of issues.

The full discussion: https://groups.google.com/forum/#!searchin/sele...

ato: That's the one.

nirvdrum the biggest being independent of the platform: most issue reports are terrible

Please note that I was in favour of staying at GC.

Reality has changed since then.

heh, resurrect the vote?!

GC is no longer an option!

jleyba: Fair enough. Selenium is just sorta carved out into sections with people responsible for each. It has extremely poor support for that sort of ownership. It's also terrible to try to plan anything in or really categorize issues.

But if smart people here don't think that's likely to be a real problem, then I'm cool rolling with it.

my hunch is the tagging system will work well enough

maybe.. i don't know

:)

slightly off-topic… i'm in favor of splitting the ruby and python bindings into their own repos on github

lukeis: don't think i would object to that (and pushing c# to its own as well)

what is the advantage of doing that?

would there still need to be submodules in the main repo? Or are we talking forks?

i don't know what c#'s dependencies are… python and ruby only depend on the firefoxdriver xpi

Titus: essentially forks

history spliced of course to just limit the files each care about :)

I like that we can build everything that is officially supported in one go.

simonstewart: ohai

yo

welcome to project hosting debate day!

lukeis: actually, i take it back. splitting would end up being more hassle in the long run

simonstewart: read the log? https://botbot.me/freenode/selenium/  (the new one that is)

simonstewart, oh hai! random buck question for you: is there a way to build a src jar?

simonstewart, took them long enough to post that

Yeah. //some:target#src

awesome, thanks

Any care to provide a synopsis of the discussion from today?

GC -> github; get 'er done

:)

Not bitbucket?

I'm starting to enjoy using mercurial

someone said you'd start suggesting that...

you also may enjoy walks on hot coal

never used it

Sorry, I have bigger fish to fry today.

Like end-of-quarter deliverables.

ato: ?

What's the ownership thing that nirvdrum was talking about?

Ah. Bugs in GC's issue tracker

Right then

I'll post this to the group later, but questions we need to answer:

1) Do we port our existing issues to GC

2) What do we do with the existing wiki?

oh and Wiki on GC is permissioned on edits… Github is open to the world

3) Do we want to use GC's issue tracker? I know Atlassian foks, so we could try and get an OSS JIRA hosted instance

For question 1 --- s/GC/GH/

1) declare issue bankruptcy and close it all

2) port it

I think it's a no-brainer that we'll end up on GitHub

3) suck it up and use GH to keep everything in one place

1) What happens if we close all Core and RC issues in our existing tracker?

they die forever (i hope)

simonstewart: I think you figured it out, but I meant IE bugs go to jimevans, Firefox bugs go to AutomatedTester, etc.

I mean, how many open issues would be left?

lots

tons even

only ~20% are RC / core

We declared issue bankruptcy when we moved from JIRA to GC

I'm okay doing it again

If we do it, I'd like us to establish an actual issue SLA documented on the wiki

Which would be?

your issue may be bit-rotted ;)

No more "NeedsClarification" bs. If you're issue report is vague, it gets closed

+1

I started one for the JS bindings: https://www.npmjs.com/package/selenium-webdrive...

Guess that's not really a SLA... covers my main grievances with the current process though

i'm going to create a bookmarklet that automatically populates a message "Issues are for actual Selenium Project issues. Please use the mailing list for support" & close.

need to make one for selendroid… i've written that too many times

Github "mirroring" removed https://github.com/SeleniumHQ/selenium

lukeis: as in "it's already done"?

correct

Or "it needs doing"?

Who did that?

i did

nothing's changed

we still "mirror" via :ci

Jenkins: http://ci.seleniumhq.org:8080/

Did you see the email I just sent to the selenium-developers list?

:)

the Github "mirroring" didn't work to our satisfaction

:)

simonstewart: you didn't see in IRC log that i made the request, did you ;)

Nope

ok, then check! done.

I modified the strap line

nice

some more implications… do we "merge" Pull Requests through the Github UI? (allowing the 'merge' commit) or do we still employ our current process?

NO

We keep our current process, since the merges violate the single logical change principle

sgtm

What if everything is squashed down first?

you still get a merge commit

ok

github needs a "Squash / rebase / sign-off (don't 'merge')" button for us :)

I've asked before

Bug migration vote open on the se-dev mailing list

I vote for option 5

5! vote 5!

I like that one.

is there a way to change "origin" on my local repo to point to GH?