IRC log of git on 2015-12-17

Timestamps are in UTC.

13:52:01 [RRSAgent]
RRSAgent has joined #git
13:52:01 [RRSAgent]
logging to
13:52:26 [Ralph]
meeting: Using GitHub for W3C Specifications
13:52:33 [Ralph]
rrsagent, please make record public
13:56:12 [koalie]
koalie has joined #git
13:56:44 [tzviya]
tzviya has joined #git
13:57:58 [shawn]
shawn has joined #git
13:58:28 [Mike5]
we doing the project review here, right?
13:58:57 [koalie]
present+ Coralie
13:59:04 [Ralph]
present+ PLH, Wendy, Ralph
13:59:06 [plh]
Slides are:
13:59:13 [ivan]
Present+ ivan
13:59:14 [plh]
(same ones as Google slides)
13:59:24 [tzviya]
present+ TzviyaSiegman
13:59:24 [Ralph]
agenda+ scribe
13:59:25 [ChrisL]
ChrisL has joined #git
13:59:29 [Ralph]
agenda+ permission to record
13:59:34 [wseltzer]
wseltzer has joined #git
14:00:05 [jeff]
jeff has joined #git
14:00:07 [Ralph]
agenda 2 = notification of recording
14:00:50 [xiaoqian]
xiaoqian has joined #git
14:00:53 [Bert]
Bert has joined #git
14:01:11 [plh]
present+ plh
14:01:28 [shawn]
14:02:06 [dom]
Present+ Dom
14:02:11 [phila]
phila has joined #git
14:02:32 [Bert]
14:02:39 [phila]
present+ phila
14:02:45 [ted]
ted has joined #git
14:02:46 [tidoust]
Present+ Francois
14:02:47 [jeanne]
present+ jeanne
14:02:48 [ted]
Present+ Ted
14:02:55 [xiaoqian]
present+ xiaoqian
14:03:00 [Ian]
Ian has joined #git
14:03:09 [xueyuan]
present+ xueyuan
14:03:12 [Ian]
present+ Ian
14:03:13 [r12a]
zakim, who's here?
14:03:13 [Zakim]
Present: Coralie, PLH, Wendy, Ralph, ivan, TzviyaSiegman, shawn, Dom, Bert, phila, Francois, jeanne, Ted, xiaoqian, xueyuan, Ian
14:03:16 [Zakim]
On IRC I see Ian, ted, phila, Bert, xiaoqian, jeff, wseltzer, ChrisL, shawn, tzviya, koalie, RRSAgent, Zakim, xueyuan, plh, jeanne, Ralph, laurent, ivan, yatil, tidoust, dom, r12a,
14:03:16 [Zakim]
... nikos, Mike5, renoirb|aw, Yves, denis
14:03:23 [r12a]
present+ r12a
14:03:52 [nigel]
nigel has joined #git
14:03:57 [nigel]
Present+ nigel
14:03:58 [MichaelC]
MichaelC has joined #git
14:04:06 [MichaelC]
present+ MichaelC
14:04:26 [jeff]
scribenick: jeff
14:04:34 [shepazu]
shepazu has joined #git
14:04:56 [kaz]
kaz has joined #git
14:05:09 [jeff]
[slide 1]
14:05:13 [yatil]
present+ EricE
14:05:18 [Ian]
scribenick: Ian
14:05:31 [Ian]
-> slides
14:05:41 [Ian]
Presented by Phililppe Le Hégaret
14:05:43 [kaz]
present+ Kaz
14:05:54 [koalie]
14:06:21 [Ian]
NOTE: This is being recorded by AUDIO
14:06:24 [Ian] will be made public
14:06:28 [jeff]
14:06:34 [Ian]
...if you do not wish to participate by audio.....don't speak up!
14:07:06 [shepazu]
present+ shepazu
14:07:30 [Ian]
14:07:34 [Ian]
....gaining experience
14:07:40 [Ian]
...want to show how relates to process and patent policy
14:07:57 [Ian]
...ultimately we'd like to harmonize how we use the tool across groups to make it easier to participate across groups
14:08:07 [Ian]
14:08:15 [Ian]
What is this NOT about?
14:08:30 [virginie]
virginie has joined #git
14:08:52 [wseltzer]
zakim, what color is the bikeshed?
14:08:52 [Zakim]
I think Magenta haze
14:09:12 [Judy]
Judy has joined #git
14:09:16 [denis]
present+ Denis
14:11:39 [Ian]
14:12:00 [Ian]
PLH: large developer community using it...makes sense for w3c community to use it as well
14:12:30 [Ian]
...lots of tools (like post processing) available
14:12:53 [Ian]
14:13:29 [tzviya]
tzviya has joined #git
14:14:32 [Ian]
14:17:18 [Ian]
Repository manager ->
14:17:28 [Ian]
help track contributions from non-participants
14:17:42 [Ian]
14:18:22 [koalie]
-> Tool to facilitate integration: github-notify-ml
14:18:49 [Ian]
14:20:10 [koalie]
14:20:55 [jeff]
present+ jeff
14:21:12 [Ian]
14:21:27 [Ian]
PLH: Reduce notifications to not be overwhelmed
14:21:28 [vivien]
vivien has joined #git
14:22:03 [Ian]
14:23:31 [Ian]
14:23:50 [dom]
(re import: that applies if you have separate github repo, but also if you have a spec coming from any other separate version control system)
14:24:15 [Ian]
14:24:24 [Ian]
Note about force
14:24:34 [Ian]
Protect your main branch.
14:25:11 [vivien]
present+ vivien
14:25:12 [Ralph]
Don't Use the Force
14:25:16 [Ian]
14:25:25 [Ian]
PLH: One or several repositories?
14:25:33 [Ian]
(pros and cons)
14:25:54 [Ian] target 1 spec per rep
14:25:57 [Ian]
14:26:41 [Ian]
1 repo advantage is easier migration of material and 1 issues list
14:26:52 [dom]
[it's not so much that it is harder to move stuff from one spec to another; it's just that it's less revision history friendly]
14:27:16 [Ian]
14:27:24 [Ian]
PLH: when migrating to github DO NOT LOSE HISTORY of your spec
14:27:49 [dom]
[FWIW, I've done that transfer a number of times, and can help with imports if people need guidance]
14:27:58 [dom]
[and big +1 on not losing history]
14:28:07 [Ian]
14:28:28 [Ian]
PLH: invite comments!
14:28:31 [Ian]
...easy to track
14:28:37 [Ian]
...advertize your repo
14:29:17 [Ian] may accept them or not
14:29:18 [Ian]
14:29:29 [nigel]
zakim, q+ to ask if anyone creates a group repo to replace wiki with gh-pages and to manage actions as issues?
14:29:29 [Zakim]
I see nigel on the speaker queue
14:29:41 [Ian]
Use labels to differentiate between bugs, enhancement, etc.
14:30:11 [ivan]
Note that issues can be searched alongside labels, and those generate stable URI for a combination
14:30:35 [AdrianHB]
AdrianHB has joined #git
14:30:50 [Ian]
PLH: It's easy to give a +1 on github to support a proposal (there's even an emoji)
14:31:54 [Ian]
PLH: See dashboard for example of cross repo (?) issue tracking
14:32:11 [dom]
-> Web Perf Issue dashboard
14:32:25 [Zakim]
nigel, you wanted to ask if anyone creates a group repo to replace wiki with gh-pages and to manage actions as issues?
14:32:26 [dom]
[I've started looking at making it more generic and customizable]
14:33:02 [Ian]
nigel: One of problems that's arisen is what to do with actions
14:33:09 [Ian]'ve not mentioned so far interaction with tracker
14:33:20 [Ian]
...tracker had been used previously for issues and actions
14:33:32 [dom]
[in the WebRTC WG, we mostly assign people to issues instead of creating actions]
14:33:42 [Ian]
...has someone created a repo for the group that is separate from the group's documentation deliverables for actions?
14:33:43 [dom]
[but yes, some groups have done what Nigel describes]
14:33:44 [ivan]
14:33:53 [Ian]
...have they used ghpages functionality to replace the wiki?
14:33:58 [plh]
ack ivan
14:34:15 [Ian]
ivan: In our group we made use of explicit assignment of issues to people
14:34:31 [Ian]
also, the search facilities through the issues is very rich...see separate search page on github
14:34:39 [dom]
[the tag manages e.g. its agendas via github ]
14:34:46 [Ian]
...that let's you find easily which issues have been assigned to a particular person
14:35:13 [Ian]
nigel: I hear other technique is to use labels on issues
14:35:21 [Ian] question: should you just replace the w3c wiki in github?
14:35:30 [Ian]
[Web Payments WG uses the github wiki]
14:35:32 [dom]
[the TAG wiki ]
14:35:42 [Ian]
plh: home page of web platform WG is actually on github
14:35:49 [Ian]
...there's a separate repo for the home page of the WG
14:36:15 [vivien]
-> Web Platform Working Group homepage
14:36:15 [Ian]
...that repo does not contain any deliverables
14:36:17 [AdrianHB]
[Web Payments WG has a repo for the group and will create new ones for specs when we start writing them]
14:36:19 [Ian]
Nigel: that answers my question!
14:36:54 [koalie]
14:36:55 [AdrianHB]
[... and therefor, as Ian says, uses the wiki on that repo]
14:36:55 [vivien]
-> Web Platform Working Group homepage repository on GitHub
14:36:57 [Ian]
closing issues
14:38:24 [Ian]
PLH: Some groups allow closing by individuals, some close after X days,
14:38:32 [Ian]
...pick a closure policy that matches your group size, culture
14:38:39 [Ian]
..make it as easy as possible to close typo fixes
14:38:49 [Ian]
14:38:54 [Ralph]
[I interpret "allowed to close" and "can close" on the slide as "entitled to ..."
14:38:58 [Ian]
wide review / horizontal issues
14:39:15 [vivien]
14:39:36 [Ian]
PLH: My thinking is that we can use labels (e.g., "Security")
14:39:48 [Ian]
(Slide 19 includes a list of proposed labels)
14:40:18 [Ian]
Web platform forms team for reviewers
14:40:32 [Ian]
if you are interested in issue of type X you should join a team
14:40:40 [r12a]
14:41:01 [plh]
ack r12a
14:41:09 [virginie]
+1 on the team and label idea \o/
14:41:15 [Ralph]
+1 to support these labels
14:41:25 [Ian]
r12a: We'd like to be able to label issues
14:41:35 [Ian]
...e.g., browsing repo we'd like to be able to track some issues
14:41:38 [Ian]
...can we add labels?
14:42:04 [Ian]
PLH: I think Teams may work since you can comment if you are part of a team
14:42:06 [Ralph]
+1 to entitle other groups to label issues that impact them
14:42:13 [dom]
[anyone with admin rights can add labels; but that means only staff by default]
14:42:23 [Ian]
...since non participants may not have authorization to label arbitrary repos
14:42:37 [Ian]
..thanks for the feedback
14:42:42 [dom]
[we could also build a tool specifically to label issues on w3c repos]
14:43:07 [Ian]
14:43:32 [Ian]
PLH: I prefer pull requests to fix typos to email telling me to do so...just push a button
14:43:36 [Ian] encourage pull requests
14:44:16 [Ian]
...encourage all WG to edit a spec!
14:45:29 [Ian]
...enable peer reviews early
14:45:32 [dom]
q+ to mention continuous integration as a tool to facilitate pull request management
14:45:39 [Ian]
...Don’t require a pull request to fix a simple typo and allow direct commits for those
14:46:04 [dom]
-> Continous integration scripts for WebRTC specs
14:46:06 [plh]
ack dom
14:46:06 [Zakim]
dom, you wanted to mention continuous integration as a tool to facilitate pull request management
14:46:27 [Ian]
dom: With regard to pull requests, we've found useful in WebRTC to use continuous integration systems
14:46:32 [Ian]
...Travis is such a system
14:46:41 [Ian]
...we have a number of scripts that run when someone submits a pull request
14:46:59 [Ian]
...easy for submitters to know that there's a missing reference, or WebIDL not well-formed, or markup bug
14:47:06 [Ian]
....helps include quality of pull request
14:47:17 [dom]
14:47:21 [r12a]
i think a validator check tool would be useful !
14:47:51 [Ian]
14:48:09 [dom]
r12a, my script does html validation among other checks (both pre- and post-respec processing)
14:48:17 [Ian]
...Merge pull requests as you close issues
14:48:21 [r12a]
14:48:28 [Ian]
PLH: Tip - Consider squashing the commits to maintain a clean commit history for your specification
14:48:44 [Ian]
...this useful when there are multiple commits before getting to consensus
14:49:13 [dom]
[also, regarding commit messages, I've found to be a useful approach to writing commit messages]
14:49:16 [tidoust]
q+ to wonder whether there is an easy online way to "view" the result of a PR on an an HTML spec
14:49:32 [Ian]
14:49:34 [Ian]
Github Team
14:49:50 [Ian]
PLH: Empower as many as possible in your group to be part of github team
14:49:59 [Ian]
...avoid silos
14:50:05 [dom]
tidoust, should work for that too, shouldn't it?
14:50:50 [Ian]
14:50:54 [Ian]
PLH: Ideally you use ghpages
14:51:03 [Ian]
...github service makes easy to see on the web
14:51:08 [Ian]
...if you do that you don't need master branch
14:51:09 [tidoust]
It should, dom, but is directly available from a PR page?
14:51:16 [Ian] can make ghpages the default
14:51:31 [tidoust]
14:52:24 [Ian]
PLH: we are thinking about git branches as a way to manage flow on rec track
14:52:33 [dom]
[some groups (for better or for worse) use master as a staging branch, and gh-pages when they want to release an editors draft]
14:53:17 [Ian]
PLH: If your edits reflect group consensus, your editors drafts are really like WDs
14:53:29 [Ian] through Travis you can publish your docs (using Travis) to /TR automatically
14:53:40 [Ian]
...we know how to do this with respec
14:54:02 [nigel]
[you can view the outcome of a pull request by viewing the changed file, choosing 'View', then 'Raw', then adjusting the URL from to]
14:54:05 [r12a]
14:54:32 [plh]
ack r12a
14:54:34 [Ian]
...note that pubrules checker tool we know will be retired by Aug checker in development
14:55:10 [Ian]
r12a: Re ED-as-WG...we've encountered the issue of making comments and then patching them back to original document
14:55:22 [Ian]
...we've had a lot of trouble having to look at Editor's drafts that have constantly changed
14:55:30 [Ian]
...hard to track back to the draft that was the source of problems
14:55:38 [Ian] I18N WG publishes to TR often
14:55:49 [Ian]
...we ask people to review the TR version since the editor's draft evolves
14:55:54 [Ian]
PLH: Coming up in the talk!
14:56:01 [Ian]
14:56:06 [Ian]
[Automatic publication workflow]
14:56:47 [Ian]
PLH: Question of "when to publish"
14:57:11 [Ralph]
"significant change"
14:57:11 [Ian]
..pros and cons to "on every commit" or "every significant commit" or "on demand"
14:57:22 [Ian]
...but in any case, please don't let more than 6 months elapse between TR publications
14:57:51 [ivan]
14:57:57 [Ian]
..that's a process requirement
14:58:21 [Ian]
14:58:27 [Ian]
[W3C process]
14:58:40 [Ian]
PLH: regarding the wide review requirement...suggest using labels
14:58:53 [Ian]
...also note use of
14:59:07 [Ian]
...if you mention "wide review" in your doc, we have a tool that will automatically advertise your spec to that list
14:59:15 [Ian]
[What tool is that?]
14:59:49 [Ralph]
-> Process-2015 Wide Review
15:00:10 [Ian]
PLH: You don't have to do anything to use the tool
15:00:16 [Ian]
...just put "wide review" in the spec
15:00:36 [Ian]
PLH: Note that if you say "we don't want wide review' in your spec you'll hear from me...that's not coded!
15:00:54 [Ralph]
-> Process-2015 6.3.2 Revising Public Working Drafts
15:01:14 [Ian]
PLH: We don't yet have a recommended way to generate a disposition of comments from an issues list
15:01:28 [Ian]
..there is a group that used labels to do that
15:01:42 [Ralph]
[the CSV on the Web Working Group]
15:01:47 [Ian]
...but we don't yet have a "recommended" way yet (their way worth consideration)
15:01:52 [Ian]
15:01:57 [Ian]
Additional tooling
15:02:05 [Ian]
PLH: Web Payments WG is using Wiki to build agenda
15:02:36 [Ian]
...we do not have a tool yet to generate activity summaries
15:02:42 [Ian]
...see modern tooling doc for more info:
15:02:49 [Ian]
...and ideas
15:02:49 [Ian]
15:03:10 [Ian]
...for example, there's a tool called "gitter" that is a chat tool integrated with github
15:03:23 [Ian]
..I don't know yet of any group that's decided to switch from IRC to bitter
15:03:37 [Ian]
15:03:41 [Ian]
[Going forward]
15:04:33 [Ian]
PLH: If you are interested in a project review on details of using github (e.g., pull requests) or auto publishing, let me know.
15:04:42 [Ian]
[Slides done]
15:04:43 [Ian]
15:04:44 [plh]
15:04:51 [Ian]
zakim, close item 1
15:04:51 [Zakim]
agendum 1, scribe, closed
15:04:52 [Ian]
zakim, close item 2
15:04:52 [Zakim]
I see 1 item remaining on the agenda:
15:04:52 [Zakim]
2. notification of recording
15:04:53 [Zakim]
agendum 2, notification of recording, closed
15:04:54 [Zakim]
I see nothing remaining on the agenda
15:05:01 [Ian]
Reminder: Audio recording
15:05:15 [Ian]
ack ivan
15:05:16 [plh]
ack ivan
15:05:26 [Ian]
ivan: Back on slide 26
15:05:26 [jeanne]
I am very grateful for this session. This is the kind of information that is very useful and not documented in other places.
15:05:41 [Ian]
ivan: We've not used Echidna in our IG (not yet for IGs)
15:05:58 [Ian]
...can I select a specific github branch to be used for automatic publishing
15:05:59 [Ian]
PLH: Yes.
15:06:37 [Ian]
..the only problem with that (mentioned by tidoust) is that if you use something other than ghpages branch, you need to use the rawgit view...and if you have materials other than that page, there can be problems.
15:06:42 [dom]
[I think echidna should be improved to retrieve content via git rather than just http]
15:07:18 [Ian]
ivan: if we use automatic publishing tool, sounds like it mostly only works with alternative 1 on (publish on every commit)
15:07:22 [Ian]
PLH: Yes, that's correct.
15:07:39 [Ian]
...we don't yet have a way to do that with 2 or 3 (on slide 26) ... but mostly people haven't looked yet into doing it
15:07:42 [Ian]
ivan: Are there plans yet?
15:07:50 [Ralph]
[the alternatives described are on ]
15:08:08 [Ian]
plh: I'm not looking into it
15:08:10 [Ralph]
15:08:33 [plh]
15:08:35 [Ralph]
15:08:42 [shepazu]
15:08:53 [xiaoqian]
15:08:54 [koalie]
thanks Ian for scribing and plh for the review! bye all!
15:09:07 [Ian]
rrsagent, make minutes
15:09:07 [RRSAgent]
I have made the request to generate Ian
15:09:10 [Ian]
rrsagent, set logs public
15:09:27 [koalie]
koalie has left #git
15:10:49 [Ian]
rrsagent, bye
15:10:49 [RRSAgent]
I see no action items