20:33:27 RRSAgent has joined #auto 20:33:27 logging to http://www.w3.org/2015/04/29-auto-irc 20:34:37 Meeting: Automotive WG - Repository Layout 20:35:09 paul: would have a separate repository for the spec and test resources 20:35:37 ted: there is a repo for the Auto BG 20:36:11 ... would make sense to have a repo "Automotive" 20:36:30 ... can migrate the resources within the BG's one to the new repo 20:36:48 Present: Paul, Dave, Ryan, Kaz, Ted 20:37:16 dave: moving resources from automotive-bg to Automotive? 20:37:29 s/Automotive/automotive/ 20:38:15 ... clone the automotive-bg and create automotive 20:38:44 ... so that all the history can be tracked 20:39:19 paul: this is the repo Ted has created 20:39:29 (goto meeting display) 20:39:48 (go through w3c/automotive-bg) 20:40:20 paul: let's keep all this stuff 20:40:36 ryan: what about media tuner stuff? 20:40:53 ... will we create a new directory? 20:41:10 ted: just created "automotive" 20:41:18 (w3c/automotive) 20:41:39 ted: will create vehicle api, data, etc. 20:41:46 ... and put other stuff as well 20:42:20 paul: this guy (w3c/automotive-bg) is for the BG 20:42:50 [paul starts to draw an image of the directory structure] 20:42:59 [on the goto meeting display] 20:43:11 [[ 20:43:13 automotive 20:43:22 --media_tuner 20:43:34 --api 20:43:40 --use_cases 20:43:43 --tests 20:43:47 --vehicle_data 20:43:53 --api 20:43:54 ]] 20:44:21 paul: there are two specs, Vehicle API and Data 20:44:31 ... and tests may need multiple files 20:45:13 ... any opinions on the structure and name? 20:45:23 ... we're using wiki for use cases 20:45:36 ... seems wiki is a good way to gather use cases 20:46:15 ... we need two implementations for each spec 20:46:28 ... all the implementations have to be public? 20:46:35 ted: commercial ones are ok 20:46:55 paul: does this basic structure make sense? 20:47:22 dave: depends on each group 20:48:39 [paul shows the repo of the Web Application Security Working Group] 20:48:47 [webappsec] 20:56:50 (some discussion on possible concern from the viewpoint of IPR) 21:08:00 -> https://github.com/w3c/presentation-api presentation api 21:09:29 -> https://github.com/w3c/csswg-drafts CSS 21:11:21 -> https://github.com/w3c/tvapi tv control cg 21:11:39 -> https://github.com/w3c/tv-ggie tv ig 21:15:44 kaz: @@@ 21:15:56 paul: don't think we should have multiple repositories 21:16:07 dave: we could use tags 21:16:44 paul: can identify distribution version as a snapshot 21:17:36 s/@@@/when I suggest we think about repositories for BG and WG, I was thinking about having automotive-bg for a developer version and automotive or automotive-wg for a distribution version/ 21:17:43 https://github.com/w3c/echidna/wiki/How-to-use-Echidna 21:18:42 ted: we can use the automatic publication tool above 21:19:17 dave: feature branch? 21:21:26 [dave put https://github.com/w3c/svgwg on the goto meeting] 21:21:44 [also put "v.2.8.4" is a tag: https://raw.githubusercontent.com/mcavage/node-restify/v2.8.4/README.md] 21:22:15 pb has joined #auto 21:22:20 djensen47 has joined #auto 21:22:30 Hello 21:22:33 hello 21:23:15 ted: three issues on the spec 21:24:05 ... accessibility consideration, etc. 21:24:11 s/accessibility/extensibility/ 21:24:54 paul: you don't have to implement all the data properties 21:25:22 +1 to FAQ 21:25:52 ted: best practices 21:28:45 ... will move the resources from the automotive-bg repo to the new automotive repo 21:29:16 ... Ted and Kaz will check the best practices of GitHub usage 21:30:17 [adjourned] 21:30:24 rrsagent, make log public 21:30:30 rrsagent, draft minutes 21:30:30 I have made the request to generate http://www.w3.org/2015/04/29-auto-minutes.html kaz 21:32:31 ted: moving with history might be tricky 21:33:13 i'm looking around for techniques, this one is incomplete afaict http://gbayer.com/development/moving-files-from-one-git-repository-to-another-preserving-history/ 21:33:42 Yeah, I was looking at that one too. 21:34:58 rrsagent, draft minutes 21:34:58 I have made the request to generate http://www.w3.org/2015/04/29-auto-minutes.html kaz 21:35:31 s/new automotive repo/new automotive repo with all the history (if possible)/ 21:35:38 rrsagent, draft minutes 21:35:38 I have made the request to generate http://www.w3.org/2015/04/29-auto-minutes.html kaz 21:39:43 okay, you can do this without losing history but it will look like history is lost on github 21:40:39 well, after a `git mv` it will appear that way 21:42:57 subtree merge should worl 21:43:03 s/worl/work/ 21:50:22 maybe not :-/ 21:57:26 https://github.com/djensen47/automotive 21:57:48 preserves history, leaves only the two files we want 21:58:59 $ git remote add bg git@github.com:w3c/automotive-bg.git 21:59:10 $ git pull bg master 21:59:30 This will result in a conflict in README.md 21:59:54 resolve conflict and $ git add README.md 22:00:06 $ git commit -m "Merged automotive-bg" 22:00:39 $ git rm -r 22:01:20 $ git commit -m "Removed files." 22:01:50 $ mkdir vehicle_data 22:02:04 $ git mv *.html vehicle_data 22:02:20 $ git commit -m "Moved spec files" 22:03:07 It turns out that after you `git mv` a file history is preserved but won't show up unless you use `git log --follow ` 22:03:25 s/a file history/a file, history/ 22:11:46 http://scottwb.com/blog/2012/07/14/merge-git-repositories-and-preseve-commit-history/ 22:11:57 i think the problem resulted from wanting the new dir 22:13:02 Yup, I think that's what I described above. 22:13:17 ah maybe i'm not seeing history because of the git mv 22:13:22 yes 22:13:43 if you clone the repo and `git log --follow ` then you see the "file" history 22:14:17 it would be nice if github supported --follow but they don't seem to 22:14:32 ok so then what i have in automotive directory is right 22:14:39 confirmed with git log 22:15:19 yup 22:15:39 the 164 commits is a good sign :-) 22:19:47 https://github.com/w3c/automotive-bg has breadcrumbs to new repo 22:20:14 and the readme for vehicle_data tip for getting full history https://github.com/w3c/automotive/tree/master/vehicle_data 22:21:44 thanks djensen47 22:22:48 no problem; I've done this before — was just looking for a "better way" but I guess there's no getting around what "git mv" does