- From: Robin Berjon <robin@w3.org>
- Date: Thu, 13 Dec 2012 12:05:50 +0100
- To: Dominique Hazael-Massieux <dom@w3.org>
- CC: Arthur Barstow <art.barstow@nokia.com>, public-script-coord@w3.org, public-webapps-testsuite@w3.org
On 12/12/2012 18:07 , Dominique Hazael-Massieux wrote: > Le mercredi 12 décembre 2012 à 18:35 +0200, Arthur Barstow a écrit : >> Yes, please do. I would say create a new directory in the following but >> would defer to Travis' recommendation: >> >> <http://dvcs.w3.org/hg/webapps/WebIDL/tests/submissions/> > > OK, I've put them in a "W3C" subdirectory, with a succinct readme to > explain how they're expected to be used. > http://dvcs.w3.org/hg/webapps/file/7810408d4c8e/WebIDL/tests/submissions/W3C Cool, thanks a lot for those tests, they're very useful. > I'd like to get my tools to rely on the official test suite over time > (so that I don't have to maintain them in several settings), but the > current "submissions" structure, and the lack of wider agreement on the > format for these tests make that difficult. I hope this can be solved > sooner rather than later :) . Ditto. What I do currently with WebIDL2.js is that I use widlproc as a git submodule in order to import your tests and keep them in sync. It works like a charm so I plan to keep doing so in the hope that you'll keep both sides in sync ;-) In terms of resolving this whole issue of clunky directory structures and obsolete version control systems, my understanding from what Art has said before has been that WebApps was waiting on the HTML TS to reach agreement on how to handle this, at which point it would follow. It so happens that we've been actively working on this and it would be *very* useful to benefit from additional input. Here are some checkpoints on how the proposal is shaping up (not all of it is applicable outside HTML, but the overall organisation is): http://www.w3.org/mid/50C7211D.9080909@w3.org http://www.w3.org/mid/50C75406.4070705@w3.org http://www.w3.org/mid/50C85D09.6050800@w3.org The topics that are still open and for which external input would be welcome: • What branches should be called when there are multiple versions of a specification being developed in parallel? (Yeah, this is die hard bikeshedding, but when you're stuck on that you're stuck on that.) • Whether we should move to GitHub. To be honest, I have yet to hear an argument against doing so, but there is some modicum of continued opposition nevertheless. I encourage those who are interested in this topic to weigh in so that we can solve these boring logistical issues. -- Robin Berjon - http://berjon.com/ - @robinberjon
Received on Thursday, 13 December 2012 11:06:06 UTC