W3C home > Mailing lists > Public > public-webrtc-editors@w3.org > November 2015

New CI system for WebRTC spec

From: Dominique Hazael-Massieux <dom@w3.org>
Date: Wed, 18 Nov 2015 11:24:04 +0100
To: "public-webrtc-editors@w3.org" <public-webrtc-editors@w3.org>
Message-ID: <564C51C4.7000802@w3.org>
Hi,

As discussed some weeks ago, I've been developing with Vivien a new set 
of scripts for our continuous integration with Travis, with the 
following goals:
* use Travis new and faster infrastructure
* simplify use of the validity checks in one's local setup
* incorporate optional line-wrapping check (per Martin's suggestion)
* make Travis failures due to external factors less likely and easier to 
ignore; in particular:
  - we've re-enabled link checking, but only looking for internal 
anchors (avoiding the errors due to intermittent network issues we've 
had in the past)
  - by default, the script uses the default ReSpec version, but if 
ReSpec introduces a bug, we can make the script uses the last known good 
version of ReSpec

The CI script is now essentially a makefile that can be called either 
from Travis or from your local setup; the various commands usable with 
the Makefile are documented in the README on 
https://github.com/w3c/webrtc-respec-ci

The main changes this implies for editors are:
* when adding a new subresource to the spec (typically an image), that 
resource needs to be listed in W3CTRMANIFEST; this file will eventually 
also be used for automatic publication on /TR/ (a topic for another day)

* with this setup, both the validity of the source (pre-respec 
rendering) and the result (post-respec rendering) will be checked; so 
far, only the latter was being checked

The proposal for incorporating this at this stage is to merge PR #315 if 
the editors confirm they are satisfied with this new approach. The pull 
request includes fixes to webrtc.html to make sure it doesn't fail the 
CI check.

At the moment, I'm not proposing to incorporate line-length enforcement 
in our usage of the CI script, as it was suggested earlier we should 
wait for the spec to stabilize before enabling this.

I'll also prepare a pull request for getUserMedia.

Dom
Received on Wednesday, 18 November 2015 10:24:08 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 19:19:03 UTC