12:57:03 <RRSAgent> RRSAgent has joined #pub
12:57:03 <RRSAgent> logging to https://www.w3.org/2021/03/18-pub-irc
12:57:05 <Zakim> Zakim has joined #pub
12:58:18 <sidvishnoi> sidvishnoi has joined #pub
12:59:13 <dom> dom has joined #pub
13:00:03 <plh> present+ plh
13:00:08 <dom> Present+
13:00:11 <wseltzer> wseltzer has joined #pub
13:00:13 <xfq> present+
13:00:14 <denis> present+
13:00:19 <cwilso> cwilso has joined #pub
13:00:29 <cwilso> present+
13:00:30 <weiler> weiler has joined #pub
13:00:53 <florian> florian has joined #pub
13:01:10 <florian> present+
13:01:18 <xueyuan> xueyuan has joined #pub
13:01:18 <Ralph> Ralph has joined #pub
13:01:22 <atsushi> atsushi has joined #pub
13:01:25 <atsushi> present+
13:01:25 <Ralph> present+
13:01:42 <weiler> present+
13:02:08 <xiaoqian> xiaoqian has joined #pub
13:02:25 <wseltzer> present+
13:02:53 <sidvishnoi> present+
13:03:06 <xiaoqian> present+
13:03:08 <Bert> Bert has joined #pub
13:04:08 <jeff> jeff has joined #pub
13:04:16 <jeff> present+
13:04:29 <tidoust> present+
13:05:06 <plh> NOTE: the presentation is recorded
13:05:10 <denis> https://www.w3.org/2021/03/18-echidna/?full#1
13:05:13 <plh> scribe+
13:05:32 <plh> Denis: main goal is help maintain /TR up-to-date
13:05:41 <plh> ... it should represent what the Web should be
13:05:51 <plh> [Slide 2]
13:06:03 <plh> Denis: in the past, the only way was to submit a request to the webmaster
13:06:12 <plh> ... outside of moratorium, etc.
13:06:17 <Yves> Yves has joined #pub
13:06:22 <plh> ... this created delays, back-n-forth, etc.
13:06:43 <caribou> caribou has joined #pub
13:06:44 <plh> ... given how heavy the Process, the specs were not published often, resulting in an outdated /TR
13:07:04 <plh> ... so we came with an automatic system
13:07:14 <plh> ... it checks if the resources can be downloaded
13:07:23 <plh> plh has changed the topic to: Slides: https://www.w3.org/2021/03/18-echidna/ (Recording in progress)
13:07:32 <plh> [Slide 3]
13:07:47 <plh> Denis: now the system has limitations, such as can't be used for FPWD
13:08:20 <plh> ... for CRs, limitation like if the doc changed WG
13:08:31 <MikeSmith> MikeSmith has joined #pub
13:08:42 <plh> ... Echidna can't handle shortname changes, some backward moves
13:08:46 <plh> [Slide 4]
13:09:21 <plh> Denis: you need your spec on GH. A group decision to use echidna.
13:09:35 <plh> ... need a token for each spec, generated by team contact/chair
13:09:42 <plh> ... should be kept private
13:09:57 <plh> ... you'll need admin rights to your GH repo
13:10:06 <plh> [Slide 5]
13:10:21 <plh> Denis: for CR snapshots, you'll need Director and Comm approval
13:10:34 <plh> ... w3c/transitions is used to get Director's approval
13:10:55 <plh> ... the template to raise a transition can be found in the README
13:11:04 <plh> ... when echidna will receive a request to publish
13:11:17 <plh> ... it will check the requirements
13:11:25 <xfq> https://github.com/w3c/transitions/issues/199
13:11:26 <xfq> https://github.com/w3c/transitions/issues/261
13:11:33 <plh> ... see examples linked from the slides
13:11:36 <plh> [Slide 6]
13:11:59 <plh> Denis: introduced how to setup the repo
13:12:18 <plh> ... Sid Vishnoi set up a GH action spec-prod to help
13:12:39 <plh> ... spec-prod can generate snapshots from respec/bikesheds
13:12:39 <xfq> https://github.com/w3c/spec-prod/
13:12:51 <plh> ... run checks like borken links
13:13:15 <plh> ... if the document is a compound one (like images, css, scripts), spec-prod will take care of them
13:13:32 <plh> [Slide 7]
13:13:45 <plh> Denis: first is to add the generated token
13:13:54 <plh> ... added as a repo secret
13:14:05 <plh> ... can be done in the settings of your repo
13:14:17 <xfq> https://github.com/w3c/{REPO_NAME}/settings/secrets/actions
13:14:19 <plh> [Slide 8]
13:14:39 <plh> Denis: [shows screenshot of adding a token]
13:14:48 <plh> [Slide 9]
13:14:59 <plh> Denis: you're now ready to create the GH action workflow
13:15:06 <plh> ... it's a simple YAML file
13:15:22 <plh> ... [shows example]
13:16:04 <plh> ... here, the workflow runs on any pull requests and push on the main branch
13:17:13 <plh> ... with this workflow, spec-prod will generate snapshots of the source doc (index.html or index.bs by default)
13:17:31 <plh> ... if it's a push on the main, it will trigger echidna
13:17:37 <plh> [Slide 10]
13:17:49 <plh> Denis: parameters to customize the workflow
13:18:15 <plh> .... (going through the list of parameters)
13:18:42 <plh> ... if you want to receive an email after Echidna is done, use W3C_NOTIFICATIONS_CC
13:18:51 <xfq> https://github.com/w3c/spec-prod/blob/main/docs/options.md
13:18:51 <plh> [Slide 11]
13:19:06 <plh> Denis: using those variables in YAML example
13:19:29 <plh> ... should read W#C_NOTIFICATIONS_CC on the slide
13:19:40 <plh> [Slide 12]
13:19:59 <plh> Denis: mainly focusing on /TR but spec-prod can be used for GH pages
13:20:12 <plh> [Slide 13]
13:20:19 <plh> Denis: checking the results from Echidna
13:20:28 <plh> ... the action tab will show you those
13:20:48 <plh> [Slide 14]
13:21:06 <plh> Denis: the screenshot sends a request to Echidna via a curl command
13:21:21 <plh> ... and then query for results
13:21:38 <plh> ... until the process returns for a result
13:21:45 <xfq> https://lists.w3.org/Archives/Public/public-tr-notifications/
13:21:56 <marcosc> marcosc has joined #pub
13:22:01 <plh> [Slide 15]
13:22:13 <plh> Denis: gives an overview on what is done and different steps
13:22:35 <plh> ... for a CR Snapshot, you'll need a transition request
13:23:31 <plh> ... [going through the workflow]
13:23:44 <plh> [Slide 16]
13:24:03 <plh> Denis: we presented spec-prod, which was built a few months ago
13:24:13 <plh> ... it's the easiest to get started
13:24:21 <plh> ... for alternatives, check the wiki
13:24:26 <plh> [Slide 17]
13:24:33 <florian> q?
13:24:36 <florian> q+
13:24:37 <plh> Denis: if issues arise, get help
13:25:12 <plh> [Slide 18]
13:25:27 <plh> Denis: for the future, want to remove more limitations
13:25:41 <plh> ... some are related to the old process document
13:25:59 <plh> ... also look at facilitating revising W3C Recommendations
13:26:19 <plh> ... new things will get announced to spec-prod@w3.org
13:26:31 <MikeSmith> q+ to ask
13:26:48 <denis> ack florian
13:27:13 <plh> Florian: thank for Echidna. My workflow is different. we don't publish on every commit in CSS
13:27:20 <Caroline_> Caroline_ has joined #pub
13:27:23 <Caroline_> Present+
13:27:29 <dom> [you could have a different branch where you only push WG-agreed changes]
13:27:32 <plh> ... substantive changes need to be explicitely approved for each publication
13:27:39 <dom> [and use spec-prod for pushes on that branch]
13:27:43 <plh> ... so the editor's draft hasn't been approved yet
13:27:53 <plh> ... so we use Echidna in a manual way
13:28:09 <plh> ... is there some intermediate way we could use?
13:28:20 <dom> qq+
13:28:24 <plh> ... like flagging some drafts for publications but not others
13:28:49 <denis> ack dom
13:28:49 <Zakim> dom, you wanted to react to florian
13:29:15 <xfq_> xfq_ has joined #pub
13:29:23 <plh> Dom: simplest way would be a dedicated branch and you push on that branch. spec-prod would react only to that branch
13:29:45 <plh> Denis: an other event to listen is the one triggering the run manual
13:29:45 <dom> "workflow_dispatch" is another approach indeed
13:29:54 <denis> https://docs.github.com/en/actions/reference/events-that-trigger-workflows#workflow_dispatch
13:29:54 <florian> q?
13:30:17 <plh> Francois: workflow_dispatch works perfectly for that indeed
13:30:31 <plh> ack Mike
13:30:31 <Zakim> MikeSmith, you wanted to ask
13:30:43 <weiler> Denis, that was a great presentation.  Thank you very much!
13:31:12 <plh> Mike: sounds like an issue for spec-prod. other groups might have the same process than the CSS Process.
13:31:15 <dom> [documenting workflow_dispatch as an example in spec-prod sounds like a good addition indeed]
13:31:37 <sidvishnoi> It's on roadmap: https://github.com/w3c/spec-prod/issues/6
13:31:43 <plh> .... so would benefit from builtin support
13:32:04 <plh> Mike: what about the case when somebody pushes more than once a day
13:32:32 <plh> ... I found that Echidna cannot republish to the same URL in /TR in multiple times on the same day
13:32:46 <vivien> vivien has joined #pub
13:32:51 <vivien> present+
13:33:04 <plh> ... echidna checks for the same URL
13:33:09 <plh> ... and intentionally fails
13:33:14 <dom> [I thought echdina explicitly allowed several publications a day?]
13:33:37 <MikeSmith> https://github.com/w3c/spec-prod/pull/33#discussion_r595664294
13:33:55 <plh> Denis: Echidna should be able to publish multiple times on the same day. sounds like a bug
13:34:04 <plh> ... please report it
13:34:12 <dom> [it may be a bug in spec-prod too]
13:34:26 <dom> [I recall spec-prod has special handling for this]
13:34:51 <plh> Florian: the new one will replace the old one
13:35:10 <plh> Denis: indeed, the previous version should be the one published days before
13:35:34 <plh> Mike: that's the bug I'm running into
13:35:58 <dom> [there may be something about change of status shortnames (CR vs CRD vs WD)]
13:36:35 <plh> Denis: sounds like a spec-prod
13:36:55 <plh> Mike: but pubrules check those URLs
13:37:22 <plh> Florian: spec-prod fails to check if it's a same day pubrules
13:37:38 <plh> q?
13:38:07 <plh> Mike: but those links are generated from biblio
13:38:34 <plh> Dom: might a mis-configuration between bikeshed/spec-prod
13:39:04 <florian> q?
13:39:06 <plh> Denis: I'll follow up on this
13:39:28 <tidoust> [Relevant entry in Bikeshed doc: "Previous Version: from biblio" https://tabatkins.github.io/bikeshed/#metadata-previous-version ]
13:39:30 <dom> [you don't have to do that]
13:39:39 <plh> Mike: we shouldn't have to manually set the previous version
13:39:42 <plh> Denis: agreed
13:40:15 <plh> Denis: regarding different working group needs, I discussed with Sid about some WGs having different documents
13:40:28 <dom> https://github.com/w3c/spec-prod/pull/27 re previous version handling
13:40:31 <sidvishnoi> In progress: https://github.com/w3c/spec-prod/issues/31
13:40:33 <plh> ... I believe Sid is monitoring the issue
13:40:55 <plh> ... so that's in the plan
13:41:01 <plh> q?
13:41:11 <plh> Florian: that would be nice indeed
13:41:27 <dom> [btw, SO MANY THANKS TO SID!!!]
13:41:30 <tidoust> q+ to ask about current status re link checker
13:41:36 <plh> ack tid
13:41:36 <Zakim> tidoust, you wanted to ask about current status re link checker
13:41:56 <plh> Francois: the linkchecker has been disabled in spec-prod
13:42:08 <plh> ... it's useful but prone to find false/positive
13:42:43 <plh> Sid: yes, it was generating too many false positive. looking into fixing some other issues next week as well
13:43:00 <MikeSmith> q?
13:43:04 <plh> Sid: it should work for a publication
13:43:19 <denis> https://www.npmjs.com/package/href-checker
13:43:24 <plh> Denis: the linkchecker is in GH
13:44:09 <plh> plh: why 2 linkcheckers?
13:44:22 <plh> Denis: the new one will check javascript links
13:44:23 <dom> I maintain the linkchecker on an irregular basis https://github.com/w3c/link-checker/
13:45:07 <plh> Denis: Echidna doesn't check the links
13:45:24 <plh> ... I'm thinking about further integration of the new linkchecker
13:45:26 <MikeSmith> q+ to say thanks to Sid
13:45:31 <plh> ack Mike
13:45:31 <Zakim> MikeSmith, you wanted to say thanks to Sid
13:45:38 <dom> +1
13:45:41 <denis> +1
13:45:43 <plh> Mike: Thank you Sid
13:45:50 <wseltzer> +1
13:45:51 <tidoust> +1
13:45:54 <xfq_> +1
13:45:55 <plh> ... this has been a massive timesaver
13:45:57 <xiaoqian> +1
13:46:06 <plh> ... for /TR and github.io URLs
13:46:30 <wseltzer> MikeSmith++
13:46:35 <plh> ... /me is looking forward to deploy spec-prod for pointerevents
13:46:37 <dom> sidvishnoi++
13:46:40 <wseltzer> sidvishnoi++
13:46:47 <plh> Denis: thank you Sid
13:46:51 <florian> q?
13:46:54 <tidoust> q+ to wonder about additional files during publication to gh-pages
13:46:54 <florian> q+
13:47:07 <plh> ack tid
13:47:07 <Zakim> tidoust, you wanted to wonder about additional files during publication to gh-pages
13:47:26 <plh> Francois: one clarification: publication in GH pages/branches
13:47:53 <plh> ... if there is a markdown document along, should I worry about them or will they get published
13:48:19 <plh> Sid: sped-prod is only for specifications, so unless it's embedded, it won't. Can look into it
13:48:36 <plh> Francois: ok, this means I need to manually add to the action
13:48:44 <dom> ack tidoust
13:49:23 <plh> ... the good thing with using GH pages, you end up with links not using the name of the branch
13:49:30 <plh> ... and GH translates from md to HTML for you
13:49:39 <plh> Sid: I'll add an option
13:49:44 <plh> ack florian
13:50:49 <plh> Florian: for transitions to CR or FPWD, can we redirect to a form to file the transition?
13:51:06 <plh> Denis: we'd thinking about improving the echidna report
13:51:18 <plh> ... since it's not always easy to understand it
13:51:28 <plh> ... we can point folks to the right direction
13:51:43 <dom> [anchoring the transition request flow into the publication request flow rather than the other way around]
13:53:38 <wseltzer> denis++
13:53:55 <plh> rrsagent, generate minutes v2
13:53:55 <RRSAgent> I have made the request to generate https://www.w3.org/2021/03/18-pub-minutes.html plh
13:54:36 <plh> rrsagent, make logs public-visible
13:55:50 <caribou> caribou has left #pub
13:57:43 <Yves> Yves has left #pub
15:57:42 <Zakim> Zakim has left #pub
15:58:36 <plh> RRSAgent, bye
15:58:36 <RRSAgent> I see no action items