Drupal 10 readiness meeting / 6 July 2020 [#3153934]

Meeting will happen in #d10readiness on drupal.slack.com.

Hello and welcome to this Drupal 10 readiness meeting!

This meeting:
➤ Is for core and contributed project developers as well as people who have integrations and services related to core. Site developers who want to stay in the know to keep up-to-date for the easiest Drupal 10 upgrade of their sites are also welcome.
➤ Usually happens every other Monday at 18:00 UTC.
➤ Is done over chat.
➤ Happens in threads, which you can follow to be notified of new replies even if you don’t comment in the thread. You may also join the meeting later and participate asynchronously!
➤ Has a public agenda anyone can add to: `https://www.drupal.org/project/drupal/issues/3153934`
➤ *Transcript will be exported and posted* to the agenda issue. For anonymous comments, start with a :bust_in_silhouette: emoji. To take a comment or thread off the record, start with a :no_entry_sign: emoji.

0️⃣ Who is here today? Comment in the thread below to introduce yourself.

atul4drupal	Atul from India
shaal	Ofer Shaal, Florida :sunglasses:
andypost	Andy
Kristen Pol (she/her)	Kristen, Coarsegold, CA, USA
surabhi.gokte	Surabhi, India :wave::skin-tone-2:
dww	Derek, US, Black Lives Matter :fist:
xjm	doh I was waiting for activity in the other channel, sorry
dan2k3k4	Dan, Zurich
mixologic	Ryan, DA Infra /Testbot/Packages/etc
Gábor Hojtsy (he/him)	Gábor from Hungary
voleger	Oleh, Ukraine
catch	Nat, UK, very late.
larowlan	Lee, Aus, laterer
Hemant Gupta	Hemant, India, Syncing through the threads :slightly_smiling_face:
pratik_kamble	Pratik, Pune India, Syncing now....
hardik_patel_12	Hardik, India, Syncing now
hansa11	Hansa, India
kimb0	Kim, Australia, laterest

1️⃣ Do you have suggested topics you are looking to discuss? Post in this thread and we’ll open threads for them as appropriate.

andypost	Php 8 progress and #2928856: Adopt the PSR-12 standard for PHP return types
surabhi.gokte	DrupalCon is almost here, how excited are you??!!
Kristen Pol (she/her)	Will there be an open social group for Drupal 10 readiness (I see the Drupal 9 porting group) or people will work in topic specific groups instead?
shaal	Are there things we should prepare before that?#3155358: [policy, no patch] Drop IE11 support from Drupal 10.0.x
xjm	Should we talk about how to proceed with jQuery vanillication?
xjm	#3052002: [meta] Replace JQuery with vanilla Javascript in core
xjm	Maybe we should spin out a separate thread around CKE5
xjm	(the strategic decision piece)
andypost	++ to split as cke5 means to build custom editorbut issue for "modern editor" stuck in bikeshed
voleger	What is the state of Drupal JS Initiative in Drupal 10 https://www.drupal.org/about/strategic-initiatives/admin-ui-jsSpecially React admin initiative? Is it on he roadmap for Drupal 10 release?
xjm	Major releases don't have roadmaps
xjm	We can talk about the general topic in a thread, but major releases are based on scheduling needs and security requirements, not features :slightly_smiling_face: We don't ever block releases on features
voleger	Got it. Thanks.
andypost	Moreover React moving much faster then Drupal and it will be deadend to maintain

2️⃣ Drupal 10 is planned for summer 2022 #3118143: [meta] Release Drupal 10 on December 14... or 15... 2022

Kristen Pol (she/her)	Looks like the Drupal 7 EOL date needs to be updated in the image in that issue
Gábor Hojtsy (he/him)	@Kristen Pol (she/her) yeah, can you post a comment to that effect? thanks
Kristen Pol (she/her)	Sure
dww	So this is the meta tracking major efforts that need to happen?
xjm	I have an updated image already that I can post
xjm	And yes, that's the meta-of-metas. Most things should be children of the already-defined children. Direct child metas should probably be defined by committers.
Kristen Pol (she/her)	https://twitter.com/kristen_pol/status/1280205697557258240?s=20
kimb0	Summer? Which hemisphere? :wink: (edited)
Gábor Hojtsy (he/him)	So right @kimb0, sorry for that.
kimb0	np

3️⃣ PHP 8! @alexpott has been doing some impressive work at #3156595: Make Drupal 9 installable on PHP8

andypost	Would be great to promote somehow php8 docker image to get more eyes and hands to try it
andypost	Next steps could be make performance tests because jit settings are flux (could change soon) but affects seriously - https://beberlei.de/2020/07/05/what_to_look_out_for_when_testing_php_jit... https://stitcher.io/blog/jit-in-real-life-web-applications (edited)
alexpott	The only reason this isn’t passing tests is because of a bug that would affect PHP 7 too if we didn’t have the pecl yaml extension installed :)
andypost	I gonna try build it, will report the state
mixologic	@alexpott is there a need for pecl yaml-less php7 env?
andypost	Just checked - yaml 1.2 installs and working with php8 (I used pecl install yaml) (edited)
xjm	Do we require pecl yaml?
xjm	I didn't think we did
andypost	no, but our tests making sure that if extension enabled nothing broken
andypost	btw here's related discussion https://drupal.slack.com/archives/C51GNJG91/p1593736274271600?thread_ts=...
andypost	FTR the same issue we have with upload_progress extension, which is badly maintained but status report using to check it
mixologic	I failed to notice ` New Features:- Basic PHP8 (master branch) compatiblity` that in the changelog and only saw the PHP Version: PHP 7.1.0 or newer and didnt think to just try it. I can add that to the php containers.
mixologic	But are this saying that having the optional extention enabled currently is hiding a bug when it isnt enabled?
xjm	Yeah that's what I'm trying to ask also I think
andypost	@mixologic are you about https://github.com/php/pecl-file_formats-yaml which php7 branch is current and actually maintainer responsive enough (edited)
mixologic	@andypost I was saying that I misinterpreted the pecl yaml page to say that it wasnt php8 compatible
mixologic	In any case, #2843553: Need containers with only minimum php requirements, as well as containers with all optional requirements compiled in. is sort of the drupalci issue that factors in the fact that we have a lot of ways things can be configured, and testing all of those ways needs a better mechanism than we have now.
andypost	Using "minimum" for testing may lead to endless tests (without opcache for example)
alexpott	Having the pecl yaml extension installed is hiding a bug in our test framework. It’s not hiding any “real” bug
mixologic	opcache is a defacto drupal requirement.
mixologic	apcu, the key/value cache is the optional thing that might be problematic, for tests.
andypost	@alexpott is there a link to bug?
alexpott	Here’s the issue that will hopefully fix the bug - #3156998: Using @requires extension_name in PHPUnit unit tests fails if extension is not present - but this is seriously gnarly code.
mixologic	@alexpott should I deploy php8 containers with pecl yaml, or wait for this fix ? seems like you can fix it without that
andypost	@mixologic OTOH core9 using composer autoloader which has option for APCu, I bet we never testing ithttps://getcomposer.org/doc/articles/autoloader-optimization.md#optimiza... (edited)
alexpott	Ideally we’d get a PR into symfony to support skipping listed deprecations and to support setting an expected deprecation programmatically. I had a PR for latter sometime back but I lost focus on it.
alexpott	@mixologic happy either way
mixologic	I think I will, just because it simplifies the m4 macro to not skip the pecl step.
mixologic	Im starting a big push to change drupalci in a big way, so environment configs will be much more simplified.
mixologic	and by simplified I mean, configurable.
andypost	@mixologic iirc only xdebug has few compatibility issues, but last weekend I used to build/run it from master without notices
mixologic	we dont really need xdebug on the containers.. its just there as a sort of last step if you really, really need to get deep and debug a test running inside a container… IIRC its a very difficult thing to even get setup, isnt documented either.
mixologic	I guess I can try apcu too: https://pecl.php.net/package-changelog.php?package=APCu&release=5.1.18
mixologic	5.1.17 says fixed compatibility with php8
andypost	yes, it works fine (I used to build it fro 5.1.18)
xjm	Summoning @timplunkett (he/him) to this thread
xjm	@timplunkett (he/him) TLDR #3109885: [meta] Ensure compatibility of Drupal 9 with PHP 8.0 (as it evolves) has a bunch of stuff in the IS now
xjm	And we're close to testability on CI

4️⃣ Adopt the PSR-12 standard for PHP7 return types

Gábor Hojtsy (he/him)	@andypost suggested #2928856: Adopt the PSR-12 standard for PHP return types
andypost	Yes, looks the issue hangs and how to speed-up it
andypost	Basically for new functionality in core we already doing it, so who's attention it waiting for?
xjm	There are a few related issues for typehinting... one mo
xjm	#3154762: [policy, no patch] All new Drupal 10 code should use appropriate type hints wherever possible
xjm	#3050720: [Meta] Implement strict typing in existing code
xjm	#1158720: Improve text for parameter type hinting in function declaration
xjm	All of these issues are related because while we can add the standards for new code, we can't require it for existing code in a lot of cases because of BC, and so we ahve to come up with a plan/process for adding typehinting for existing code with BC in the continuous upgrade path
xjm	One question I have about PSR-12 is can we retain our existing bracing on the same line rather than the newline? The standard does not seem to mention the opening curly brace.
xjm	But the examples have it on a separate line
andypost	Hm... yep, that's strange...
andypost	Then #3154762: [policy, no patch] All new Drupal 10 code should use appropriate type hints wherever possible should be first priority because PHP8 already showed that our DB-layer needs split for pre-php8 and type-hinted-future

5️⃣ DrupalCon plans, DrupalCon Global is next week!

Gábor Hojtsy (he/him)	@Kristen Pol (she/her) and @surabhi.gokte raised this
Gábor Hojtsy (he/him)	I posted a rundown at http://hojtsy.hu/blog/2020-jul-02/learn-about-and-shape-future-drupal-dr... of sessions, BoFs and contribution
Gábor Hojtsy (he/him)	@Kristen Pol (she/her) asked if there will be a “Drupal 10” table/topic at contribution
Gábor Hojtsy (he/him)	I think there should be more individual topics, like CKEditor or Symfony 5 or PHP 8 :slightly_smiling_face:
Gábor Hojtsy (he/him)	There already is Olivero which counts towards Drupal 10 :wink:
Kristen Pol (she/her)	Sounds good
Gábor Hojtsy (he/him)	there is also a LOT of space in the BoFs that is yet unoccuppied
Gábor Hojtsy (he/him)	some great “Drupal 10” BoFs are https://events.drupal.org/global2020/bofs/drupal-9s-accessibility-wcag-2...
Gábor Hojtsy (he/him)	and https://events.drupal.org/global2020/bofs/discussion-drupal-initiative-l...
Gábor Hojtsy (he/him)	and https://events.drupal.org/global2020/bofs/standardize-rector-rules-part-...
surabhi.gokte	Wow!
xjm	I still do not understand what sessions are when so I am completely at a loss about what I can schedule BoF or table-wise
xjm	I do think a generic "D10 readiness" table would be good. We could split off into a "CKEditor 5 table" if stakeholders were there for that but we would need at least someone to be the table lead.
Gábor Hojtsy (he/him)	@xjm based on the tech for tables, it would be more practical to have more of the tables defined up front, as its hard to expect people to show up at the right time across various timezones without pre-planning it a bit
xjm	Right, so if we define tables up front, then we need to ensure that there are people who are interested in working at each table up front
xjm	Like just saying there's a CKEditor table is going to be bad UX if no one is there to collaborate with.
Gábor Hojtsy (he/him)	@xjm yes, absolutely, then they can define when they would be available, which they can make events out of, so people will not wonder into it at random times expecting people be there and radio silence, but expectations would be managed :slightly_smiling_face:
xjm	I actually think CKEditor should be a Hard Problems discussion with the product managers, Wim, and Lauri first. Because we need to commit to CKE5 first (or not).
xjm	I was already thinking of setting up a meeting next week but just forgot it was DrupalCon.. maybe the week after is better
Gábor Hojtsy (he/him)	Yeah hard problems meetings at DrupalCon is not a good time probably. Enough video time already.

6️⃣ Drop IE11 support from Drupal 10.0.x

Gábor Hojtsy (he/him)	raised by @shaal
Gábor Hojtsy (he/him)	#3155358: [policy, no patch] Drop IE11 support from Drupal 10.0.x
xjm	I reviewed this yesterday. I think @andrewmacpherson raised a good point about assistive tech users and I think that's a consideration point we need to include in our heuristics
xjm	Not just global usage, but regional availability, assistive tech, etc.
shaal	I was wondering if there's anything we can prepare in advance for when it would happen (either Drupal10 or Drupal11)for example - reconsider what HTML, CSS, and Javascript functionality we've been avoided until now, that we can finally benefit from.
shaal	Screen reader survey is only done every 2 years :disappointed:But so far, just like users worldwide, the decline of IE11 seems steady
catch	There's been more discussion on #3080068: [policy, no patch] Define usage heuristics for browser support, including the screenreader survey.

7️⃣ Replace JQuery with vanilla Javascript in core

Gábor Hojtsy (he/him)	raised by @xjm
Gábor Hojtsy (he/him)	#3052002: [meta] Replace JQuery with vanilla Javascript in core
shaal	@nod_ ^^
xjm	So @larowlan had an idea about converting just path's JS first as a novice issue, but I wanted to be very careful to make sure people don't go creating issues to clean it up on a per-module basis
xjm	If we can define a list of pattern replacements than we can do them as cross-core cleanups as per https://www.drupal.org/core/scope
xjm	But we could do a few parts as test balloons
xjm	We should also be strategic about what we clean up first, e.g. if we end up rewriting a lot of CKEditor integration then there's no point also churning that first with de-jQuerification
xjm	Also, can we define linting rules for (not jQuery thing, but vanilla ES6 thing)?
Gábor Hojtsy (he/him)	@nod_ posted on twitter that he plans to contribute to this :slightly_smiling_face:
Gábor Hojtsy (he/him)	https://twitter.com/nod_/status/1276981835181043712
shaal	We should figure out if there are certain patterns that needs to be addressed in a global way, like the use of jQuery.once #2402103: Add once.js to core
xjm	We also should file a meta issue listing the core/assets/vendor assets that depend on jQuery, and determine the solution for each (whether to drop it, track the upgrade for it, etc.)
Gábor Hojtsy (he/him)	Who wants to take this on? ^^
nod_	to some extend the "simple" replace, like once without jquery, some selector stuff can be scripted with something like https://github.com/codemod-js/codemod or https://github.com/facebook/jscodeshift (edited)
catch	jquery.once has an issue #2402103: Add once.js to core
larowlan	I think we need an issue to pass along the raw element to behaviors as a second argument too
larowlan	otherwise all the behaviors will be using context.get(0) to get the non jquery context
larowlan	i.e its hard to migrate off jquery when all behaviors get given a jquery object for their detach/attach function argument
nod_	context is already a dom element
nod_	no jquery in behaviors for a while now
larowlan	ah, I was working on some code for d.o a few weeks back, that’s d7, sorry for the noise
nod_	haha got worried for a bit :stuck_out_tongue:
dww	@larowlan Did you intend for #3157328: Use vanilla JS instead of jQuery in path module to be assigned to yourself?
larowlan	Yep
larowlan	Going to walk a new FE contributor through it
nod_	nice, though I'm not sure by module is the right way to do it at the moment. as long as once still needs jQuery almost all of core js require jquery. we tried the "per module" approach for the js clean-up, that took like 2 years just to clean up simple things.
nod_	other issue is lack of dedicated js comitter, as soon as we pick up some momentum with js stuff, it all comes crashing down due to a lack of progress in the code
larowlan	feel free to ping me on that front
nod_	good to know, thx

8️⃣ CKEditor 4 is nearing EOL, need a replacement for Drupal 10

Gábor Hojtsy (he/him)	@xjm raised this
Gábor Hojtsy (he/him)	#3091226: [META] Select the best modern editor for Drupal 9
xjm	So some folks have proposed that, since CKE5 is a rewrite, we could switch to a different editor instead
xjm	However, @wimleers (he/him) has already defined a plan for CKE5 (as of last year) and thinks we don't have time to adopt a completely different editor
xjm	(Like CKE5 is still less work to upgrade than researching a brand new thing, and we're at risk on getting it stable in core in time already anyway)
andypost	Related issue is #2966864: Add optional support for CKEditor 5 in D9 so we can remove CKE 4 from Drupal 10
xjm	I think the decision about whether to use CKEditor 5 or a different editor for D9 has these decision-makers:Product managersRelease managersEditor subsystem maintainersFEFM(edited)
Gábor Hojtsy (he/him)	We should regroup on that and decide soon so we are out of the decision limbo.
xjm	So I was thinking of having a real-time meeting with those stakeholders to discuss it, and possibly either just make the decision to switch to CEK5 (while encouraging other experimental editors to happen in core or contrib if they want) or timebox the feedback time on it
xjm	Anyone have opinions/thoughts on this?
andypost	As there's working patch from "editor maintainer" looks it's good idea
xjm	Uhh there's not a working patch to my knowledge, just a plan based on CKE5 as of a year ago
xjm	Interesting cross-thread point that came up somewhere -- does CKE5 not support IE11? So if core did, we'd either need to explicitly exempt CKE5 (it's not required, after all) or write a shim
shaal	yes, CKE5 does not support IE11
Gábor Hojtsy (he/him)	https://ckeditor.com/docs/ckeditor5/latest/builds/guides/support/browser... not supported yet
Gábor Hojtsy (he/him)	ticket has been open for 4 years https://github.com/ckeditor/ckeditor5/issues/330
xjm	Nice archaeology (edited)
wimleers (he/him)	I think a different text editor is possible if and only if there is a dedicated group of people working on that, with significant sponsored time.
Gábor Hojtsy (he/him)	@wimleers (he/him) what do you think of the relative effort of CK5 vs. something else (I know that is a wild question)
Gábor Hojtsy (he/him)	it sounded like even CK5 needs a dedicated group of people
wimleers (he/him)	Yes, but at least that has A) working relationships, B) known concepts in CKE4, C) Drupal has integrated with those concepts in CKE4, and we can work with the CKE team to accelerate insight in how to map that to CKE5 concepts.All of that needs to be redone entirely with other editors.

9️⃣ State of Drupal React admin UI initiative

Gábor Hojtsy (he/him)	raised by @voleger
Gábor Hojtsy (he/him)	from https://www.drupal.org/about/strategic-initiatives/admin-ui-js
xjm	As I mentioned in the general thread, initiatives/features are added in minor releases under the continuous upgrade path, not major releases
xjm	So if a given initiative's work were added to core in 9.2 for example then it'd be available in 10.0, but otherwise it would be added later in 10.1+
xjm	(Like we did not block 9.0 on Claro or Layout Builder or anything like that either)
voleger	https://github.com/jsdrupal/drupal-admin-uiI don't know the reason why this initiative stopped (based on the repo activity).Maybe there are some prerequirements before continuing working in that direction? (Like stabilize Claro admin theme or so)
Gábor Hojtsy (he/him)	@voleger unfortunately that project did not succeed, last commit was over a year ago and major activity was last March even before that
Gábor Hojtsy (he/him)	https://github.com/jsdrupal/drupal-admin-ui/graphs/contributors
xjm	It's more to do with the contributors' availability and interest, initiative funding, etc.
xjm	I don't think "did not succeed" is quite fair; it uncovered a lot of useful info as well as core tech debt and limitations
Gábor Hojtsy (he/him)	Well it did not reach its goals was what I meant.
andypost	IIRC it's all stuck on waiting for shadow dom and #2915262: Define the new DBlog UI
Gábor Hojtsy (he/him)	I would be surprised if that were the case. That issue had activity last in January 2018
andypost	yes, there were few implementations on different frameworks (in jsdrupal repo) but then more serious question raised about form-api implementation (client side)
andypost	I guess it was main reason that contentacms stuck as well
andypost	But looks there's some work still happens https://github.com/contentacms
berdir	last update I remember is that at least some of the active people there switched to help with claro to help finish that
andypost	And #2915311: JavaScript VDOM library evaluation: integrating web components still has no solution