Meeting will happen in #d10readiness on drupal.slack.com.
| Kristen Pol (she/her) |
Kristen, Santa Cruz, California, contributor (edited) |
| xjm |
I am here! Got distracted briefly by, uhhhh, the lyrics to a Moana song cough |
| FeyP |
Patrick, contrib module developer |
| shaal |
Ofer Shaal, DrupalPod :sunglasses: |
| Ilcho Vuchkov (vuil) |
Ilcho, Bulgaria EU |
| mglaman |
Matt. WI,US. drupal-check/phpstan-drupal/php8 |
| longwave |
Dave, UK, core contributor |
| Gábor Hojtsy (he/him) |
Gábor, Drupal 10 coordinator |
| catch |
Nat from the UK |
| larowlan |
hey there, lee from AU |
| kimb0 |
Kim, au :wave: |
| joshmiller |
Hey all. Josh Miller, lives in Indiana, works on 26 different Urban Institute sites. Majority are D7, 1/3 are D8, and 2 are D9 :smile: Just keeping up with the progress on D10 |
| Gábor Hojtsy (he/him) |
It seems to be that time now that we need to start tracking these. |
| Gábor Hojtsy (he/him) |
See https://www.drupal.org/about/core/policies/core-release-cycles/schedule#... |
| Gábor Hojtsy (he/him) |
The Drupal 10 branch will be open in 2-3 months from now, when major dependency updates become available. Especially Symfony 5.4 and 6.0. |
| Gábor Hojtsy (he/him) |
Similar to Drupal 9, Drupal 10 has 3 scenarios based on when beta requirements are complete:If beta is done by March 18, release is June 15, 2022. In 317 days! :open_mouth: If beta is done by May 13, release is August 17, 2022.Last resort: If beta is done by September 9, release is December 14. |
| Gábor Hojtsy (he/him) |
For Drupal 9, we made the first scenario :slightly_smiling_face: |
| xjm |
It is also worth noting that the deprecation deadline for D10 deprecations is the beta deadline of 9.3 which is on November 8 in 98 days. We also hope to finalize the PHP and DB requirements around then. |
| Gábor Hojtsy (he/him) |
@xjm raised this |
| Gábor Hojtsy (he/him) |
see #3222947: Decide whether to move Quick Edit to contrib |
| xjm |
See #3222947: Decide whether to move Quick Edit to contrib |
| Gábor Hojtsy (he/him) |
Feedback to removing it from core has been almost entirely supportive so far (to my personal surprise BTW). |
| xjm |
Next thing we need is someone willing to maintain it in contrib, or at minimum help with any security issues with it for the next two years or so |
| xjm |
(FWIW I am also going to try to help us get any known sec issues with it fixed before we move it) |
| Gábor Hojtsy (he/him) |
We also need product manager signoff on the issue that I’ve seen, neither Angie nor Dries chimed in yet. |
| xjm |
(Says the product manager?) :P |
| Gábor Hojtsy (he/him) |
I pinged the question to Dries, not sure how much that will help :slightly_smiling_face: |
| Gábor Hojtsy (he/him) |
I think this is a question with definite impartiality for all core product managers, since Dries was involved as a funder in making QuickEdit happen while Angie was a manager on it and I was an employee involved with it. |
| Gábor Hojtsy (he/him) |
So I am seeking wider input from all of them :slightly_smiling_face: |
| xjm |
Yeah Angie might not have much time at the moment and Dries is traveling too. I see your point though. Maybe we should reach out to yoroy as the one PM who was not involved in its development? |
| xjm |
And maybe write up a succinct summary of the "why" for them so they don't have to read the whole issue (although the flood of +1s are also useful info in themselves I guess) |
| Gábor Hojtsy (he/him) |
I think the issue summary is pretty good already? Do you think of something else to be written up? |
| xjm |
@Gábor Hojtsy (he/him) I dunno; catch and I wrote most of the IS and I am never sure if my writing is sufficiently an executive-level summary or not :slightly_smiling_face: |
| xjm |
But if you think it's sufficient then that probably means it is :slightly_smiling_face: |
| longwave |
what happens if nobody steps up to maintain it in contrib? (not quite a parallel, but look at the complete lack of activity on simpletest in contrib) |
| xjm |
We've never not found a sucker, err, maintainer before, so that would be interesting to try to solve |
| joshmiller |
FWIW, we have a staff of 30 Staff content editors at the Urban Institute. 1 is Drupal-specialist. All of them interact with Drupal weekly. None of them use Quick Edit. |
| xjm |
@joshmiller Could you post that to the issue? #3222947: Decide whether to move Quick Edit to contrib |
| joshmiller |
@xjm :you-got-it-dude: |
| Gábor Hojtsy (he/him) |
Palantir.net has been sponsoring @mglaman to work on Drupal Rector |
| Gábor Hojtsy (he/him) |
Thanks to this effort, a HUGE part of deprecated API uses in contrib are now covered by rector. |
| shaal |
:tada: |
| Gábor Hojtsy (he/him) |
Screenshot 2021-08-02 at 20.26.05.png |
| xjm |
That graph is legit fantastic |
| Gábor Hojtsy (he/him) |
That and other graphs are at https://dev.acquia.com/drupal10/deprecation_status/graphs as well as contrib readiness results in other tabs. |
| Gábor Hojtsy (he/him) |
Overall 60% of contribs only exhibit info.yml file incompatibility as much as we can detect currently. |
| Gábor Hojtsy (he/him) |
We are still going to introduce deprecated APIs and extensions in Drupal 9, so this will get worse a bit still. |
| xjm |
But only for 98 more days aside from possibly certain critical-but-easy-for-site-owners deprecations of legacy path modules and themes :slightly_smiling_face: |
| xjm |
Right now I think PHP 8.1 is adding a lot more disruption than core itself in terms of deprecations |
| larowlan |
Awesome work |
| larowlan |
We are getting close to committing #3223209: deprecate file_save_data, file_copy and file_move and replace with a service which will upset that a bit |
| mglaman |
PHPStan is missing a lot of checks. We have a meta issue going |
| Gábor Hojtsy (he/him) |
I’ve been unsuccessfully experimenting with this but failed so glad @mglaman picked it up :smile: |
| mglaman |
One of my first steps was to get internal functions recognized as deprecated by parsing stubfiles. |
| mglaman |
Like each, create_function, etc |
| Gábor Hojtsy (he/him) |
@mglaman hm, is it feasible to do though? |
| mglaman |
Yep :slightly_smiling_face: |
| mglaman |
I’ve got the PR into BetterReflection, copied a fix from the upstream AND support for JetBrain’s own Deprecated PHP attribute they use |
| mglaman |
Basically I’m walking around like a fool and Ondrej is the parent correcting my path |
| Gábor Hojtsy (he/him) |
Raised by @FeyP |
| Gábor Hojtsy (he/him) |
The process is the same, the tools are the same. Upgrade Status, Drupal Rector, etc. |
| mglaman |
I’d say the same and easier.easier because:round 2, lessons learnedComposer isn’t dropping a major release or rewrite |
| FeyP |
Very nice. Thank you! And when would be a good time to start replacing APIs in contrib? |
| Gábor Hojtsy (he/him) |
Well, I would add the caveat that Drupal 9 to 10 will have more frontend changes. So those may or may not be as easy to follow for contrib. Especially around CKEditor 4 to 5. |
| Gábor Hojtsy (he/him) |
@FeyP I will post the contrib question in its own number :smile: (edited) |
| FeyP |
Ah, okay, sorry :slightly_smiling_face: |
| xjm |
And also:Two years of new APIs and deprecations instead of fiveWe already added the important module compatibility APIs for the info files that are 60% or whatever of contrib modules' upgrades |
| Gábor Hojtsy (he/him) |
Raised by @FeyP |
| Gábor Hojtsy (he/him) |
Two things IMHO |
| Gábor Hojtsy (he/him) |
The deprecated APIs list is still trending up: https://dev.acquia.com/drupal10/deprecation_status/graphs bottom graph |
| Gábor Hojtsy (he/him) |
2. Most deprecated API uses we found in contrib are Drupal 8 APIs deprecated for removal in Drupal 10. These can be resolved as soon as you are ready to drop Drupal 8 support. Such as when Drupal 8 goes EOL in 3 months. |
| Gábor Hojtsy (he/him) |
So there is a lot you can already do. |
| xjm |
It depends I think on whether you want to do a few small updates, or one big one |
| xjm |
Some module maintainers might like to continually improve their D10 compatibility, and that's OK to do now |
| xjm |
If you only want to do one or two passes I would wait until December |
| Gábor Hojtsy (he/him) |
Also as a separate effort, making your project compatible with PHP 8 is something you can do already. If you still support Drupal 8 make sure you still support PHP 7 though as Drupal 8 cannot be updated to support PHP 8. See also thread 5️⃣ for our automation efforts to help find PHP 8 incompatibility errors even on PHP 7. (edited) |
| xjm |
And if you really can only look at it once, wait until D10 hits a beta |
| xjm |
(Which is March at the earliest) |
| xjm |
So depends on your maintainer style basically :slightly_smiling_face: |
| xjm |
Oh I would also definitely wait if your module interacts with core JavaScript APIs in particular |
| xjm |
That is one area where there will probably be many new deprecations in the next few months |
| FeyP |
Thank you, very helpful! |
| shaal |
@xjm re: core JavaScript APIs changes, do you have an example for such change? |
| xjm |
@shaal Anything that integrates with CKEditor, Autocomplete, the Dialog UI, or Backbone especially |
| shaal |
Would we need tools beyond Rector to catch these deprecations ? |
| Gábor Hojtsy (he/him) |
Well, CKEditor 5 will be entirely different, so if one has a CK4 integration contrib module, I am not sure if we can offer an automated rectory-style mapping as to what to do about it. We’ve been asking people to try to port smaller projects, but I am not sure its been hapening yet. (edited) |
| xjm |
@Gábor Hojtsy (he/him) I forget which tools detect JS deprecations; do you remember? I couldn't find issues related to it on the roadmap issue or our internal spreadsheet |
| xjm |
We have a standard for them and an API in core to notify the developer |
| xjm |
But I can't remember beyond that where they get flagged |
| xjm |
https://www.drupal.org/about/core/policies/core-change-policies/drupal-c... that's the core docs |
| Gábor Hojtsy (he/him) |
I thought we would have such tools listed in #3096679: JavaScript calls to deprecated APIs are not found (for Drupal 10) if we identified any, but they are not listed there |
| xjm |
@shaal To a large extent it will be deprecations of whole modules or libraries, which Upgrade Status detects for sure |
| xjm |
Or will detect in the case of modules and already does detect for libraries, that is |
| xjm |
Like we are wholesale deprecating jquery_ui.* and ckeditor.module |
| xjm |
But I am going to make a meta for #3105708: Properly deprecate Drupal.Ajax.element_settings and friends and add them all properly to the issue hierarchy |
| shaal |
:+1: thank you |
| Gábor Hojtsy (he/him) |
@shaal raised this |
| Gábor Hojtsy (he/him) |
the bot is https://www.drupal.org/u/project-update-bot and so far it posted Drupal 8 to 9 patches generated automatically based on Upgrade Status and Drupal Rector |
| Gábor Hojtsy (he/him) |
some maintainers are not active (anymore) though and even reviewed patches are not getting committed |
| Gábor Hojtsy (he/him) |
three is an issue at #3168047: Commit Project Update Bot patches to under-maintained projects for Drupal 9 compatibility |
| Gábor Hojtsy (he/him) |
@xjm was about to write up a policy doc based on discussions with the DA about what is going to happen (automated commits to unmaintained projects under certain conditions) |
| xjm |
Yes, I got mentally blocked a bit on the fact that we don't have an actual docs page other than the project page explaining Project Update Bot. But working on this finally today. |
| xjm |
Will share it with others this week so we can get moving on the actual process. |
| xjm |
Initially we will only commit RTBC patches, created by the bot itself, where the maintainer has not responded in 4 weeks. And create a new branch and tag for that commit. Maintainers will be emailed before it happens and can opt out (TBD how, maybe the same way as opting out of the bot generally). |
| xjm |
We will not auto-commit patches created by humans or ones that are not RTBC |
| xjm |
Or ones that fail tests if there is a test suite |
| xjm |
BUT if you are a human who has created a D9 patch for a module, consider offering to maintain the module and following the abandoned module process if you don't hear back from current maintainers |
| FeyP |
Will the commit bot detect, if there already is another compatible branch without release? And what would happen in that case? |
| FeyP |
Also, if the commit bot creates a new branch, will it also create a relase form that branch or just post about the new branch in the issue? |
| Gábor Hojtsy (he/him) |
The both should pick up the latest branch I believe. |
| xjm |
@FeyP We were thinking that it would tag an alpha on the new branch, so that there is something to test, but it's clear that it is very tentative and doesn't provide API/upgrade path/etc. promises. |
| drumm |
Draft query for issues that match the criteria so far. |
| xjm |
Just one note on the end bit of the query, the ^9 string will not correctly match once we commit >=8.9 versions. Also project could already technically have that syntax for an info string though it's probably not common. |
| drumm |
Right, nothing indexes semver compatibilities, which is why we push back on a lot of “show things that are compatible with X” requests - they aren’t at all practical to write. In this case, can look for new things showing up in the core_version_requirements result column and update the query as-necessary. |
| xjm |
Sounds like a good approach to me |
| xjm |
@drumm Gábor pointed out that we should also add a condition to the query that the issue creation date is at least four weeks ago. (The maintainer would not have responded in 4 weeks if the bot were to open an issue for something tomorrow that wasn't fixable before.) |
| drumm |
That’s easy enough to add |
| Gábor Hojtsy (he/him) |
I think this will not limit the number currently as the bot did not run in four weeks I think but it will limit once the bot starts to run again. |
Comments
Comment #2
gábor hojtsyComment #12
gábor hojtsySaving notes. Thanks all!
Comment #13
gábor hojtsy