Gábor Hojtsy (he/him) |
Drupal 10 uses PHPUnit 9, PHPUnit 10 was released in the meantime and PHPUnit 11 was also recently released. |
Gábor Hojtsy (he/him) |
The main issue is #3217904: [meta] Support PHPUnit 10 in Drupal 11 where @mondrake and @neclimdul and @longwave have been working to piece apart the required changes. |
Gábor Hojtsy (he/him) |
A major part recently landed in #3402444: Deprecate Test Suites, no longer available in PHPUnit 10 but was unfortunately rolled back as it conflicted with how the performance tests are ran. (edited) |
Gábor Hojtsy (he/him) |
There are a lot more in the main issue that people can get involved with :slightly_smiling_face: |
mondrake |
...and even more if you look at issues tagged PHPUnit 10, https://www.drupal.org/project/issues/search?projects=&project_issue_fol... |
mondrake |
...and yes, still a LOT to do, so all hands on deck :grinning: |
longwave |
this is perhaps both the hardest and least rewarding of all the D11 tasks, but we still need to do it! |
Gábor Hojtsy (he/him) |
@Spokje, @andypost and @longwave were involved in this recently :slightly_smiling_face: #3394694: [Meta] Symfony 7 compatibility -- this one definitely needs an issue summary update with more of the problems identified based on the attempted update. |
andypost |
Summary looks updated, 2 issues are actionable |
Gábor Hojtsy (he/him) |
@andypost hm, @longwave commented a few days ago that "The list of tasks is incomplete, there are more than those listed." |
andypost |
Probably for new depreciations |
longwave |
ContainerAware and the missing void return types break everything right now but once those are done we are close I believe |
Gábor Hojtsy (he/him) |
Please post specific issues where people can help if that is useful :slightly_smiling_face: |
longwave |
Void returns issue is RTBC, people can help on #3397519: [META] Reduce use of ContainerAware classes where possible where each usage needs to be looked at individually |
andypost |
Btw probably the composer's upgrade to 2.7.x is required to be backported even to 10.2 https://github.com/composer/composer/security/advisories/GHSA-7c6p-848j-... |
Gábor Hojtsy (he/him) |
@andypost but composer 2.2.23 has this backported as per that advisory? (edited) |
andypost |
Yes, because 2.2 is the current LTS release |
Spokje |
#3421371: Security update composer/composer (CVE-2023-43655) |
Gábor Hojtsy (he/him) |
See #3265680: Deprecate dependencies, libraries, modules, and themes that will be removed from Drupal 11 core for a list of those modules and other smaller pieces planned to be depreacted. |
Gábor Hojtsy (he/him) |
One process improvement recently suggested by @catch was to bundle the database dump removals for multiple modules in #3414563: Add new 10.3.x database dump fixtures, without modules deprecated for removal in 11.x |
andypost |
I bet it mostly blocked on depreciation of plugins #3039240: Create a way to declare a plugin as deprecated |
Gábor Hojtsy (he/him) |
@andypost I don't think modules themselves are blocker on that(?) but deprecating other things may be possible once there is a way to deprecate plugins indeed :slightly_smiling_face: that said, there is not that much time left at this point? |
catch |
Yeah entire modules can be deprecated, we don't deprecate individual APIs because the contrib version provides them. A lot of these are very close now. |
catch |
Field layout shouldn't be on the list here, it's supposed to be replaced in core then marked obsolete eventually. |
catch |
It should be on a list, just not this specific one. |
andypost |
Yep, not replaced but incorporated into core |
Gábor Hojtsy (he/him) |
It would be good to get done at one point though, its a lot more complicated than moving to contrib... |
andypost |
Then it needs a product manager review |
Gábor Hojtsy (he/him) |
ok I provided "product manager review" in #3007167: [META] Enable layout builder for form displays, and deprecate field_layout#comment-15438456 which is more about accepting the bleak situation the module is in |
andypost |
Maybe @timplunkett (he/him) can mention about sense to reroll [#2844302]As this is only blocker |
andypost |
And 150k patch surely needs split to be reviewed |
Gábor Hojtsy (he/him) |
Last time Tim posted on the issue was 5 years ago, so I think its safe to say its not a priority :smile: |
Kristen Pol (she/her) |
That’s a lot! |
quietone |
The last issues for Book, Forum and Tour are for removing them from themes. I am not much use on these so asking for help and reviews. |
dww |
I see Archive_Tar on the list for research. That's only used for Update Manager AFAIK. I don't believe we have an issue yet for formally deprecating the "Update Manager" parts of update.module (the install-new-code UI, authorize.php, etc) when autoupdates is stable. I don't think there's much, if anything, to deprecate. Very hard to believe contrib has extended any of that weird plumbing. :sweat_smile: But do we need to put this on the radar for things we're planning to deprecate / remove? |
dww |
There's #3285191: [meta] Only support Drupal core installs managed by composer, but that's not the same scope of what I'm talking about |
dww |
Ahh, clicking through enough rabbit holes, I find #3417136: Remove adding an extension via a URL -- that's most of it |
dww |
Tried to summarize at #3285191: [meta] Only support Drupal core installs managed by composer#comment-15442566 |
Gábor Hojtsy (he/him) |
@Björn Brala (bbrala) and I worked on updating the bot to post Drupal 10 to 11 patches now. @Björn Brala (bbrala) tested it on jsonapi_extras :slightly_smiling_face: However now we need to make it open MRs instead. Any takers? :slightly_smiling_face: |
Björn Brala (bbrala) |
I wrote an outline in an issue for that. |
Björn Brala (bbrala) |
#3286278: Project Update Bot should create MRs instead of using patches |
Björn Brala (bbrala) |
in a comment, is update NOT done. was just dumping an overview, sorry (edited) |
Gábor Hojtsy (he/him) |
ah commented there then :slightly_smiling_face: |
Björn Brala (bbrala) |
Edited, IS was not updated |
Gábor Hojtsy (he/him) |
Exciting that this may be within reach :slightly_smiling_face: Do you have time to start experimenting with this or should I? :slightly_smiling_face: |
Björn Brala (bbrala) |
The hoest aanser is that it will be hard for me to find the time and energy right now. |
Björn Brala (bbrala) |
(Autocorrect was in dutch oops) |
Gábor Hojtsy (he/him) |
Ok I'll take a look, should I generate a "random" ssh key and add it to the user? ;) |
Björn Brala (bbrala) |
YES |
Björn Brala (bbrala) |
thats how i would do it |
Björn Brala (bbrala) |
then add the PK to the secret in the repo |
Kristen Pol (she/her) |
Thanks for the work on this :heart: |
Björn Brala (bbrala) |
I've done some more initial work on this as a morning exercise. All untested until now, but the flow is getting clear. Still need to do IS maintenance lol :stuck_out_tongue_winking_eye: |
Gábor Hojtsy (he/him) |
Ah sorry I did not get around to this yet. |
Gábor Hojtsy (he/him) |
Thanks for your work! |
Björn Brala (bbrala) |
No worries. Had some spare time for some simple stuff |
Gábor Hojtsy (he/him) |
Suggested by @larowlan |
larowlan |
What's the next steps here, are we just going to do them one issue per plugin type? |
Gábor Hojtsy (he/him) |
Summoning those who were involved may help to discuss this. Who are those people? :) |
larowlan |
@longwave @Björn Brala (bbrala) |
catch |
I think it's one issue per plugin type yeah. We might not even want to deprecate annotations in Drupal 10 assuming we remove in 12 but supporting everywhere would be good. |
longwave |
we should start opening some new issues for plugin types, ideally the easiest ones first that don't have many properties to consider |
larowlan |
Will discuss this internally at work to see if we can take a lead on this |
larowlan |
Added child issues for all of them #3396165: [meta] Convert all core plugin types to attribute discovery |
Gábor Hojtsy (he/him) |
Topic proposed by @larowlan |
larowlan |
The issue is RTBC and has been approved by Dries |
larowlan |
I think the next step is an MR to update governance repo then we call for nominations and work with drupal.org maintainers on getting a new role (if needed) |
Gábor Hojtsy (he/him) |
#3406954: Proposal - create a 'Project Update Working Group' |
Gábor Hojtsy (he/him) |
Assuming the MR will be the text proposed in the issue I am happy to commit to speed this up :) |
larowlan |
It will take me away from the RTBC queue but I'll try to get that MR up today |
larowlan |
Sooner we get this rolling etc |
Kristen Pol (she/her) |
Fantastic work :heart_eyes: |
larowlan |
opened https://git.drupalcode.org/project/governance/-/merge_requests/19 for review |
greg-boggs |
Looks great. When can we start applying? :) |
larowlan |
Once it has been committed to governance project I'll do an announcement |
greg-boggs |
Super exciting. Thanks larowlan |
Gábor Hojtsy (he/him) |
@mglaman proposed himself and @Björn Brala (bbrala) for #3400254: Add a developer tooling topic to Drupal core |
Gábor Hojtsy (he/him) |
More feedback would be welcome there to help craft this :slightly_smiling_face: There was/is some tinkering with what to name it exactly to make it clear and avoid confusion. |
mglaman |
I proposed this in a heat of issues. But I'm not sure how this would really work |
mglaman |
It seems like it'd be better with some kind of CI job than personal manual intervention. Maybe now that we have GitLab CI it would be easier |
Gábor Hojtsy (he/him) |
What would the CI job look for / test? |
mglaman |
:melting_face: I guess I don't know. Shoot. I guess it's more of a test-driven approach by flagged deprecations and verifying the PHPStan job fails |
mglaman |
So it's still a process problem. If that fails we need phpstan-drupal work |
Gábor Hojtsy (he/him) |
@mglaman you are saying the CI job would use some kind of baseline and check if a @deprecated or some other marker was added in the MR, that phpstan now reports more than the baseline? Or or based on expected deprecations introduced in tests or somesuch given that it is more likely to fail where not existing deprecation infrastructure is used (such as with annotations) |
mglaman |
I need to think on how it could work. The problem is that the deprecation is usually added and verified as “fix” by removal and a unit test. Not necessarily that it was found/fixed via PHPStan. And that if it wasn’t caught by PHPStan, an issue isn’t necessarily opened or mentioned on the CR |
mglaman |
It’s like a test-only run… but only of the added deprecations :sweat_smile: which is impossible to know. Especially if it’s a big one like the enforced accessCheck |
Gábor Hojtsy (he/him) |
Well, it is possible to know if there was added test coverage for the new deprecation? |
Gábor Hojtsy (he/him) |
Which I think may be a usual scenario? |
mglaman |
That should be a usual scenario. I just can’t fathom how we could run PHPStan with only the deprecation tags or such add… I guess it’s really “hey this patch introduced trigger_error but no @deprecations added to code! |
Gábor Hojtsy (he/him) |
Well, that is a simple regex on the diff? :stuck_out_tongue: |
mglaman |
yeah, that one seems a lot easier than something complex |
mglaman |
I’m trying to think of a really “dumb” solution that could have helped catch things. Like “oh, this failed. it’s not a simple deprecation, we should make sure the tooling for detecting this is provided” |
Gábor Hojtsy (he/him) |
Issue at #3214954: [11.x] [meta] Set Drupal 11 platform and browser requirements at least six months before the release |
Gábor Hojtsy (he/him) |
Topic proposed by @Juraj Nemec |
Gábor Hojtsy (he/him) |
My understanding there is that MySQL/MariaDB, SQLite and Postgres are all set? What's missing @Juraj Nemec? |
Juraj Nemec |
I see the child issues (PostgreSQL, SQLite and MySQL) as RTBC, so probably close to be done, but not fixed yet? |
Juraj Nemec |
But it is great that we have PHP requirements already set, which is more important I think. |
Gábor Hojtsy (he/him) |
I think all of the child issues are RTBC so this would also be either marked fixed or RTBC? cc @catch |
catch |
I kept wanting to mark fixed but I think we need to update d.o docs first. |
Gábor Hojtsy (he/him) |
platform requirements pages? |
Gábor Hojtsy (he/him) |
this one for example? https://www.drupal.org/docs/getting-started/system-requirements/overview... |
catch |
Lol I didn't know it was there. So we could just update that for now. Eventually it would be nice if the top level page had a clear path for 10 Vs 11 |
Gábor Hojtsy (he/him) |
I'll look into this tomorrow :slightly_smiling_face: |
andypost |
Sqlite docker image split needs opinions as sqlite is bundled #3419616: Create PHP 8.3 image containing sqlite 3.45 |
quietone |
Doing the work for these has been on my list since last week. I've just been distracted with other things in my life. But it looks like I can get to it today. |
quietone |
The Overview page listed above means that there are two documents in the same guide that must be changed. Honestly, I don't want to maintain the info in two place. I would prefer if committers only updated the specific requirement pages. |
quietone |
the docs are updated but not the Overview. The Overview has links to the individual requirements pages anyway. The policy issues are Fixed, and follow ups exist. |
quietone |
For implementation, waiting on SQLite 3.45 and MariaDB 10.6 environments. |
Gábor Hojtsy (he/him) |
Yay thanks! @hansfn also noted in the mysql issue that we need an INSTALL.txt update issue postponed to when the 11.x branch is open for 11.x things :slightly_smiling_face: |
Gábor Hojtsy (he/him) |
Opened #3421175: Update INSTALL.txt and hook_requirements() etc. with remaining Drupal 11 platform requirements |
larowlan, Juraj Nemec, Kristen Pol, Gábor Hojtsy, catch, longwave, mondrake, andypost, Spokje, quietone, dww, Björn Brala (bbrala), greg-boggs, mglaman
Comments
Comment #2
gábor hojtsyComment #4
mikelutzComment #15
mikelutzComment #16
mikelutzComment #17
smustgrave CreditAttribution: smustgrave at Mobomo commentedWent into #d11readiness channel on Feb 12th 2024 and verified all threads have been captured.
Everyone that participated in a thread appears to be credited as well.