damienmckenna → created an issue.
Technically the MR adds extra attributes to the Views "parent" field definition. Don't get me wrong, this is awesome to have for Views displays. But it doesn't integrate with Search API so I think we should either re-open this or open a separate issue to handle Search API integration; your call, smustgrave.
Sorry, #10 patched the wrong function. This one should work.
Here's a patch for the 2.0.x branch.
damienmckenna → created an issue.
FYI I'm not seeing this in 2.0.0-beta1?
I think it would be useful to add some test coverage.
I don't know why BookConditionTest is failing..
getAllBooks() already has some sort logic, so lets replace that.
Should the logic actually be in BookOutlineStorage::getBooks() or even BookOutlineStorage::loadMultiple()?
damienmckenna → created an issue. See original summary → .
FYI the hierarchical_term_formatter module might also work instead of this, and doesn't require any hackery to add it to a project.
I also added a composer.json file so that it can be added to a project.
To use this version via composer.json, add this to the top of the "repositories" block:
{
"type": "vcs",
"url": "https://git.drupalcode.org/issue/referenced_term_formatter-3404534.git"
},
Then require the project like this:
$ composer require drupal/referenced_term_formatter:dev-3404534-composer-fix
That should work.
The problem is that the dependencies list in referenced_term_formatter.info.yml said "taxonomy:taxonomy" instead of "drupal:taxonomy". Because there is no composer.json, the Drupal packagist system automatically generates the composer.json file from the contents of the info.yml file, which leads to the problem described here.
The fix was to change the dependency.
damienmckenna → made their first commit to this issue’s fork.
I ran into this where the Media tag was added to the "Manually editable HTML tags" field, but the error message didn't display anything.
Debugging the SourceEditingRedundantTagsConstraintValidator::validate() shows that $tags_plugin_report is empty, and this is the variable passed as an argument to the error message.
FWIW in my scenario this is the value of the tag definition:
<drupal-media data-entity-type data-entity-uuid alt data-view-mode data-caption data-align class>
It's not hardcoded to only work with one vocabulary, but the logic's a little odd if there's more than one enabled.
Removed the 'break' line.
This is the current code:
public function getTree() {
if (!isset($this->tree)) {
$this->tree = [];
/** @var \Drupal\taxonomy\TermStorageInterface $term_storage */
$term_storage = $this->entityTypeManager->getStorage('taxonomy_term');
$tree = [];
foreach ($this->configuration['vocabularies'] as $vocabulary_id) {
if ($vocabulary = Vocabulary::load($vocabulary_id)) {
$tree[$vocabulary_id][$vocabulary_id] = [
'label' => $vocabulary->label(),
'depth' => 0,
'parents' => [],
'weight' => 0,
'description' => $vocabulary->label(),
'path' => $vocabulary->toUrl('overview-form')->toString(TRUE)->getGeneratedUrl(),
];
// @todo It is possible that this will return a filtered set, if
// term_access is applied to the query.
$data = $term_storage->loadTree($vocabulary_id);
$this->tree = $this->buildTree($vocabulary_id, $data, $tree);
}
}
}
return $this->tree;
}
Isn't this hardcoded to only work with one vocabulary?
Improved MR in patch format for 2.0.1 and 2.0.4.
This patch is for 2.0.4, #4 was for 2.0.1.
Fixed patch; I'll update the MR in a moment.
Kinda breaking the concept of semantic versioning to add it to the upcoming 1.1.3 release, but I've committed this. I did have to rework the URL to the docs page, as it kept failing phpcs, hopefully it's acceptable in this format. Thank you.
damienmckenna → made their first commit to this issue’s fork.
The MR adds a setting for controlling whether the buttons are shown at the top of the node form.
Should this one be closed and another opened to fix the tests?
*shrug*
I combined both approaches.
Thanks everyone!
There are a ton of problems with how HTML fields are converted to text for tokens, I've started using the Smart Trim module because of this. FYI the summary token is also going to be removed as it's no longer relevant now that there's no default "body" field in 11.3.
Done:
https://rumble.com/v712oqk-contrib-half-hour-2018-01-04.html
https://rumble.com/v712ory-contrib-half-hour-2018-01-11.html
https://rumble.com/v712osi-contrib-half-hour-2018-01-18.html
https://rumble.com/v712oti-contrib-half-hour-2018-01-25.html
https://rumble.com/v712ou2-contrib-half-hour-2018-02-01.html
https://rumble.com/v712ouc-contrib-half-hour-2018-02-08.html
https://rumble.com/v712owq-contrib-half-hour-2018-02-15.html
https://rumble.com/v712oxu-contrib-half-hour-2018-03-01.html
https://rumble.com/v712oyg-contrib-half-hour-2018-03-08.html
https://rumble.com/v712p1k-contrib-half-hour-2018-03-15.html
https://rumble.com/v712p2i-contrib-half-hour-2018-03-22.html
https://rumble.com/v712p2y-contrib-half-hour-2018-03-29.html
https://rumble.com/v712p3g-contrib-half-hour-2018-04-05.html
https://rumble.com/v712p4g-contrib-half-hour-2018-04-19.html
https://rumble.com/v712p52-contrib-half-hour-2018-05-03.html
https://rumble.com/v712p5e-contrib-half-hour-2018-05-10.html
https://rumble.com/v712p60-contrib-half-hour-2018-05-17.html
https://rumble.com/v712p72-contrib-half-hour-2018-05-24.html
https://rumble.com/v712p7m-contrib-half-hour-2018-05-31.html
https://rumble.com/v712p9k-contrib-half-hour-2018-06-07.html
https://rumble.com/v712pag-contrib-half-hour-2018-06-14.html
https://rumble.com/v712pbe-contrib-half-hour-2018-06-21.html
https://rumble.com/v712pcc-contrib-half-hour-2018-06-28.html
https://rumble.com/v712pda-contrib-half-hour-2018-07-05.html
https://rumble.com/v712pdy-contrib-half-hour-2018-07-12.html
https://rumble.com/v712pek-contrib-half-hour-2018-07-26.html
https://rumble.com/v712pfo-contrib-half-hour-2018-08-02.html
https://rumble.com/v712pg2-contrib-half-hour-2018-08-16.html
https://rumble.com/v712pgm-contrib-half-hour-2018-08-23.html
https://rumble.com/v712phk-contrib-half-hour-2018-08-30.html
https://rumble.com/v712pi6-contrib-half-hour-2018-09-06.html
https://rumble.com/v712pp0-drupal-contrib-half-hour-2018-09-20.html
https://rumble.com/v712piy-drupal-contrib-half-hour-2018-09-13.html
https://rumble.com/v712ppu-drupal-contrib-half-hour-2018-10-04.html
https://rumble.com/v712pqq-drupal-contrib-half-hour-2018-10-11.html
https://rumble.com/v712pri-drupal-contrib-half-hour-2018-10-18.html
https://rumble.com/v712psc-drupal-contrib-half-hour-2018-10-25.html
https://rumble.com/v712pt4-drupal-contrib-half-hour-2018-11-01.html
https://rumble.com/v712ptm-drupal-contrib-half-hour-2018-11-08.html
https://rumble.com/v712puc-drupal-contrib-half-hour-2018-11-15.html
https://rumble.com/v712pv6-drupal-contrib-half-hour-2018-11-29.html
https://rumble.com/v712pvs-drupal-contrib-half-hour-2018-12-06.html
https://rumble.com/v712pwc-drupal-contrib-half-hour-2018-12-13.html
https://rumble.com/v712px2-drupal-contrib-half-hour-2018-12-20.html
I provided an MR that adds basic logic for the else_pass_through option, let me know what you think and I can refine & expand it.
damienmckenna → created an issue. See original summary → .
@fjgarlin: Is there a current plan on when you & the team will start migrating projects? Thank you.
I used convertmoduleoop, will see how the tests work.
Bumping the priority for this as it's currently nearly impossible to index an existing site.
Thank you for the MR.
The tests appear to be failing badly, it might be worth some manual testing to confirm there aren't unintended problems.
Also, I noticed that if I break the command (or the ssh connection breaks and I have to reconnect to the server) and re-run the command it starts over from the start again, rather than continuing from where it was. Is this intended?
FYI patch #8 resolves the issue, it might be worth committing it and refactoring in a separate issue.
FWIW this is working well for us on a prod site, thank you.
FYI I ran into a similar problem and realized that ProcessedText::preRenderText() was doing a whole load of work if the string was blank, which seemed unnecessary: ✨ Return early in preRenderText() if text is blank Active
Related: 🐛 PHP8.1 deprecation: str_replace(): Passing null to parameter #3 Needs work
The MR simply returns early if the string is empty.
My suggestion is that for now release this as-is with the test line commented out and then work on the one failing test separately, just to be able to move on.
The test that fails was last edited in #3255904: Remove the WebformAssertLegacyTrait → , but it actually comes from #3094270: Add support for webform variants → .
WebformVariantOverrideTest passes locally but fails on gitlabci. Hrm..
Ok, I see that now.
Is there normally such a variation in how long the tests take to run? When I added the _PHPUNIT_CONCURRENT line initially the tests ran in <20 minutes, now it's hitting twice that and timing out.
The dependency issue showed up when I ran all of the tests locally, maybe it was a dependency of another dependency so it was my fault it failed locally.
The fact that it's failing at 189 tests and not finishing them all suggests it's not running the tests concurrently per the _PHPUNIT_CONCURRENT setting - does the issue fork need to be updated?
I added a missing test dependency and changed the tests to run concurrently, that allowed them to finish in ~18 minutes instead of timing out after ~30 minutes. The tests are now green.
This is related to a security issue, so adding the "Security improvements" tag for visibility.
damienmckenna → made their first commit to this issue’s fork.
Given that in gitlab the README.md file will become the project page's description, the README.md needs to be updated too.
You never know.
On my site I was actually having problems getting v2 to run on a site with 11.1, but solved it by hardcoding composer.json to require 2.0.1. For now.