This patch is not currently working. The problem is that the install hook simply updates the database, but it doesn't update the schema definition. I have an update update hook that works for my use case, but it's not the correct solution, I'm sure. So we're going to need to correctly handle this.
Here's my custom update hook that creates a backup table, drops the old redirect source column, add the new redirect source column, and then update the new column with the backup table.
/**
* Add source fragment column.
*/
function redirect_update_8110() {
$database = \Drupal::database();
$database->query('CREATE TABLE {redirect_bak} SELECT * FROM {redirect}');
$field_storage_definition = \Drupal::entityDefinitionUpdateManager()->getFieldStorageDefinition('redirect_source', 'redirect');
\Drupal::entityDefinitionUpdateManager()->uninstallFieldStorageDefinition($field_storage_definition);
$field_storage_definition = \Drupal\Core\Field\BaseFieldDefinition::create('redirect_source')
->setLabel(t('From'))
->setDescription(t("Enter an internal Drupal path or path alias to redirect (e.g. %example1 or %example2). Fragment anchors (e.g. %anchor) are <strong>not</strong> allowed.", ['%example1' => 'node/123', '%example2' => 'taxonomy/term/123', '%anchor' => '#anchor']))
->setRequired(TRUE)
->setTranslatable(FALSE)
->setDisplayOptions('form', [
'type' => 'redirect_link',
'weight' => -5,
])
->setDisplayConfigurable('form', TRUE);
\Drupal::entityDefinitionUpdateManager()->installFieldStorageDefinition('redirect_source', 'redirect', 'redirect', $field_storage_definition);
$database->query('UPDATE {redirect} r JOIN {redirect_bak} rb ON r.rid = rb.rid SET r.redirect_source__path = rb.redirect_source__path, r.redirect_source__query = rb.redirect_source__query');
$database->query('DROP TABLE {redirect_bak}');
}
j_ten_man → created an issue.
Here's how I got this installed via composer on Drupal 10 (10.2.5):
Prerequisites: require cweagans/composer-patches
In your composer.json file, add the following to your "repositories" (before the drupal.org repository):
{
"type": "git",
"url": "https://git.drupalcode.org/issue/bibcite_crossref-3388342.git"
},
{
"type": "git",
"url": "https://github.com/alvar0hurtad0/cache.git"
},
{
"type": "git",
"url": "https://github.com/dmlb2000/simple-cache-bridge.git"
}
Add the following to the "require" section:
"cache/cache": "dev-patch-1 as 1.2",
"cache/simple-cache-bridge": "dev-master as 1.2.0",
"drupal/bibcite_crossref": "dev-3388342-drupal-10-version"
Add the following patches:
"drupal/bibcite_crossref": {
"#3388342-25: Fix json_encode call": "https://www.drupal.org/files/issues/2024-04-26/3388342-25.patch"
},
"renanbr/crossref-client": {
"#3388342-24: Fix bad cache check": "https://patch-diff.githubusercontent.com/raw/renanbr/crossref-client/pull/24.patch"
}
Hopefully the upstream changes are merged soon, but this let me install everything via composer.
j_ten_man → made their first commit to this issue’s fork.
Just ran into this issue on a views search page. The user.current_user_context service was getting the current user set from the search API. This would then load the user (doing a full entityTypeManager load of the user) to supply the user account. Blocks were then no longer rendering on the page that were limited to specific roles since the current_user was user 0. Using -2 as the uid didn't help either. I've created a merge request which uses the uid of the current user. Not sure if there are other negative consequences from doing this, but the site that we're using this on shouldn't have any negative consequences from doing this.
A more elegant solution to this would be to reset the current_user and user.current_user_context services after everything is rendered, but wasn't sure how to accomplish that.
j_ten_man → made their first commit to this issue’s fork.
j_ten_man → created an issue.
j_ten_man → created an issue.
j_ten_man → created an issue.
j_ten_man → created an issue.
j_ten_man → made their first commit to this issue’s fork.
Make a couple of changes to the patch:
- Made parameters before the config factory non-nullable in order to match parent class and avoid warnings from php8.
- Changed ConfigFactory to ConfigFactoryInterface
j_ten_man → made their first commit to this issue’s fork.
Potentially that's a better solution - it allows other event subscribers to correctly handle the text that way. I wasn't sure of other unintended consequences of doing it that way as it would potentially be an API change, and this solution was good enough for my needs.
Patch attached.
The issue stems from the fact that the ChannelPostRender event casts the $output variable to a string, so we lose the Markup class on it.
j_ten_man → created an issue.
Here's the corrected patch with the interdiff.
Access granted
And here is a lightly tested patch. Did some code re-organization to re-use the same logic for calculating the dimensions when the img tag is generated and when the image is actually resized.
This patch is experiencing the issue that was pointed out in #17. I'll work on a fix for it, but the patch on 21 and 23 don't work because we need to implement the transformDimensions method and correctly handle when we're not scaling the image.
Changing all of the >= checks to >. This was causing issues for me where if the image is already scaled to the correct width or height, then it was still recalculating a new value. The comments in the patch seem to indicate that it should actually be just > and not >=.
j_ten_man → made their first commit to this issue’s fork.
j_ten_man → made their first commit to this issue’s fork.
j_ten_man → created an issue.
j_ten_man → made their first commit to this issue’s fork.
j_ten_man → made their first commit to this issue’s fork.
j_ten_man → created an issue.
j_ten_man → made their first commit to this issue’s fork.
Re-roll of #20 for latest 10.1.x updates.
Went with a different solution as the patch here wasn't working for me. Forced the order to be set on the line item while collecting them.
j_ten_man → created an issue.
This patch needs work, but don't want to spend a bunch of time on it if others think that we should go a different route (for example, using an Event vs alter hook).
An example implementation using this hook:
<?php
/**
* Implements hook_commerce_stripe_payment_intent_create_alter().
*/
function hook_commerce_stripe_payment_intent_create_alter(&$intent_array, $order, $payment = NULL) {
// Allow a free order to process.
if (!$intent_array['amount']) {
// Force manual capture - a free order shouldn't automatically be charged.
$intent_array['capture_method'] = 'manual';
// Get the price before discounts as the authorization amount.
$intent_array['amount'] = \Drupal::service('commerce_price.minor_units_converter')->toMinorUnits($order->getSubtotalPrice());
// $.50 is the minimum amount required by Stripe.
if (!$intent_array['amount']) {
$intent_array['amount'] = 50;
}
}
}
Rerolled against latest dev.
j_ten_man → made their first commit to this issue’s fork.
j_ten_man → made their first commit to this issue’s fork.
j_ten_man → created an issue.