I think for the guidance on the message fields we want something like "The pattern should begin with 'What's wrong". Followed by "How to make it right"
example: The alt text contains the filename. Please reword your alt to not include the filename."
Here is the updated look with the addition of the list of violations as well as the accompanying filter.
Crediting maggiewachs for the idea.
@mgifford I am not sure what to surface for date. I struggle with including it because it may largely misrepresent what it is.
- Created date - Using the created date would indicate when the node, term or media entity was created, but the image in its current state might not have been there. So in terms of error, the created date would err only on the side of representing a date that was too early.
- Updated date - would indicate when the entity was last updated, but that does not mean when the image field was last updated. Using the updated date would err on the side of being too recent a date.
- revision where the image first appeared - This would tell when it was added, but would never change. The revision processing that is required to determine when an image first appeared would be intensive and likely cause the refresh of this report to take too long to process.
- revision where the image last changed - This is perhaps the most useful information, but would be the most intensive to search the revision histories. for each entity.
Released as part of 1.0.8 →
I think this report was from an older version. The current version is showing this correctly
This is a great find @skyriter. We definitely did not adequately defend against old processes rusting.
I think this can be solved by wrapping the call to `->gatherItemsToProcess()` in a try - catch and then surface the error notice to the column if it catches one. The UI calls that function to see how many items is has to show in the display for each BatchOperation.
We will only be storing the violations, not the warnings. This is to reduce the noise.
Crediting MaggieWachs for the idea.
Note: The entity id link on the ids will take you to the entity where the image is referenced.
swirt → created an issue. See original summary → .
Since the auditor to populate the audit table has not been added yet, in order to see the View populated with sample data you will have to run: `drush atvfat` This is a temporary Drush command that populated the audit table.
This View is operational now. Of course the Auditor that builds the entries is not running yet, but the View works with test content.
Current View from PR has an issue
SQLSTATE[42000]: Syntax error or access violation: 1064 You have an error in your SQL syntax; check the manual that corresponds to your MariaDB server version for the right syntax to use near 'FROM "alt_text_validation_audit" "alt_text_validation_audit" LIMIT 11 OFFSET 0' at line 3: SELECT FROM "alt_text_validation_audit" "alt_text_validation_audit" LIMIT 11 OFFSET 0; Array ( )
For some reason the db table name is repeating.
I've re-thought this a bit. Rather than using a custom config entity which would need entity references and a lot of extra overhead, this should just be a custom field (using the field API) to create a field that is made of of subfields. The field instance on the form will have unlimited cardinality. That solves the editorial experience nightmare that we would have with a config entity through entity references.
Here is a great video on the process making of custom fields
write, truncate and generate test data are working.
Oddly enough, I can not get the drush command to register.
swirt → created an issue.
As the process runs, it can add entries to ✨ Create DB table for storing audit details Active
The approach I am going to take for reporting is to put all alts found into a table then generate the report once per day on cron (maybe later with a drush command). This way I can store them and denote whether they pass validation or not. Initially we will start with just the image fields, but then later add processing to include the long html fields too (including body).