Problem/Motivation
Not yet using Gitlab CI.
Steps to reproduce
Proposed resolution
Start using Gitlab CI.
Remaining tasks
User interface changes
API changes
Data model changes
Issue fork upgrade_status-3403194
Show commands
Start within a Git clone of the project using the version control instructions.
Or, if you do not have SSH keys set up on git.drupalcode.org:
Comments
Comment #3
gábor hojtsyThe single Drupal 10 fail is due to placement of the module:
While after merging that the test will work, it will not work for future MRs either, so we need to fix the assertion to be more MR-proof.
Comment #4
gábor hojtsyThe 9 pipeline had similar fails, but uncovers one more fail point:
Comment #6
marvil07 commented@Gábor Hojtsy, it took a few iterations to get it right, but I think the current code in the merge request is now ready for inclusion.
At
tests/modules/upgrade_status_test_error/fatal.php, I added a couple of comments, one for php linting, and the other for coding standards.The linting one is the important, that avoids a lint fail, which is expected in that case.
To avoid it I declared that it should be only linted on versions lower than php 7, and given composer requires higher than php 7 indirectly via dependencies, it basically will not happen ever.
That last bit needed some changes on
UpgradeStatusAnalyzeTest, since a few messages there contain the line number of the parse error.Finally, the directory assumptions on test, as you mentioned on #3 and #4, needed a few changes.
To solve it, I am relying on the extension object related to the module where the tested template lives.
There are multiple suggestions from phpcs, phpstan, and stylint, but they are all optional, so I have not really dive into them.
Marking now as NR.
Comment #8
gábor hojtsyLanded in 4.x, thanks for the fantastic work on this!
Now trying to backport to 3.x. I think the getPath() may need to be rewritten as drupal_get_path(), but let's make the tests tell us that :D
Comment #11
gábor hojtsyI checked and the extension class based path methods are already used, so that should be fine. Committed to 3.x. Thanks!