Closed (fixed)
Project:
Drupal core
Version:
8.8.x-dev
Component:
phpunit
Priority:
Normal
Category:
Bug report
Assigned:
Unassigned
Issue tags:
Reporter:
Created:
27 May 2020 at 18:46 UTC
Updated:
12 Jun 2020 at 21:09 UTC
Jump to comment: Most recent, Most recent file
Comments
Comment #2
mondrakeComment #3
mohrerao commentedOn it
Comment #4
mohrerao commentedAdding a patch
Comment #5
longwaveI think it might be safe to replace
$this->config('system.site')->get('name')with the hardcoded string "Drupal"?Comment #7
mohrerao commented@longwave, That makes sense. But if someone wants to test this on local with a different site name, then it would fail. I see tests using both hardcoded text 'Drupal' as well as $this->config('system.site')->get('name'). I have now replaced it to Drupal.
Comment #8
mondrakeThe changes to assertText calls are out of scope here.
This is ending up assigning $title_source twice, and not using it also.
No, it won't, because for every functional test a default Drupal installation is prepared and the test run on that, to avoid that the local site info leak in the test.
Comment #9
mohrerao commentedThanks for the comments @mondrake.
Thanks for the info. Did not know that Drupal uses a fresh installation for test runs. That answers lot of time taken for running tests!!
Comment #10
mondrakeThanks @mohrerao, this looks good, just there's a stray use statement left causing CS check failure.
Nit:
This could be just
$this->assertTitle('0 | Drupal');, probably.Comment #11
mondrakeComment #12
mohrerao commentedThanks for that. Updated.
Comment #13
mondrakeThanks. This is outstanding:
Comment #14
mohrerao commentedAah missed it. Updating patch.
Comment #16
mondrakeThanks! This patch is cleaning up Simpletest cruft from calls to AssertLegacyTrait::assertTitle and WebAssert::titleEquals - ensuring the first argument is always a string and no further arguments are passed in.
Prep for #3139412: Replace usages of deprecated AssertLegacyTrait::assertTitle().
Comment #17
xjmComment #21
xjmSo much cleaner, and also actually better test coverage in a couple places. Thanks for the fast turnaround!
The test author was feeling a bit existentialist, apparently.
(* In case someone needs to say something about how Samuel Beckett was not actually an existentialist, OK, fine.)
I've committed this to 9.1.x and cherry-picked it to 9.0.x and 8.9.x. It does not cherry-pick cleanly to 8.8.x, so if it's straightforward, we should create an 8.8.x version. (8.8.x will go into security-only mode starting with the code freeze on June 2, FYI, but we may be able to get this and the issue it's blocking backported by then.) Thanks!
Comment #22
bunty badgujar commentedComment #23
bunty badgujar commentedbackport to 8.8.x version.
Comment #24
mohrerao commented3143339-23.patch looks fine. Moving to RTBC again for commiting against 8.8.x
Comment #25
bunty badgujar commentedComment #27
xjmCommitted the backport to 8.8.x. Thanks!