Every time the GitLab CI-job 🧹 CSS linting (stylelint) runs, there's a warning:
Browserslist: caniuse-lite is outdated. Please run:
npx update-browserslist-db@latest
Why you should do it regularly: https://github.com/browserslist/update-db#readme
Find a MR that triggers the 🧹 CSS linting (stylelint) job and observe the warning.
Here's one I've prepared/found earlier: https://git.drupalcode.org/issue/drupal-3042127/-/jobs/3294813#L269-271
Do what's recommended in the warning, commit and move on to greater things in life.
$ npx update-browserslist-db@latest
$ yarn build:css
Now the warning should be gone.
Backport this to the gazillion current branches, TBD by the core committers.
Show commandsStart 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:
changes
Comments
Comment #3
spokjeComment #4
spokjeComment #5
smustgrave commentedJust picking 1 change
Shouldn't this be
Comment #6
spokjeTBH: I don't know, but much more important, these are all "automagically" made changes by
yarn build:cssafter I did the update ofcanisuse.So even if it is indeed better to make that change, and again, I really am not a FE-person, so I have no clue, I believe that would be something for a follow-up issue?
(And I'm not even really sure about that one either...)
Comment #7
smustgrave commentedAsked #frontend channel. My concern is the current position is only checking that child selector and not the top level dir attribute.
Comment #8
markconroy commented:dir(rtl)should work perfectly fine for us.It's a CSS function that has been supported for quite a while.
https://developer.mozilla.org/en-US/docs/Web/CSS/:dir
Comment #9
smustgrave commentedMy only question is the position of :dir() does that function only apply to that child selector or will it check that top level id dir=rtl?
Comment #10
markconroy commentedFrom MDN
Looks like it's document-level.
It seems
[dir="rtl"]is only be valid if that is part of the HTML of the page, but:dif(rtl)can be inferred from other things like user-agents, for example.https://drafts.csswg.org/selectors/#the-dir-pseudo
Comment #11
smustgrave commentedin that case I have no objections.
Comment #12
spokjeComment #13
spokjeMirroring status of parent issue, which became critical.
Comment #16
nod_Committed f39f923 and pushed to 11.x. Thanks!
Comment #17
spokjeThanks @nod_, do we want to backport this down to other branches?
Comment #19
nod_yes makes sense to backport to 10.4
Comment #20
nod_Comment #21
spokjePersonally I think we might wanna backport this to 11.1.x, 10.5.x and 10.4.x
(Oh, the joy off multiple active main branches...)
I'll put MRs up for all three of them and let y'all decide.
Comment #24
spokjeSo it already landed in
11.1.x, so making an MR for10.5.xand am pretty hopeful this can be backported to10.4.xeasily.Comment #27
nod_nice thanks
Comment #28
spokjeWell, the bump in 10.5.x has surprisingly little, as in none at all, effect on the CSS files in there.
Comment #29
spokjeOops
Comment #33
nod_