users: Add screen reader accessible column

[Venefilyn]

Current users page lists the users in a table together with a column for actions. However, the table header doesn't include the last actions column and instead shows the table header line ending preemptively.

As we likely don't want a visibly named column for the actions this will instead visually just extend the line so it is consistent while also adding a screen reader friendly text to indicate what the column is portraying.

This does mean we need to include
@patternfly/patternfly/patternfly-addons.css in the PatternFly CSS import.

---

This line seems only to occur on Chrome and not Firefox, but definitely a bug

PR applied:

[martinpitt]

Fun! Our current pixel test reference is correct, the separator lines properly go to the end. Triggering the f-41 tests to see if it makes any difference.

I cannot reproduce the bug with Chromium 132.0.6834 (current F41 version). Nevertheless, this looks harmless/ok, so I'm fine with applying it. However, I have a question:

We don't use the pf-v5-u-screen-reader anywhere in our projects so far, so that's new to me. Documentation.

There are only two other similar usages:

pkg/packagekit/updates.jsx: {_("Danger alert:")}
pkg/static/login.html:

The latter seems to be an alias according to the documentation. @Venefilyn do you know the difference between the variant with and without -u? And I think the first one is just entirely outdated and ought to be removed?

Thanks!

[martinpitt]

@Venefilyn Ugh, this breaks mobile screen size quite badly -- it seems to allocate a lot of extra space for the new column and then keep it empty.

This failure is also curious -- at first sight it's totally unrelated, but then again the font change doesn't feel like a flake. I've never seen that kind of flake before.

So this is a typical Yak shave 😢

[Venefilyn]

@Venefilyn Ugh, this breaks mobile screen size quite badly -- it seems to allocate a lot of extra space for the new column and then keep it empty.

I'll take to make sure mobile view works as expected, thanks!

This failure is also curious -- at first sight it's totally unrelated, but then again the font change doesn't feel like a flake. I've never seen that kind of flake before.

It could be related either way since this PR does include a CSS import. It shouldn't do anything given they are helper CSS stuff but the helper classes might be used here and there without them doing anything atm

[Venefilyn]

@martinpitt the failing pixel test look fine now to me if I read them correctly

[martinpitt]

@Venefilyn Agreed, this looks reasonable. I need to run, please ask @jelly or someone else in the channel to push the pixel test. Otherwise I'll do it this evening. 👍 thanks!

[martinpitt]

I pushed the pixel test (you need repo permissions for that -- it will come soon enough 😁 ). Thanks!