In D25709#19636, @valerio.bozzolan wrote:

OK.

@avivey what do you think about a new silly commodity method PhabricatorTransactions::findOneByType(array, string) ?

(BTW, in D25698 we started deprecating remarkup.diviner for RemarkupReferenceController)

what will this do?

I'm assuming you're talking about dropping a file in the text area when editing the page, right? Not just dropping it when viewing a wiki page?

Presumably resolved by D25134.

maybe just add if ($value === null) { continue; } somewhere before the this?

grep for PHUIDiffTableOfContentsItemView in the codebase to see where we might be using this class.

If we can't find the problem, that probably means we have some dead code here that should be removed.

Is there something that's not working here?

:shrug:

(re-opening, there's more work left here)

Do we still need the auto-complete feature? In modern times, the OS has the "insert emoji" feature built-in (at least, Windows and Android and KDE do).

Ok, this looks much better then what we have today, so yay.

Looks good! I want to play with it locally before accepting.

It also collected lots of commits, since May 2023.

In D25697#19176, @20after4 wrote:

I obviously failed to run ./bin/celerity map one more time before running arc diff.

There should really be some way to detect and prevent that mistake. It's one I've made probably 100 times before.

What's the current behavior for fallback from cs_CZ? Would it make sense to set the default to en_US in the base class?

What if $cdn has some evil characters? Do we have some escaping, or can it cause anything bad, or do we not care? Looks like we're including a bunch of other config as-is here, so maybe it's ok.

Looks good to me.

I still don't like the solution of using place-holder text for that; But this implementation also adds another DB access which I think is redundant.

Instead of this diff, can you add a modern endpoint that uses the EditEngine/transactions? e.g. maniphest.edit, badge.edit, etc.
That will need a different name (phriction.document.edit?), but will allow full editing capability.

So PHP, very wow

As T15224 suggest, this menu should have the same content as the side-menu on non-mobile view; And in most places, the framework handles that, by actually moving that menu.
So the "right" fix is to find what creates the current side-menu and make sure we call that, or figuring out why they're different here and normally the same.

There may be some room for extra clarity for users in complex organization setups, but this is a rare enough operation (outside of dev) that it's probably not worth worrying about.

Please add some negative tests - things that could trigger this thing, but shouldn't.

I think the EditEngine is what's used to create the actions form, and it has some sense of the object's status (see for example the available actions on Revisions - these change based on the revision's state).
Maybe it can get an additional "field" for this warning, and display it based on task status.

Very cool!

Don't we have something like phutil_is_integer method somewhere?

This is spelling, not formatting.

Using a password in 2024 is a bad advice in general.

need to also test the result of renderPhabricatorLogo() when no logo set, when a custom logo is set, and when a custom logo is set but hidden (maybe. Maybe the logo file should be Attached to something public, so this situation can't really happen?)

That "don't sign a CLA" article should be titled "don't sign away your copyrights", to be more accurate and less click-bait-y. A CLA may ask you to sign away copyrights (which I feel makes the project non-open-source) ; It doesn't have to.

(It's about spam, but it doesn't have anything explicitly sensitive. I've made it public now)

On a more meta level, Maniphest isn't well-suited to be an entry-form to be filled by a non-expert user; Nuance is/was intended for this use-case.

on the technical level, Herald can't block object creation - it runs after the fact, by the Daemons.

I think the 404 is possibly better then "silently fail without error".

How can we end up with null here?

Rebase

In T15096#12185, @valerio.bozzolan wrote:

Another problem dramatically frequent for newcomers (at least in my office).

It's relatively too much easy to start with a similar commit message:
...

(run the full tests, and then count it as my + O1 Accept).

I think the costs of the extra translations are low, but the cost of "the tests are not being translated for developers" are even lower.

I'd be worried that this might break some tests that do relate to translation/extraction, but running arc unit --everything should solve that.

This is probably the right fix (or maybe moving shouldAllowPublic() to the base PhabricatorFactController).

You can set "visibility" for each URI already.

Yeah, thad makes sense, I guess.