VOOZH about

URL: https://phabricator.services.mozilla.com/p/dao/

⇱ ♟ dao


dao (Dão Gottwald)
User

User Details

User Since
Jul 13 2018, 5:18 PM (413 w, 5 d)
Availability
Available
Review Queue
6

Recent Activity

Today

Yesterday

Mon, Jun 15

Sun, Jun 14

I tried removing , and it seemed to work. We redid the some related CSS recently, so maybe it's not necessary anymore. I don't want to complicate this patch with that though.

Sat, Jun 13

Fri, Jun 12

Thu, Jun 11

Wed, Jun 10

in child processes could be a that tunnels function calls through the actor to the parent controller. There could be a whitelist of all parent controller functions that support being called over IPC (they have to be async and have serializable parameters and return value). In the parent process, would still be a direct reference to the parent controller.
That way, instances in child processes would have access to the IPC enabled subset of the parent controller interface while instances in the parent process can easily access the whole parent controller. We also wouldn't have to create actors for moz-urlbars in the parent process.

In TypeScript, you could do something like

parentController: UrlbarParentController | UrlbarParentControllerIPCSubset

// If this returns true, TS knows that the full parent controller is available
inParent(): this is this & { parentController: UrlbarParentController } {
 return ... 
}

So we could statically check that no parent-only methods are used in a child process. We would probably also have to do the same thing but in reverse for the parent controller.

What do you think about this? Did you already have another idea how the actor communication between child and parent could look like?

The suggestion doesn't really affect the work in this patch, except that the patch creates actors even in the parent process (easy to change later). With the comments addressed, this patch looks good to me.

Tue, Jun 9

Let's get additional review from Moritz on this and then we can decide if we're comfortable enough or want an additional review from Marco or Mark.

Mon, Jun 8

Marking @mstriemer as blocking for recomp while I'm gonna ask a questions on behalf of desktop-theme-reviewers, and requesting changes on that for visibility:

dao retitled D304697: Bug 2044950 - Split UrlbarController into UrlbarParentController and UrlbarChildController, and add UrlbarParent and UrlbarChild actors. from WIP: Bug 2044950 - Split UrlbarController into UrlbarParentController and UrlbarChildController, and add UrlbarParent and UrlbarChild actors. to Bug 2044950 - Split UrlbarController into UrlbarParentController and UrlbarChildController, and add UrlbarParent and UrlbarChild actors..

Thu, Jun 4

Tue, Jun 2

Mon, Jun 1

Sat, May 30

Fri, May 29

Thu, May 28

add-circle-fill and subtract-circle-fill

The Proton versions of these icons draw a and then a on top. The Nova versions in acorn-icons draw a single describing the filled area around the + or -. That would mean the Nova version could only be used when the icon is supposed to have a transparent + or -.

Since there are existing uses for explicitly coloring the + or - (e.g. the application menu zoom controls), I manually updated the Nova version of this icons to draw a and then remove the circle drawing parts of the . This means the Nova versions in this patch diverge from the acorn-icons repo at this time. I'm talking with the icon designers about how they want to proceed, but I think in practical terms, we need these icons to work the same way as Proton.

Wed, May 27

And I think it would be fine to give them all hardcoded sizes according to the spec rather than basing their heights on the height of the urlbar.

Tue, May 26

Fri, May 22

Wed, May 20

May 18 2026

May 17 2026

May 16 2026

May 14 2026

May 13 2026

I haven't yet heard back from @Juliana or @jules. I think I still want to veto this because it feels like a semantically confusing hack on top of a suboptimal default button style.