Initial rendering support for declarative skeletons [chromium/src : main]

0 views

Skip to first unread message

Rune Lillesveen (Gerrit)

unread,

Apr 17, 2026, 2:28:15 AM (yesterday) Apr 17

to Rune Lillesveen, Vladimir Levin, Noam Rosenthal, android-bu...@system.gserviceaccount.com, Menard, Alexis, chromium...@chromium.org, devtools...@chromium.org, apavlo...@chromium.org, blink-re...@chromium.org, blink-re...@chromium.org, blink-re...@chromium.org, blink-rev...@chromium.org, blink-rev...@chromium.org, blink-...@chromium.org, devtools-re...@chromium.org, jmedle...@chromium.org, kinuko...@chromium.org

Attention needed from Noam Rosenthal and Vladimir Levin

Rune Lillesveen added 1 comment

Patchset-level comments

File-level comment, Patchset 6 (Latest):

Rune Lillesveen . resolved

This is the start of rendering a skeleton in a ::skeleton pseudo tree that we've talked about.

There are a lot of details to figure out. Now, the ::skeleton is rendered under the LayoutView at the same time as the rest of the document.

I move a skeleton document fully using an <html> root element under the ::skeleton pseudo. The question is if the author creating a skeleton should expect that an authored skeleton document should render exactly as if it was rendered standalone. If so, we need to make the ::skeleton <html> root behave as a document root with selector matching, propagation of overflow to viewport, etc.

Open in Gerrit

Related details

Attention is currently required from:

Noam Rosenthal
Vladimir Levin

Submit Requirements:

Code-Coverage
Code-Owners
Code-Review
Review-Enforcement

Inspect html for hidden footers to help with email filtering. To unsubscribe visit settings.

Gerrit

satisfied_requirement

unsatisfied_requirement

open

diffy

Vladimir Levin (Gerrit)

unread,

Apr 17, 2026, 9:58:18 PM (23 hours ago) Apr 17

to Rune Lillesveen, Noam Rosenthal, android-bu...@system.gserviceaccount.com, Menard, Alexis, chromium...@chromium.org, devtools...@chromium.org, apavlo...@chromium.org, blink-re...@chromium.org, blink-re...@chromium.org, blink-re...@chromium.org, blink-rev...@chromium.org, blink-rev...@chromium.org, blink-...@chromium.org, devtools-re...@chromium.org, jmedle...@chromium.org, kinuko...@chromium.org

Attention needed from Noam Rosenthal and Rune Lillesveen

Vladimir Levin added 3 comments

Commit Message

Line 21, Patchset 6 (Latest):

Vladimir Levin . unresolved

I'd file a bug to track implementation

File third_party/blink/renderer/core/skeleton/build.gni

Line 6, Patchset 6 (Latest): "skeleton.cc",

Vladimir Levin . unresolved

Should we have a couple of smoke tests?

File third_party/blink/renderer/core/skeleton/skeleton_loader.h

Line 62, Patchset 6 (Latest): Member<Skeleton> skeleton_;

Vladimir Levin . unresolved

The fact that there's a single skeleton but a HashSet of URLs surprises me. I can see that the implementation is to basically just fetch the skeleton as soon as we have a navigation (which I assume just races the two fetches). I would've pictured this more like a prefetch eventually where we have a URL->skeleton type of map that is opportunistically populated

Open in Gerrit

Related details

Attention is currently required from:

Noam Rosenthal
Rune Lillesveen

Submit Requirements:

Code-Coverage
Code-Owners
Code-Review

No-Unresolved-Comments

Review-Enforcement

Inspect html for hidden footers to help with email filtering. To unsubscribe visit settings.

Gerrit

satisfied_requirement

unsatisfied_requirement

open

diffy

Reply all

Reply to author

Forward

0 new messages