1546739 - {inc} Race condition with `min-width:min-content` on `<img>` elements

Status: RESOLVED FIXED

(Core :: Layout, defect, P3)

Description

[Daniel Holbert [:dholbert]]

I was just investigating some testcases for Bug 1544719 and I discovered that we don't seem to honor width:min-content reliably -- sometimes it takes a forced incremental layout for it to take effect.

Testcase coming up.

STR:

1. Load testcase.

EXPECTED RESULTS:
Image should be 800px wide (plus 2px border)

ACTUAL RESULTS:
Image is 50px wide.

If I take additional action to trigger an incremental reflow (e.g. resizing browser horizontally, or opening devtools with F12), then the rendering switches to EXPECTED RESULTS.

Interestingly, Chrome seems to have this bug as well. I can force them to switch from "actual" to "expected" results by toggling one of the img's sizing properties off and then on again, in devtools.

Comment 1

[Daniel Holbert [:dholbert]]

Attachment: (image for use in testcase)

Comment 2

[Daniel Holbert [:dholbert]]

Attachment: testcase 1

Comment 3

[Daniel Holbert [:dholbert]]

Attachment: screencast of bug

Comment 4

[Daniel Holbert [:dholbert]]

Regression range:
https://hg.mozilla.org/integration/autoland/pushloghtml?fromchange=94682e23ba118dcfab3843a3280618b2e1686ecb&tochange=fdc7b66d76a4ded62d0292bef4de0ffb090d9587

fdc7b66d76a4 Emilio Cobos Álvarez — Bug 1395964 - Don't clear the previously painted image on intrinsic size changes. r=tnikkel
3f1e9dae2467 Emilio Cobos Álvarez — Bug 1395964 - Make image loading changes not reframe. r=tnikkel

--> regression from bug 1395964

Note: sometimes I don't see the broken behavior on first load, but if I do a "light" reload (with a simple Ctrl+R) a few times, then I inevitably see the bug at least once, within the first ~5 pageloads. (And at least 3-4 times within the first ~10 reloads)

Comment 5

[Emilio Cobos Álvarez (:emilio)]

Attachment: Test-case that's broken at least as far as 63

Was working just by chance, but sure, can fix.

Comment 6

[Daniel Holbert [:dholbert]]

Thanks!

Comment 7

[Emilio Cobos Álvarez (:emilio)]

Attachment: Bug 1546739 - Fix detection of intrinsic size dependence of images. r=dholbert

Fixed width and height is not a strong enough condition.

min/max-width with intrinsic size keywords makes the final size of the image
also depend on the intrinsic size. Don't optimize away reflows when the
intrinsic size changes if they're used.

Comment 8

[Pulsebot]

Pushed by [e-mail]:
https://hg.mozilla.org/integration/autoland/rev/d128c93735ff
Fix detection of intrinsic size dependence of images. r=dholbert

Comment 9

[Web Platform Test Sync Bot (Matrix: #interop:mozilla.org)]

Created web-platform-tests PR https://github.com/web-platform-tests/wpt/pull/16530 for changes under testing/web-platform/tests

Comment 10

[Bogdan Tara[:bogdan_tara | bogdant]]

https://hg.mozilla.org/mozilla-central/rev/d128c93735ff