Daniel Holbert [:dholbert] Reporter
	Description • 8 years ago

Right now, for content-sized flex items (with default values "flex-basis:auto; width:auto"), we give them a flex base size of "fit-content", basically. I think this is what the spec used to require.

But now, we should use "max-content", per this chunk of current spec:
 # Otherwise, size the item into the available space
 # using its used flex basis in place of its main size,
 # treating a value of 'content' as 'max-content'

('content' there is a value we don't explicitly support, but it's the value that corresponds to the "flex-basis:auto; width:auto" behavior)
 https://drafts.csswg.org/css-flexbox/#hypothetical-main-size


See also this Chrome bug, where Chrome's behavior is actually correct and we're wrong, I believe:
 https://bugs.chromium.org/p/chromium/issues/detail?id=623504
(Edge matches Chrome on this, too.)

	yisibl
	Comment 1 • 8 years ago

Opera 12.16 and Safari 9 matches Chrome on this, too.

	Christian :Biesinger (don't email me, ping me on IRC)
	Comment 2 • 8 years ago

Daniel, do you think http://codepen.io/anon/pen/KbEzC is the same issue? The four flex items should not have the same size because flex-basis: auto and a different max-content size should lead to different flexing being applied.

(from https://bugs.chromium.org/p/chromium/issues/detail?id=414197)

	Daniel Holbert [:dholbert] Reporter
	Comment 3 • 8 years ago

Yes.

	Daniel Holbert [:dholbert] Reporter
	Comment 4 • 6 years ago

(I think this is a prereq for bug 1377253 -- to be able to accurately cache/reuse content measurements across multiple reflows, as I'd like to do there, it will be important that these measurements aren't influenced by the size of the container.)

	Daniel Holbert [:dholbert] Reporter
	Comment 5 • 6 years ago

Actually I think this is a dupe of bug 1374540 (whose fix landed a few days ago) -- and the attached testcase and comment 2 each give EXPECTED RESULTS in Nightly.

(Comment 0's spec-text-quote is the same as bug 1374540 comment 3.)

--> duping and transferring blocking relationship

	Will Bamberg [:wbamberg]
	Comment 6 • 6 years ago

The note reviewed here: https://github.com/mdn/browser-compat-data/pull/2151 is now live on MDN: https://developer.mozilla.org/en-US/docs/Web/CSS/flex, so I'm marking this dev-doc-complete.

	Chris Mills [:cmills]
	Comment 7 • 6 years ago

(In reply to Will Bamberg [:wbamberg] from comment #6)
> The note reviewed here: https://github.com/mdn/browser-compat-data/pull/2151
> is now live on MDN: https://developer.mozilla.org/en-US/docs/Web/CSS/flex,
> so I'm marking this dev-doc-complete.

Rel note added: https://developer.mozilla.org/en-US/Firefox/Releases/61#CSS