Previously absolutely positioned boxes could only have a % height if
their parent had a absolute height (a height in pixels, em, etc).
This broke some websites/demos such as the "Francine CSS oil painting",
which starts to appear after this commit.
Francine: https://diana-adrianne.com/purecss-francine/
Previously we were checking if values were "auto" after resolving the
"auto"-ness out of them, which didn't work. There's still a bunch of
work to do on this algorithm, but now we can at least resolve some basic
automatic height scenarios.
Percentage sizes of flex items are relative to the flex container, but
even when the flex container is automatically sized, we still have to
support them.
To make this work, we first do a pass where percentage sizes are ignored
(treated as "auto", basically) in order to get a "reference" value.
Then we do a second pass where percentages can be resolved against this
reference value.
We were neglecting to clamp flex items when calculating the intrinsic
main size of the flex container. This was covered by a FIXME, which we
can now remove. :^)
We had an older incomplete implementation of flex container max-content
sizing that we used for unconstrained main size determination.
This patch replaces the old implementation with a call to the new
max-content sizing code.
Note that this isn't a complete implementation of flex container sizing
still, but at least we've deduplicated some code.
The CSS-FLEXBOX-1 spec gives us two situations in which flex item cross
sizes should be considered definite. Both of them happen *during* flex
layout, which is super finicky but it is what it is.
When calculating intrinsic sizes, we don't need to recurse into *every*
box and layout its insides. IIUC, we can skip any unconstrained box with
definite sizes in both axes. So this patch does exactly that.
When we decide that a box has definite width or height based on its
containing block's corresponding size, we'll want to resolve the
current box's size as well. Otherwise anyone querying the size on
this box will get the bogus message of "yes, this definite, and its
value is zero."
Previously, we considered all LengthPercentage values for used flex
basis to be definite. This is not accurate, as percentages should only
be considered definite if the reference value they resolve against is
a definite size.
Fix this by checking the flex container's main definite size flag.
There's no need to override the sizes before calculating the cross size.
Besides, by the time we're calculating the hypothetical cross size of
flex items, we may already have established a definite main size anyway,
so overriding it would be wrong.
This state is less static than we originally assumed, and there are
special formatting context-specific rules that say certain sizes are
definite in special circumstances.
To be able to support this, we move the has-definite-size flags from
the layout node to the UsedValues struct instead.
If we wait until after the parent context has laid out the flex
container, abspos children are able to use the final results of the
parent sizing the flex container.
This makes `height:auto` work on abspos children of a flex container.
This disables this system progress bar, and instead creates one
out of pseudo elements, that can be selected and styled with the
::-webkit-progress-bar/value selectors.
Before performing intrinsic sizing layout on a box, we now check if its
containing block has automatic size in the relevant axis, and if so,
we fetch the size of the nearest containing block ancestor with a size.
`min-width: auto` and `min-height: auto` have special behavior on flex
items. We already handled it in many cases but there were two places
where it was incorrectly treated as 0. This fixes that.
This was a leftover from when ComputedValues stored sizes in Optionals.
Now that we've gotten rid of the "undefined" state, there's no need for
this helper, we can just access the size values directly.
Due to a missing `return` statement, we were creating two anonymous
wrapper blocks around each piece of inline content inside a flex
container.
This had no visual impact, since they ended up with 0x0 dimensions,
but we were wasting a fair bit of time running layout on them.
Each of these strings would previously rely on StringView's char const*
constructor overload, which would call __builtin_strlen on the string.
Since we now have operator ""sv, we can replace these with much simpler
versions. This opens the door to being able to remove
StringView(char const*).
No functional changes.
This commit moves the length calculations out to be directly on the
StringView users. This is an important step towards the goal of removing
StringView(char const*), as it moves the responsibility of calculating
the size of the string to the user of the StringView (which will prevent
naive uses causing OOB access).
