Prefixed intrinsic size value
<!-- Please describe your changes on the following line: -->
This is a PR for https://bugzilla.mozilla.org/show_bug.cgi?id=1355402
- [X] `./mach build -d` does not report any errors
- [X] `./mach test-tidy` does not report any errors
- [X] These changes fix#16788
- [X] These changes do not require tests because it's for stylo
<!-- Reviewable:start -->
---
This change is [<img src="https://reviewable.io/review_button.svg" height="34" align="absmiddle" alt="Reviewable"/>](https://reviewable.io/reviews/servo/servo/16962)
<!-- Reviewable:end -->
This enum type used to contain the result of parsing
one CSS source declaration (`name: value;`) and expanding shorthands.
Enum types are as big as the biggest of their variant (plus discriminant),
which was quite big because some shorthands
expand to many longhand properties.
This type was returned through many functions and methods,
wrapped and rewrapped in `Result` with different error types.
This presumably caused significant `memmove` traffic.
Instead, we now allocate an `ArrayVec` on the stack
and pass `&mut` references to it for various functions to push into it.
This type is also very big, but we never move it.
We still use an intermediate data structure because we sometimes decide
after shorthand expansion that a declaration is invalid after all
and that we’re gonna drop it.
Only later do we push to a `PropertyDeclarationBlock`,
with an entire `ArrayVec` or nothing.
In future work we can try to avoid a large stack-allocated array,
and instead writing directly to the heap allocation
of the `Vec` inside `PropertyDeclarationBlock`.
However this is tricky:
we need to preserve this "all or nothing" aspect
of parsing one source declaration,
and at the same time we want to make it as little error-prone as possible
for the various call sites.
`PropertyDeclarationBlock` curently does property deduplication
incrementally: as each `PropertyDeclaration` is pushed,
we check if an existing declaration of the same property exists
and if so overwrite it.
To get rid of the stack allocated array we’d need to somehow
deduplicate separately after pushing multiple `PropertyDeclaration`.
Introduce ComputeDistance trait, which implement compute_distance and
compute_squared_distance.
For vector, compute_squared_distance is necessary because we use Euclidean
distance as the distance between two values. The easier way to implement
compute_squared_distance is to square the result from compute_distance, but
for some property values, they may have many components, e.g. (v1, v2, v3).
If we just square the result from compute_distance, the computation is
(sqrt(v1^2 + v2^2 + v3^2))^2. There are two redundant operators:
"square-root" and then "square". In order to avoid this, we should
implement compute_squared_distance separately for these types.
MozReview-Commit-ID: LmmrUXYlDb6
scriptlevel is a property that affects how font-size is inherited. If scriptlevel is
+1, for example, it will inherit as the script size multiplier times
the parent font. This does not affect cases where the font-size is
explicitly set.
However, this transformation is not allowed to reduce the size below
scriptminsize. If this inheritance will reduce it to below
scriptminsize, it will be set to scriptminsize or the parent size,
whichever is smaller (the parent size could be smaller than the min size
because it was explicitly specified).
Now, within a node that has inherited a font-size which was
crossing scriptminsize once the scriptlevel was applied, a negative
scriptlevel may be used to increase the size again.
This should work, however if we have already been capped by the
scriptminsize multiple times, this can lead to a jump in the size.
For example, if we have text of the form:
huge large medium small tiny reallytiny tiny small medium huge
which is represented by progressive nesting and scriptlevel values of
+1 till the center after which the scriptlevel is -1, the "tiny"s should
be the same size, as should be the "small"s and "medium"s, etc.
However, if scriptminsize kicked it at around "medium", then
medium/tiny/reallytiny will all be the same size (the min size).
A -1 scriptlevel change after this will increase the min size by the
multiplier, making the second tiny larger than medium.
Instead, we wish for the second "tiny" to still be capped by the script
level, and when we reach the second "large", it should be the same size
as the original one.
We do this by cascading two separate font sizes. The font size (mSize)
is the actual displayed font size. The unconstrained font size
(mScriptUnconstrainedSize) is the font size in the situation where
scriptminsize never applied.
We calculate the proposed inherited font size based on scriptlevel and
the parent unconstrained size, instead of using the parent font size.
This is stored in the node's unconstrained size and will also be stored
in the font size provided that it is above the min size.
All of this only applies when inheriting. When the font size is
manually set, scriptminsize does not apply, and both the real and
unconstrained size are set to the explicit value. However, if the font
size is manually set to an em or percent unit, the unconstrained size
will be set to the value of that unit computed against the parent
unconstrained size, whereas the font size will be set computing against
the parent font size.
MozReview-Commit-ID: 820BIWqno3L
