Many APIs have been renamed to replace points
or Points
with length
or Length
.
This new name better describes one-dimentional measure of space in some unspecified unit
which is often unrelated to the PostScript point or the CSS pt
unit.
This also removes a misleading similarity with the 2D Point
,
whose components can have any unit and are not even necessarily absolute lengths.
Example usage change:
use taffy::prelude::*;
// …
let header_node = taffy
.new_leaf(
Style {
- size: Size { width: points(800.0), height: points(100.0) },
+ size: Size { width: length(800.0), height: length(100.0) },
..Default::default()
},
).unwrap();
layout_flexbox()
has been removed from the prelude. UseFlexboxAlgorithm::perform_layout()
instead.- The following methods have been removed from the
LayoutTree
trait:parent
,is_childless
,layout
,measure_node
,needs_measure
,cache_mut
andmark_dirty
. These no longer need to be implemented in custom implementation ofLayoutTree
.
- The Flexbox algorithm has now been moved behind the
flexbox
feature. Theflexbox
feature is enabled by default. - The
justify_self
property has been moved behind thegrid
feature. taffy::node::Node
has been replaced withtaffy::tree::NodeId
. This should make it much easier to implement theLayoutTree
trait as the underlying type backing the node id now au64
rather than aslotmap::DefaultKey
.- Module organisation changes:
- The
math
module has been made private - The
axis
module has been merged into thegeometry
module - The debug module is no longer public. The
print_tree
function is now accesible underutil
. - All types from the
node
,data
,layout
,error
andcache
modules have been moved to the thetree
module.
- The
- Fix exponential blowup when laying out trees containing nodes with min and max sizes.
- Fix sizing of children when the available_space < min_size (#407)
- Fix caching bug where a cached result would sometimes be incorrectly used when the amount of available space increased (bevyengine/bevy#8111) and (bevyengine/bevy#8124)
- Fix incorrect min-content size for
flex-wrap: wrap
nodes (bevyengine/bevy#8082)
- Fix: Make
padding
andborder
floor node sizes (#372) - Fix: Prevent percentages contributing to min-content sizes (#388) (also fixes bevyengine/bevy#8017)
- Fix: Ignore
align_content
whenflex_wrap
is set tonowrap
(#383)
- Fix
display: none
when it is set on a flexbox child (#380) - Fix
display: none
when it is set on a grid child (#381)
- Fix
display: none
when it is set for the only node in the hierarchy (#377)
- Added
enable_rounding
anddisable_rounding
methods to theTaffy
struct which enable consumers of Taffy to obtain unroundedf32
values for the computed layouts if they want them. Rounding remains enabled by default.
- Fixed rounding algorithm such that it never leaves gaps between adjacent nodes (#369)
- Fixed compiling with the
grid
feature disabled (#370) - Fixed compiling with the
std
feature disabled
- Allow partial nested values to be deserialized into a
Style
using theserde
feature.
- The
serde
feature now works when thegrid
feature is enabled
See below for details of breaking changes.
We very excited to report that we now have support for CSS Grid layout. This is in addition to the existing Flexbox layout support, and the two modes interoperate. You can set a node to use Grid layout by setting the display
property to Display::Grid
.
Taffy implements the CSS Grid specification faithfully, so documentation designed for the web should translate cleanly to Taffy's implementation. If you are interested in learning how to use CSS Grid, we would recommend the following resources:
- CSS Grid Garden. This is an interactive tutorial/game that allows you to learn the essential parts of CSS Grid in a fun engaging way.
- A Complete Guide To CSS Grid by CSS Tricks. This is detailed guide with illustrations and comphrehensive written explanation of the different Grid propertie and how they work.
In addition to the usual sizing/spacing proerties (size, min_size, padding, margin, etc), the following Grid style properties are supported on Grid Containers:
Property | Explanation |
---|---|
grid-template-columns |
The track sizing functions of the grid's explicit columns |
grid-template-rows |
The track sizing functions of the grid's explicit rows |
grid-auto-rows |
Track sizing functions for the grid's implicitly generated rows |
grid-auto-columns |
Track sizing functions for the grid's implicitly generated columns |
grid-auto-flow |
Whether auto-placed items are placed row-wise or column-wise. And sparsely or densely. |
gap |
The size of the vertical and horizontal gaps between grid rows/columns |
align-content |
Align grid tracks within the container in the inline (horizontal) axis |
justify-content |
Align grid tracks within the container in the block (vertical) axis |
align-items |
Align the child items within their grid areas in the inline (horizontal) axis |
justify-items |
Align the child items within their grid areas in the block (vertical) axis |
And the following Grid style properties are supported on Grid Items (children):
Property | Explanation |
---|---|
grid-row |
The (row) grid line the item starts at (or a span) |
grid-column |
The (column) grid line the item end at (or a span) |
align-self |
Align the item within it's grid area in the inline (horizontal) axis. Overrides align-items . |
justify-self |
Align the item within it's grid area in the block (vertical) axis. Overrides justify-items . |
The following properties and features are not currently supported:
- Subgrids
- Masonry grid layout
- Named grid lines
- Named areas:
grid-template-areas
andgrid-area
grid-template
orgrid
shorthand
See examples/grid_holy_grail.rs for an example using Taffy to implement the so-called Holy Grail Layout. If you want to run this example, the don't forget the enable the CSS Grid cargo feature:
cargo run --example grid_holy_grail --features grid
Ten new helper functions have added to the taffy prelude. These helper functions have short, intuitive names, and have generic return types which allow them to magically return the correct type depending on context. They make defining styles much easier, and means you won't typically need to use types like Dimension
or TrackSizingFunction
directly.
For example, instead of:
let size : Size<Dimension> = Size { width: Dimension::Points(100.0), height: Dimension::Percent(50.0) };
you can now write
let size : Size<Dimension> = Size { width: points(100.0), height: percent(50.0) };
And that same helper function will work other types like LengthPercentage
and MinTrackSizingFunction
that also have a Points
variant. There are also generic impl's for Size<T>
, Rect<T>
and Line<T>
which means if your node is the same size in all dimensions you can even write
let size : Size<Dimension> = points(100.0);
Available style helpers:
Type(s) | Helpers that work with that type | |
---|---|---|
LengthPercentage |
zero() |
Generates a Points variant with the value 0.0 |
points(val: f32) |
Generates a Points variant with the specified value |
|
percent(val: f32) |
Generates a Percent variant with the specified value.Note that the scale of 0-1 not 0-100. |
|
LengthPercentageAuto Dimension |
All helpers from LengthPercentage and... |
|
auto() |
Generates an Auto variant |
|
MinTrackSizingFunction |
All helpers from LengthPercentageAuto /Dimension and... |
|
min_content() |
Generates an MinContent variant |
|
max_content() |
Generates an MinContent variant |
|
MaxTrackSizingFunction |
All helpers from MinTrackSizingFunction and... |
|
fit_content(limit: LengthPercentage) |
Generates a FitContent variant with the specified limit.Nest the points or percent helper inside this function to specified the limit. |
|
fr(fraction: f32) |
Generates a Fraction (fr ) variant with the specified flex fraction |
|
NonRepeatingTrackSizingFunction |
All helpers from MaxTrackSizingFunction and... |
|
minmax(min: MinTrackSizingFunction, max: MaxTrackSizingFunction) |
Equivalent to CSS minmax() function. |
|
flex(fraction: f32) |
Equivalent to CSS minmax(0px, 1fr) . This is likely what you want if you want evenly sized rows/columns. |
|
TrackSizingFunction |
All helpers from NonRepeatingTrackSizingFunction and... |
|
repeat(rep: GridTrackRepetition, tracks: Vec<TrackSizingFunction>) |
Equivalent to css repeat() function. |
|
Vec<TrackSizingFunction> |
evenly_sized_tracks(count: u16) |
Equivalent to CSS repeat(count, minmax(0px, 1fr) |
AvailableSpace |
auto() |
Generates an Auto variant |
min_content() |
Generates an MinContent variant |
|
max_content() |
Generates an MinContent variant |
|
Size<T> |
Any helper that works for T will also work for Size<T> and will set both width and height to that value |
|
Rect<T> |
Any helper that works for T will also work for Rect<T> and will set top , left , bottom , and right to that value |
AlignContent
andJustifyContent
has been merged.JustifyContent
is now an alias ofAlignContent
and contains theStretch
variant.- This variant will be ignored (falling back to
Start
) when applied Flexbox containers. It is valid value for Grid containers.
AlignItems
andAlignSelf
have been merged.- The
Auto
variant ofAlignSelf
has been removed. You should now useOption::None
if you wish to specifyAlignSelf::Auto
. AlignSelf
is now an alias ofAlignItems
.JustifyItems
andJustifySelf
aliases have been added. These properties have no affect on Flexbox containers, but apply to Grid containers.
- The
Default
impls have been removed from all alignment types. This is because the correct default varies by property, and the types are now shared between multiple properties. TheStyle
struct still has a default for each alignment property, so this is considered unlikely to affect you in practice.
- New types
LengthPercentage
andLengthPercentageAuto
have been added.LengthPercentage
is likeDimension
but only contains thePoints
andPercent
variants, which allows us to increase type safety for properties that don't support theAuto
value.LengthPercentageAuto
is currently identical toDimension
but will allow us to expand dimension in future to support values likeMinContent
,MaxContent
andFitContent
.
- Some style properties have been updated to use either
LengthPercentage
orLengthPercentageAuto
instead ofDimension
. You will need to update your code, but it is recommended that you use the new style helpers (see above) rather than using the new types directly (although you certainly can use them directly if you want to).
- The
position
property is now renamed toinset
and is now in line with CSS inset specs - The
position_type
property is now renamed toposition
and is now in line with CSS position specs. ThePositionType
enum has been similarly renamed toPosition
.
- Added generic associated type to
LayoutTree
for aChildIter
, an iterator on the children of a given node. - Changed the
children
method ofLayoutTree
to return theChildIter
generic associated type to allow for custom tree storage implementations which do not store the children of a node contiguously. - Added
child_count
method toLayoutTree
for querying the number of children of a node. Required because thechildren
method now returns an iterator instead of an array. - Added
is_childless
method toLayoutTree
for querying whether a node has no children.
The AvailableSpace
enum has been moved from the layout
module to the style
module. If you are importing it via the prelude then you will unaffected by the change.
- Flexbox nodes sized under a min-content constraint now size correctly (#291)
- Aspect ratio is now applied correctly in many circumstances
- Absolutely positioned items now apply margins correctly
- Min/max size are now applied correctly
- Inset applied incorrectly to relatively positioned flexbox children when both
top
andbottom
orleft
andright
were specified (#348) - Fix case where column-gap style could be used in place of row-gap style (when using a percentage gap with an indefinite container size)
- Removed
top_from_points
,bot_from_points
,top_from_percent
, andbot_from_percent
methods removed fromRect<Dimension>
. These functions were incredibly specific for an unusual use case, so we would be surprised if anyone was using them. Please use the new style helpers instead. - Removed
min_main_size
,max_main_size
,min_cross_size
,max_cross_size
, andcross_size
methods fromStyle
. Use the more generalcross
andmain
methods directly on thesize
,min_size
, andmax_size
properties instead. - Removed
main_margin_start
,main_margin_end
,cross_margin_start
,cross_margin_end
fromStyle
. Use the more generalmain_start
,main_end
,cross_start
, andcross_end
on themargin
property instead.
- Border or padding on the horizontal axis could, in some cases, increase the height of nodes.
- In case of conflicts,
min_size
now overridesmax_size
which overridessize
(#261). This is the behaviour specified in the CSS specification, and was also the behaviour in Taffyv0.1.0
, but a regression was introduced in Taffyv0.2.0
. taffy::compute_layout
has been made public allowing Taffy to be used with custom storage (#263)
The gap property is now supported on flex containers. This can make it much easier to create even spacing or "gutters" between nodes.
Additionally we have a SpaceEvenly
variant to the AlignContent
enum to support evenly spaced justification in the cross axis (equivalent to align-content: space-evenly
in CSS)
Two debugging features have been added:
taffy::debug::print_tree(&Taffy, root)
- This will print a debug representation of the computed layout of an entire node tree (starting atroot
), which can be useful for debugging layouts.- A cargo feature
debug
. This enabled debug logging of the layout computation process itself (this is probably mainly useful for those working taffy itself).
A number of performance improvements have landed since taffy 0.1:
- Firstly, our custom
taffy::forest
storage implementation was ripped out and replaced with a much simpler implementation using theslotmap
crate. This led to performance increases of up to 90%. - Secondly, the caching implementation was improved by upping the number of cache slots from 2 to 4 and tweaking how computed results are allocated to chache slots to better match the actual usage patterns of the flexbox layout algorithm. This had a particularly dramatic effect on deep hierachies (which often involve recomputing the same results repeatedly), fixing the exponential blowup that was previously exhibited on these trees and improving performance by over 1000x in some cases!
Benchmark | Taffy 0.1 | Taffy 0.2 | % change (0.1 -> 0.2) |
---|---|---|---|
wide/1_000 nodes (2-level hierarchy) | 699.18 µs | 445.01 µs | -36.279% |
wide/10_000 nodes (2-level hierarchy) | 8.8244 ms | 7.1313 ms | -16.352% |
wide/100_000 nodes (2-level hierarchy) | 204.48 ms | 242.93 ms | +18.803% |
deep/4000 nodes (12-level hierarchy)) | 5.2320 s | 2.7363 ms | -99.947% |
deep/10_000 nodes (14-level hierarchy) | 75.207 s | 6.9415 ms | -99.991% |
deep/100_000 nodes (17-level hierarchy) | - | 102.72 ms | - |
deep/1_000_000 nodes (20-level hierarchy) | - | 799.35 ms | - |
(note that the table above contains multiple different units (milliseconds vs. microseconds vs. nanoseconds))
As you can see, we have actually regressed slightly in the "wide" benchmarks (where all nodes are siblings of a single parent node). Although it should be noted our results in these benchmarks are still very fast, especially on the 10,000 node benchmark which we consider to be the most realistic size where the result is measured in microseconds.
However, in the "deep" benchmarks we see dramatic improvements. The previous version of Taffy suffered from exponential blowup in the case of deeply nested hierachies. This has resulted in somewhat silly improvements like the 10,000 node (14-level) hierachy where Taffy 0.2 is a full 1 million times faster than Taffy 0.1. We've also included results with larger numbers of nodes (although you're unlikely to need that many) to demonstrate that this scalability continues up to even deeper levels of nesting.
Benchmarks vs. Yoga
Yoga benchmarks run via it's node.js bindings (the yoga-layout-prebuilt
npm package), they were run a few times manually and it was verified that variance in the numbers of each run was minimal. It should be noted that this is using an old version of Yoga.
Benchmark | Yoga | Taffy 0.2 |
---|---|---|
yoga/10 nodes (1-level hierarchy) | 45.1670 µs | 33.297 ns |
yoga/100 nodes (2-level hierarchy) | 134.1250 µs | 336.53 ns |
yoga/1_000 nodes (3-level hierarchy) | 1.2221 ms | 3.8928 µs |
yoga/10_000 nodes (4-level hierarchy) | 13.8672 ms | 36.162 µs |
yoga/100_000 nodes (5-level hierarchy) | 141.5307 ms | 1.6404 ms |
(note that the table above contains multiple different units (milliseconds vs. microseconds vs. nanoseconds))
While we're trying not to get too excited (there could easily be an issue with our benchmarking methodology which make this an unfair comparison), we are pleased to see that we seem to be anywhere between 100x and 1000x times faster depending on the node count!
taffy::Node
is now unique only to the Taffy instance from which it was created.- Renamed
Taffy.new_node(..)
->Taffy.new_with_children(..)
- Renamed
Taffy.new_leaf()
->Taffy.new_leaf_with_measure()
- Added
taffy::node::Taffy.new_leaf()
which allows the creation of new leaf-nodes without having to supply a measure function
- Renamed
taffy::Error
->taffy::error::TaffyError
- Replaced
taffy::error::InvalidChild
with a newInvalidChild
variant oftaffy::error::TaffyError
- Replaced
taffy::error::InvalidNode
with a newInvalidNode
variant oftaffy::error::TaffyError
- The following method new return
Err(TaffyError::ChildIndexOutOfBounds)
instead of panicking:taffy::Taffy::remove_child_at_index
taffy::Taffy::replace_child_at_index
taffy::Taffy::child_at_index
Taffy::remove
now returns aResult<usize, Error>
, to indicate if the operation was sucessful (and if it was, which ID was invalidated).
A new enum Taffy::layout::AvailableSpace
has been added.
The definition looks like this:
/// The amount of space available to a node in a given axis
pub enum AvailableSpace {
/// The amount of space available is the specified number of pixels
Definite(f32),
/// The amount of space available is indefinite and the node should be laid out under a min-content constraint
MinContent,
/// The amount of space available is indefinite and the node should be laid out under a max-content constraint
MaxContent,
}
This enum is now used instead of Option<f32>
when calling Taffy.compute_layout
(if you previously passing Size::NONE
to compute_layout
, then you will need to change this to Size::MAX_CONTENT
).
And a different instance of it is passed as a new second parameter to MeasureFunc
. MeasureFunc
s may choose to use this parameter in their computation or ignore it as they see fit. The canonical example of when it makes sense to use it is when laying out text. If MinContent
has been passed in the axis in which the text is flowing (i.e. the horizontal axis for left-to-right text), then you should line-break at every possible opportunity (e.g. all word boundaries), whereas if MaxContent
has been passed then you shouldn't line break at all..
- Several convenience constants have been defined: notably
Style::DEFAULT
Size<f32>.zero()
is nowSize::<f32>::ZERO
Point<f32>.zero()
is nowPoint::<f32>::ZERO
Size::undefined()
is nowSize::NONE
- Removed
taffy::forest::Forest
.taffy::node::Taffy
now handles it's own storage using a slotmap (which comes with a performance boost up to 90%). - Removed
taffy::number::Number
. UseOption<f32>
is used instead- the associated public
MinMax
andOrElse
traits have also been removed; these should never have been public
- the associated public
- Removed unused dependencies
hashbrown
,hash32
, andtypenum
.slotmap
is now the only required dependency (num_traits
andarrayvec
are also required if you wish to use taffy in ano_std
environment).
-
Miscellaneous correctness fixes which align our implementation with Chrome:
- Nodes can only ever have one parent
- Fixed rounding of fractional values to follow latest Chrome - values are now rounded the same regardless of their position
- Fixed computing free space when using both
flex-grow
and a minimum size - Padding is now only subtracted when determining the available space if the node size is unspecified, following section 9.2.2 of the flexbox spec
MeasureFunc
(and henceNodeData
and henceForest
and hence the publicTaffy
type) are nowSend
andSync
, enabling their use in async and parallel applications
-
Taffy can now be vendored using
cargo-vendor
(README.md is now included in package).
- the
order
field ofLayout
is now public, and describes the relative z-ordering of nodes - renamed crate from
stretch2
totaffy
- updated to the latest version of all dependencies to reduce upstream pain caused by duplicate dependencies
- renamed
stretch::node::Strech
->taffy::node::Taffy
- fixed feature strategy for
alloc
andstd
: these can now be compiled together, withstd
's types taking priority
- removed Javascript / Kotlin / Swift bindings
- the maintainer team lacks expertise to keep these working
- more serious refactors are planned, and this will be challenging to keep working through that process
- if you are interested in helping us maintain bindings to other languages, get in touch!
- the
serde_camel_case
andserde_kebab_case
features have been removed: they were poorly motivated and were not correctly additive (if both were enabled compilation would fail) - removed the
Direction
andOverflow
structs, and the correspondingdirection
andoverflow
fields fromStyle
- these had no effect in the current code base and were actively misleading
This is the final release of stretch
: migrate to the crate named taffy
for future fixes and features!
These notes describe the differences between this release and stretch
0.3.2, the abandoned crate from which this library was forked.
- updated assorted dependencies
- fixed an exponential performance blow-up with deep nesting
- fixed percent height values, which were using parent width
- recomputing layout no longer moves children of non-zero-positioned parent
- fixed broken Swift bindings