Avalonia

Commit Graph

Author	SHA1	Message	Date
Benedikt Stebner	acc1494308	Consolidate the Unicode segmentation trie and add UAX-29 sentence segmentation (#21875 ) * Pack the segmentation properties into one Unicode trie Grapheme, indic-conjunct, word, line and sentence break classes now share a single SegmentationTrie, together with the Emoji, Emoji_Presentation and Default_Ignorable_Code_Point flags that rode in the grapheme word. The UnicodeDataTrie word keeps general category, script and script extensions. - SegmentationTrieGenerator replaces GraphemeBreakClassTrieGenerator and emits Segmentation.trie.cs; UnicodeData.trie.cs is regenerated without the break fields it no longer carries. - SentenceBreakClass data and Codepoint.SentenceBreakClass expose UAX-29 sentence classes to callers. - WordSegment carries the source text slice alongside the code-unit and code-point readouts, matching the shape newer segment types use; the existing constructor and members stay for compatibility. Data regenerated against UCD 17.0.0. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> * Add UAX-29 sentence segmentation SentenceBreakEnumerator walks sentence boundaries over a UTF-16 span (rules SB1-SB11 and SB998), yielding SentenceSegment values of offset plus text slice. Sentence is the one segmentation granularity accessibility clients and IME navigation ask for that had no implementation, so callers had to approximate it with punctuation heuristics. The conformance suite runs against the Unicode SentenceBreakTest.txt data. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> * Make sentence segmentation linear in the text length Every rule that needed left-hand context rescanned the text backwards from the current position, so a run of spaces or closing brackets was walked once per character in it. Enumerating 40k spaces took 5.5 s; 40k closing brackets took 8.1 s, and both grew quadratically. - Fold the left-hand context into the enumerator as it advances: the last two significant classes for SB6/SB7, and a stage counter over the (STerm \| ATerm) Close* Sp* (Sep \| CR \| LF)? grammar for SB8-SB11. - Memoize the SB8 lookahead. A scan answers every start position it passed over, so the Close* Sp* run after an ATerm is scanned once, not once per position. - Classify with bitmasks, matching LineBreakEnumerator. Both pathological inputs now finish in 0.1 ms at 40k characters. Output is unchanged: the UCD SentenceBreakTest.txt suite passes, and 1.4M random strings over the break classes segment identically to the previous implementation. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> * Read the word segment length from the property valid for every segment WordSegment.Text is empty for segments built through the code-unit constructor that shipped in 12.1, so Text.Length holds only for segments that come from WordBreakEnumerator. Length is set by both constructors. - Measure the ellipsis break position from Length again. - Say on the Text property itself, where it is read, that it can be empty. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> * Benchmark word and sentence segmentation across text shapes The break enumerator benchmark covered three enumerators over random codepoint distributions, which measures trie lookup cost but not the per-character cost of the rules themselves. Long runs of one class are where a segmentation algorithm degrades, and nothing measured them. - Add the sentence enumerator to UnicodeBreakEnumeratorBenchmark. - Add UnicodeSegmentationScalingBenchmark: prose, Han, ZWJ emoji, and runs of spaces, closing punctuation, terminator-plus-spaces and regional indicators, each at two lengths so the cost per character is comparable between them. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> * Document the sentence break classes and trim the break unit SentenceBreakClass is public API with nothing but the UCD aliases to go on, so a caller reading Codepoint.SentenceBreakClass has to consult UAX #29 to learn what a value means or which rule reacts to it. - Describe the type and every member in terms of what the sentence rules do with it, keeping the UCD alias as a trailing comment. - Give SentenceBreakUnit a primary constructor and drop its Codepoint property, which nothing reads now that the class is stored directly. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> * Clean up documentation for SentenceBreakClass Removed remarks from the SentenceBreakClass documentation. --------- Co-authored-by: Claude Opus 5 <noreply@anthropic.com>	3 weeks ago
Benedikt Stebner	79ff4ff73e	[Text] Glyph Outline Part 5/13 - Shaped buffer glyph indices (#21501 ) * Fix glyph metrics; expose glyph bounding box GlyphMetrics previously stored the advance in Width/Height, leaving no field for the actual ink bounding box. Correct the contract: - Expand GlyphMetrics with AdvanceWidth / AdvanceHeight (and the bitmap / vertical-layout members XOffset / YOffset / VerticalOriginX/Y, reserved for later population). - Add GlyfTable.TryGetGlyphBounds: an allocation-free, header-only read of the glyph's control-point bounding box (xMin/yMin/xMax/yMax). Empty glyphs return true with zero bounds; out-of-range / short data return false. Composite glyphs use their header bbox without recursion. - TryGetGlyphMetrics (single + batch) now sources advances into AdvanceWidth / AdvanceHeight and the bounding box into Width / Height / XBearing / YBearing via TryGetGlyphBounds. Side bearings fall back to hmtx/vmtx when the glyph has no outline. The previously latent API had one test relying on the old behaviour (Width == advance); it's updated to assert AdvanceWidth. New tests cover TryGetGlyphBounds (range, empty, letter, descriptor parity) and the corrected metrics (advance placement, ink width distinct from advance, empty-glyph advance-without-ink, batch/single parity). * Add glyph-bounds benchmark GlyphBoundsBenchmark (Avalonia.Benchmarks): a BenchmarkDotNet comparison of the table-based bounds read against Skia's SKFont.GetGlyphWidths, in kernel (font pre-created) and per-run variants across 1 / 16 / 256 glyphs, with MemoryDiagnoser. * Add allocation-free batch glyph-bounds path Adds a bounds-only batch reader so ink-bounds computation (the GlyphRunImpl use case, which already has advances from shaping) doesn't pay for advance lookup or per-glyph indirection: - GlyphBounds: internal value type carrying the control-point box. - GlyfTable.GetGlyphBounds / GlyphTypeface.TryGetGlyphBounds: fetch the glyf and loca spans once per batch and read offsets + headers directly via BinaryPrimitives — no per-glyph ReadOnlyMemory.Span conversion, no intermediate Memory.Slice, no nested call chain. LocaTable exposes RawData / IsShortFormat for this. - Batch TryGetGlyphMetrics refactored onto the same span-cached reader. Measured against Skia's SKFont.GetGlyphWidths (bounds-only, in-process BenchmarkDotNet): the table path is faster at every run length (0.06× / 0.32× / 0.66× at 1 / 16 / 256 glyphs) and allocation-free, where Skia allocates 104 B/run for the SKFont. The earlier advance+bounds path via TryGetGlyphMetrics was ~2.2× slower than Skia at 256 glyphs; the bounds-only span-cached path is ~1.5× faster. The benchmark is updated to compare bounds-only on both sides. * Restore System.Buffers.Binary using in GlyfTable The glyf bounds readers (TryGetGlyphBounds / GetGlyphBounds) read the header int16s via BinaryPrimitives. The base pr2 commit trimmed this using as unused there, but it is needed once pr2b adds the bounds code — without it GlyfTable does not compile on this branch. * Clamp glyph bounding-box extents to non-negative GlyphBounds.Width/Height and both TryGetGlyphMetrics overloads computed the ink extent as (xMax - xMin) / (yMax - yMin). A malformed glyf header where xMax < xMin (or yMax < yMin) produced a negative value that, when narrowed to the ushort GlyphMetrics.Width/Height, wrapped to a huge positive extent instead of a safe zero. Clamp the extent to non-negative in GlyphBounds (the single source of truth for both the single and batch metric paths) and route the single TryGetGlyphMetrics path through GlyphBounds so both behave identically. For int16 coordinates the maximum extent is 65535 = ushort.MaxValue, so the clamped value always fits without overflow. Adds GlyphBoundsTests. * Allocate batch glyph-metric buffers only when their source is present Batch TryGetGlyphMetrics unconditionally sized the hMetrics, vMetrics and bounds temporary buffers to glyphIds.Length, even when the font has no hmtx, no vmtx, or no glyf table. For CFF/CFF2 fonts (no glyf) the bounds buffer is never read, and most fonts have no vmtx, so a large batch on such a font heap-allocated a per-glyph array that was pure waste. Size each buffer to zero — a free, empty stackalloc — when its source table is absent, so it only materialises (and only spills to the heap past 256 glyphs) when it will actually be filled and read. * Note the halved loca offsets in the batch bounds reader The short loca format stores every offset divided by two, which is why the batch reader doubles what it reads. LocaTable's single-glyph path already carries that note; the inline decode in the batch path did not. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> * Split the batch metrics loop on the glyf branch instead of per glyph The glyf check does not vary across the run, so it belongs above the loop rather than inside it. Splitting it also lets the bounds buffer live only in the branch that reads it, which drops the zero-length allocation that CFF / CFF2 fonts used to make just to satisfy the shared loop. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> * Add vertical glyph advance API to GlyphTypeface Mirror the existing TryGetHorizontalGlyphAdvance / ...Advances pair with vertical equivalents. Until now a vertical-layout caller (CJK, Mongolian) had no advance-only path on GlyphTypeface and had to go through TryGetGlyphMetrics, which after PR2b reads the glyf bounding box per glyph — pure waste when only the advance is needed. VerticalMetricsTable already exposes the per-glyph and batch readers internally (used by TryGetGlyphMetrics(batch)); this is a thin wrapper over them. Tests cover the Latin (no vmtx → false) and CJK (vmtx → positive advances) cases and batch/single parity using Inter and MiSans-Normal. Audit note: HorizontalMetricsTable and VerticalMetricsTable batch readers already span-cache (each fetches _data.Span once per call, reusing a BigEndianBinaryReader across glyphs). No metrics-table refactor is needed — the speculative Part 2 of the plan collapses to nothing. * Expose contiguous glyph indices span on ShapedBuffer Consumers that need a ReadOnlySpan<ushort> of glyph IDs (the upcoming GlyphTypeface.TryGetGlyphBounds batch path, SKFont.GetGlyphWidths, atlas lookups...) previously had to allocate a parallel ushort[] and project GlyphInfo.GlyphIndex into it. Carry that array on the ShapedBuffer instead: - ShapedBuffer rents a parallel ushort[] from ArrayPool alongside the existing GlyphInfo[] rental, exposed as `public ReadOnlySpan<ushort> GlyphIndices`. - The indexer setter syncs both, so HarfBuzz shaping and post-shape mutators (InterWordJustification rewrites the same GlyphIndex) keep the parallel view in lockstep with no caller change. - Split / WithBidiLevel slice the indices alongside the GlyphInfos at the same offsets — both share the parent's pooled arrays. - Dispose returns both rentals. Consumer wiring: - GlyphRunImpl: copies ShapedBuffer.GlyphIndices once via CopyTo when the source is a ShapedBuffer, dropping the per-glyph extract from the position loop. - TextFormatterImpl.CreateEmptyTextLine: switched to the pooled public ctor + indexer (was passing a heap-allocated GlyphInfo[] to the internal slice ctor, which now requires a parallel ushort slice). Tests cover indexer sync, overwrite, ascending split alignment, empty-leading split, and dispose clearing the view. --------- Co-authored-by: Claude Opus 5 <noreply@anthropic.com>	3 weeks ago
Steven Kirk	15a9c99fc6	Add typed binding expressions (#21552 ) * Update ncrunch config. * Initial groundwork for typed binding expressions. - Moved a bunch of stuff from `UntypedBindingExpressionBase` to `BindingExpressionBase` - Make various APIs accept `BindingExpressionBase` instead of `UntypedBindingExpressionBase` - Added a typed `IPropertyInfo` - Added a method on the `CompiledBindingPathBuilder` to build typed property accessors - Initial basic implementation of `TypedBindingExpression` * Track values in typed binding expression. * Support binding mode in typed binding expression. * Add typed binding expression benchmarks. Add Setup/Values benchmarks comparing the typed binding expression against the untyped CompiledBinding (as emitted by the XAML compiler today) and the reflection-based Binding. Writing the benchmarks surfaced that TypedBindingExpression never detached: it had no Dispose override, so disposing a binding (or rebinding the same property) leaked its PropertyChanged subscriptions and left it registered in the value store. Wire up disposal mirroring UntypedBindingExpressionBase: stop, detach handlers, remove from the value store and frame. This required widening IBindingExpressionSink.OnCompleted from UntypedBindingExpressionBase to BindingExpressionBase. Also fix a nullable warning that broke the Release build. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> @ * Don't box value in IBindingExpressionSink.OnChanged. Instead notify the sink of what's changed and let the sink read the value boxed or unboxed. * Move `IValueEntry` to `BindingExpressionBase` * Fall back to BindingExpression. Make `TypedPropertyElement` derive from `PropertyElement` and fall back to creating a standard boxing `BindingExpression` if any of the requirements for a typed binding expression are not met. * Don't use typed expressions for DataContext. * Emit typed binding expression from XAML compiler. When a compiled binding path is shape-eligible (single CLR property on a reference-type source, no transforms, instance getter) the compiler now emits a call to the typed CompiledBindingPathBuilder.Property<TSource, TResult> overload, producing a TypedPropertyElement that the runtime turns into a non-boxing TypedBindingExpression<TSource, TValue>. Other shapes continue to use the existing untyped emission path, and runtime fallback to BindingExpression still kicks in when modifiers (Source, Converter, StringFormat, etc.) make the typed expression unsuitable. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * Add failing tests for TwoWay typed binding source echo. TypedBindingExpression writes the source value back to the source both on attach and whenever the source raises a change: pushing the value to the target re-enters OnTargetPropertyChanged, which (in TwoWay mode) calls WriteValueToSource with the value that just arrived from the source. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> * Don't echo source value back to source in TwoWay typed binding. When TypedBindingExpression pushed the source value to the target, the resulting target PropertyChanged re-entered OnTargetPropertyChanged which, in TwoWay mode, wrote the value straight back to the source - a redundant round-trip on every attach and every source-originated change. Guard the source->target push with a flag and skip WriteValueToSource while it is set, so only genuine target changes are written back to the source. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> * Address typed binding code-review follow-ups. - Reset _isRunning when a typed binding is unsubscribed so the expression can be restarted (and re-subscribe to its source) if the value store reactivates the entry later, matching UntypedBindingExpressionBase.Stop(). - Remove a stray `using static PropertySetSnapshot` import. - Extract the duplicated IValueEntry value-unwrapping logic (shared by EffectiveValue<T> and DirectPropertyBase) into IValueEntry.TryGetValue. - Drop the unreachable UpdateSourceTrigger NotSupportedException in CreateTypedExpression; CanUseTypedBindingExpression already constrains it. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> * Fix TypedBinding_Values benchmark build error. The typed binding path was built with a single-argument `Property(propertyInfo)` call, which has no matching overload, so the benchmark project did not compile. Use the same typed three-argument overload as TypedBinding_Setup so the benchmark exercises the TypedPropertyElement path. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> * Skip typed binding expression for data-validation properties. TypedBindingExpression does not support data validation, but the eligibility check didn't account for it, so a directly-assignable single-property DataContext binding to a validation-enabled target (e.g. TextBox.Text, NumericUpDown.Value) would take the typed path and silently drop validation errors. Fall back to the untyped BindingExpression when the target property enables data validation. Data validation support in the typed expression can be added as a follow-up. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> * Clarify that data validation is not supported. * Address typed binding review comments. - React to null/empty PropertyChanged.PropertyName ("all properties changed") in the typed expression, matching the untyped path. - Swallow source getter exceptions raised during PropertyChanged. - Fall back to the untyped path for non-StyledElement targets, read-only sources and wider target types in TwoWay/OneWayToSource modes. - Fix a stray space in an exception message. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> * Share TypedBindingExpression.Attach validation across instantiations. Factor the validation logic (most notably the exception string formatting) out of the generic Attach method into a non-generic static helper. It only uses typeof(TValue), not TValue, so sharing it avoids duplicating the code per generic instantiation, a meaningful NativeAOT size saving (~3.3 KB => ~2.5 KB per instantiation). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> * Update API suppressions. These APIs are in a `[PrivateApi]`. * Use cached boolean boxes in typed binding expression. Merging main brought in the cached boxed booleans emitted for compiled binding property getters (#21065), but bindings that now take the typed binding expression path don't go through that getter: the value is boxed in TypedBindingExpression.GetUntypedValue when the target property isn't strongly typed (e.g. binding a bool to TextBlock.Tag). Box booleans via a shared cache there too, restoring the no-allocation-per-read behaviour. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01QkxawATDHa6rbPaDmPCLHZ * Extract the whole TypedBindingExpression.Attach body into AttachCore. `3198196b4c` only factored out the validation, leaving the field assignments in the generic method. As MrJul pointed out, none of those assignments depend on TSource or TValue either, so the entire body can move into a method that doesn't reference the generic parameters and be shared across instantiations rather than duplicated per instantiation. Attach is now just a forwarder passing typeof(TValue) as a Type, and AttachCore reads _sink/TargetProperty directly instead of having them threaded through as parameters, which reads better than the previous split. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01QkxawATDHa6rbPaDmPCLHZ * Move the method alongside the other privates. * Add failing tests for typed binding expression breaking changes. Each test binds the same property twice: once via a path which produces a TypedBindingExpression and once via a path which produces an untyped BindingExpression. The assertions describe the untyped behaviour, so the typed cases fail where the typed expression diverges: - A binding with no value (null or incompatible DataContext) publishes the target property's default value at the binding's priority instead of not contributing a value, overriding style setters and breaking inheritance. - Writing a value of another type, or null, to an object-typed target property throws from OnTargetPropertyChanged. - No binding error is logged when the DataContext is of the wrong type. - A throwing source getter leaves the stale value in place. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01B66teLaFQBr9GJe4a1aTZm * Fix typed binding expression divergences from the untyped path. - Don't notify when the expression has no value and had none before: doing so pushed the target property's default value at the binding's priority, overriding style setters and breaking property inheritance. - Read the target property's new value defensively. The binding value type only needs to be assignable to the target property type, so the property can hold a value which isn't a TValue; casting it threw out of the caller's SetValue. - Log a binding error when the DataContext can't be converted to the source type, and when the source getter throws. The getter also now clears the value rather than leaving the stale one in place. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01B66teLaFQBr9GJe4a1aTZm * Pass correct priority to child bindings. The priority here doesn't actually have any effect as the binding isn't in a value store, but there's no reason to not use the correct one. * Use the typed overload. * Set initial priority to default priority. Makes `TypedBindingExpressionBase` and `UntypedBindingExpressionBase` have the same behavior here. * Remove unused vars. * We now have a typed value here. * Publish typed binding values on the UI thread. The source's PropertyChanged event can be raised on any thread, so marshal the notification to the UI thread as UntypedBindingExpressionBase does. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01B1ZwmCJkThPSeViLnRxoYs * Share the property info cache lookup. Emit and EmitTyped had identical cache lookups, differing only in which dictionary they searched. Also delay constructing the generic types in EmitTyped until a cache miss, as they aren't needed on a hit. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01B1ZwmCJkThPSeViLnRxoYs * Share the property accessor delegate emission. EmitFunc was duplicated as a local function in both Emit and EmitTyped. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01B1ZwmCJkThPSeViLnRxoYs --------- Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>	1 month ago
Benedikt Stebner	aee3f68551	TextCollapsingProperties bidirectional-text (BiDi) correctness fixes (#21446 ) * Fix BiDi correctness across TextCollapsingProperties implementations Audit and fix logical-vs-visual ordering bugs in every text-collapsing implementation and the helpers they share, so trimming always operates in logical order (the consumer re-runs the BiDi reorderer in FinalizeLine, so pre-applying visual order would be reordered a second time). - LogicalTextRunEnumerator returned the first run on every MoveNext when _indexedTextRuns was null. - TextLeadingPrefixCharacterEllipsis: walked visual-order runs while splitting at logical offsets, guarded the wrong field in the ctor, hardcoded LeftToRight for the symbol, appended the suffix in reverse-logical order, and mis-tracked the width budget across multi-run lines. - TextPathSegmentEllipsis: crashed via Split(0) in the fallback path, and measured RTL segments as ~0 width (it assumed ascending cluster ids). - ShapedTextRun.TryMeasureCharacters/Backwards returned visual-order counts for RTL runs. ShapedBuffer gains logical-order, cluster-cache-backed width helpers (GetCharRangeWidth, FindLeading/FindTrailingCharCountWithinWidth) that are safe in the one-char-per-cluster simple mode (null _clusterStartChars). Adds a BiDi characterization test suite and documents the logical-order Collapse contract. * Align text width comparisons to epsilon-tolerant MathUtilities Several width-fit comparisons in the measure/collapse path used raw operators with no epsilon, while the cluster-cache hot path (MeasureCharactersThatFit) and the ShapedTextRun branch of MeasureLength already went through MathUtilities. Replace the raw float comparisons with their epsilon-aware equivalents (<= -> LessThanOrClose, < -> LessThan, > -> GreaterThan) in ShapedBuffer.FindLeading/FindTrailingCharCountWithinWidth, TextPathSegmentEllipsis, and TextLeadingPrefixCharacterEllipsis. Also unify the DrawableTextRun branch of TextFormatterImpl.MeasureLength with the ShapedTextRun branch (>= -> GreaterThan) so the "whole run fits" boundary is identical for both run types: a run fits when width <= remaining (tolerant) and overflows only when strictly greater. Integer char-offset/index comparisons and the availableWidth <= 0 positivity guards are intentionally left as raw comparisons. * Cleanup some comments * Fix Split(0) crash in TextLeadingPrefixCharacterEllipsis suffix collection When a logical-tail run fits entirely within the remaining suffix budget, TryMeasureCharactersBackwards returns suffixCount == run.Length, so Split(run.Length - suffixCount) evaluated to Split(0), which throws ArgumentOutOfRangeException. This was reachable on multi-run lines whose post-split tail contains a small later run that fully fits. Guard the split on splitAt > 0 and use the whole run as-is when it fully fits (dropping it only when suffixCount == 0) — the same boundary pattern already applied to TextPathSegmentEllipsis. Add a multi-run regression test. * Remove duplicate ShapedBuffer.MeasureCharactersThatFit MeasureCharactersThatFit and FindLeadingCharCountWithinWidth ran the identical cluster-prefix binary search; the only functional difference was the unused `out widthConsumed`. Drop the former and repoint its callers (TextFormatterImpl.MeasureLength and the benchmark) at FindLeadingCharCountWithinWidth, which also unifies the availableWidth<=0 guard. The one test that asserted the consumed width now recovers it via GetCharRangeWidth(0, fit). Also tidy GetCharRangeWidth's simple-mode bounds to use Math.Clamp.	4 months ago
Benedikt Stebner	42bba0b859	[Text] Unicode and text layout optimizations (#21400 ) * perf(text): O(1) width queries on ShapedBuffer, cached LB-class lookup - ShapedBuffer: lazy, pooled cluster-prefix cache shared across Split; adds TotalGlyphAdvance / MeasureCharactersThatFit. - GlyphRun/TextFormatterImpl/Skia GlyphRunImpl: read the cache, drop duplicate scans; SplitTextRuns now returns firstLength via out. - LineBreakEnumerator: cache Next/PreviousClass per advance; remove unused LineBreakPairTable. - Tests + benchmarks for the new paths and a dotnet-trace harness. ~1.32x faster / -5% alloc on the emoji-wrap micro-benchmark; no observable behaviour or public API change. * perf(TextFormatting): inline LineBreakEnumerator rule dispatch Replace the BreakUnitDelegate[] s_rules array dispatch in LineBreakEnumerator.ExecuteRules with a sequence of direct static calls and a `goto Done` early-exit. This removes 42 indirect calls per codepoint and lets the JIT reason across rule boundaries, which in turn makes [MethodImpl(AggressiveInlining)] meaningful — JIT cannot inline through delegate.Invoke, so the attribute was a no-op in the previous shape. Selectively apply AggressiveInlining to the small single-condition rules (LB03, LB04, LB06, LB07, LB08a, LB11–LB15d, LB18, LB20, LB21b, LB22, LB29, LB31) and let the JIT decide on the larger ones to avoid bloating the merged ExecuteRules. The static BreakUnitDelegate[] s_rules array is removed. Benchmarks (BDN default job, --inProcess, N=13–22, rel. StdDev <1.5%): UnicodeBreakEnumeratorBenchmark.LineBreakEnumerator_Sequence Ascii 154.6 µs -> 24.84 µs (6.22x) Bmp 190.2 µs -> 32.81 µs (5.80x) Supplementary 227.6 µs -> 37.56 µs (6.06x) TextLayoutProfile.BuildEmojisWrapped Before (branch, pre-inline) 804.5 µs / 570.15 KB After (branch, post-inline) 547.0 µs / 570.15 KB (-32.0%) vs upstream/master (1061 µs) -> ~1.94x total speedup No allocation change; CPU-only dispatch reshape. All LineBreakEnumerator unit tests pass (5/5). * Refactor ShapedBuffer to share pool storage via IRef + generation counter Wrap the ArrayPool-rented glyph and cluster-cache arrays in a small PooledArray<T> disposable and expose them through IRef<T>. Split children and WithBidiLevel aliases now Clone() the refs instead of borrowing raw pool arrays, so the backing storage survives until every sibling has been disposed - eliminating the UAF risk that existed when a parent was disposed before its children. Add a per-glyph-holder generation counter (Volatile.Read / Interlocked .Increment). The indexer setter bumps the counter on every write, and EnsureClusterCache compares its recorded generation against the holder's current value, rebuilding on mismatch. This lets us drop the previous "no mutation after Split/WithBidiLevel" contract: mutations performed through any sibling now propagate to the others' caches transparently. Dispose is made idempotent via a _disposed guard so overlapping cache eviction and TextLine teardown only release the IRefs once. Adds ShapedBufferSharedStorageTests covering sibling lifetime, Dispose idempotency, and generation-driven cache invalidation across Split children and WithBidiLevel aliases. * test(ShapedBuffer): cover cluster-cache sharing across Split/WithBidiLevel aliases Add a regression test hook (`ClusterPrefix`) exposing the backing cluster-prefix array reference, plus two tests that mutate a parent buffer before aliasing it and assert the alias reuses the parent's pooled prefix array instead of rebuilding. Guards against forgetting to propagate `_cacheGeneration` to alias buffers, which would silently defeat the cached-split fast path. * Correctly use MathUtilities.LessThanOrClose --------- Co-authored-by: Julien Lebosquain <julien@lebosquain.net>	4 months ago
Nikita Tsukanov	be6bc063f1	Pass original logical size to render targets (#21380 ) * Pass original logical size to render targets * Change FromSizeCeiling method to internal access * api diff again?	5 months ago
Benedikt Stebner	f41f97ee26	Implement TextRunCache (#21030 )	6 months ago
Javier Suárez	af0f61a02b	[Feature] Expand Page-type navigation: CarouselPage (#20871 ) * Added CarouselPage * Renamed DrawerBreakpointWidth to DrawerBreakpointLength * More unit tests * More tests * Improve Carousel gestures * Make SelectionChanged a RoutedEvent * Change TabbedPage IsGestureEnabled to true for consistency * Allow to clear CurrentPage * Reuse Tab Icon Path (perf) * Avoid duplicated ResolvePageAtIndex * More tests * A new showcase * Improve CarouselPage Transitions sample * More changes * Expose IsNavigating * A11y improvements * Updated showcase samples * More changes * Fix vertical carousel page gestures * More tests * More changes in showcases * More changes * Refactor SwipeGestureRecognizer * More changes * More changes * Added new CommandBar Events sample * Avoid duplicated subscriptions in CommandBar * Avoid invalid SelectedIndex values * Avoid duplicated subscription * Fix SwipeGestureRecognizer non needed allocation * Added Easing to stack and modal navigation default transitions * Preserve NavigationPage modal stack across visual tree detach/reattach * More fixes and tests * More changes * Simplify icon logic * More fixes, samples and tests * Fixed tests * Changes based on feedback * Updated test * Fixed tests * Update API suppressions --------- Co-authored-by: Julien Lebosquain <julien@lebosquain.net>	7 months ago
Nikita Tsukanov	ae6a085ebc	Don't tick with render loop when app is idle (#20873 ) * Don't tick with render loop when app is idle * Update src/Avalonia.Base/Rendering/Composition/Transport/BatchStreamArrayPool.cs Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com> * wip * wip * api diff * fixes * Address review: clear wakeupPending at tick start, guard CarbonEmissionsHack subscriptions - Clear _wakeupPending at start of TimerTick so wakeups already processed by the current tick don't force an unnecessary extra tick - Guard CarbonEmissionsHack against duplicate subscriptions using a private attached property Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com> * Fix lock-order inversion in Add/Remove vs TimerTick Move Wakeup() and Stop() calls outside the _items lock in Add/Remove to prevent deadlock with TimerTick which acquires _timerLock then _items. Add/Remove are UI-thread-only so the extracted logic remains safe. Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com> * Address review: remove unused usings, guard double Stop(), fix SleepLoop extra frame - Remove unused using directives from IRenderLoopTask.cs - Guard TimerTick Stop() with _running check to prevent double Stop() when Remove() already stopped the timer - SleepLoopRenderTimer: use WaitOne(timeout) instead of Thread.Sleep so Stop() can interrupt the sleep, and recheck _stopped before Tick Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com> * Fix DisplayLinkTimer foreground handler bypassing render loop state Only resume the display link on WillEnterForeground if the timer was calling Start() to avoid setting _stopped=false when the render loop had the timer stopped. Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com> * Fix DisplayLinkTimer thread safety, revert global NU5104 suppression - Stop() now only sets _stopped flag; OnLinkTick() self-pauses the CADisplayLink from the timer thread to avoid thread-affinity issues - Revert NU5104 global suppression in SharedVersion.props Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com> * Cap _ticksSinceLastCommit to prevent int overflow Stop incrementing once it reaches CommitGraceTicks to prevent wrapping negative and keeping the render loop awake indefinitely. Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com> * refactor: remove Start/Stop from IRenderTimer, merge into Tick setter Timer start/stop is now controlled entirely by setting the Tick property: non-null starts, null stops. This eliminates the explicit Start()/Stop() methods from IRenderTimer, making the API simpler. DefaultRenderLoop controls the timer purely through Tick assignment under its _timerLock. A new _hasItems flag tracks subscriber presence since Tick is now transient (null when idle). Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com> * fix: address review comments on timer thread safety and guards - ChoreographerTimer: add _frameCallbackActive guard to prevent double PostFrameCallback from both Tick setter and SubscribeView - ServerCompositor: cap _ticksSinceLastCommit at int.MaxValue - SleepLoopRenderTimer: make _tick volatile, remove _stopped recheck (guard moved to DefaultRenderLoop) - DefaultRenderLoop: add _running check at tick start to drop late ticks - ThreadProxyRenderTimer: add lock for internal state manipulation - DisplayLinkTimer: add lock for all internal state manipulation - Re-add NU5104 suppression to SharedVersion.props Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com> * fix: make _hasItems volatile for cross-thread visibility Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com> * fix: guard against redundant starts in DefaultRenderTimer, make _tick volatile across all timers Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com> * Remove CarbonEmissionsHack, revert iOS/Android timers to always-ticking - Delete CarbonEmissionsHack class and its XAML reference - Revert DisplayLinkTimer (iOS) to original always-ticking implementation - Revert ChoreographerTimer (Android) to original always-ticking implementation - Add TODO comments for future start/stop on RenderLoop request Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com> * Fix DirectCompositionConnection WaitOne not respecting process exit cancellation Use WaitHandle.WaitAny with both _wakeEvent and cts.Token.WaitHandle so the loop can exit when ProcessExit fires while the timer is stopped. Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com> * timers * XML docs * Cache delegate --------- Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com> Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>	7 months ago
Javier Suárez	7bc087c658	[Feature] Page-based navigation system (#20794 ) * Added CommandBar * Added ContentPage * Added NavigationPage * Added TabbedPage * Added DrawerPage * More showcases * More tests * Added Connected Animations * More samples * More fixes * More fixes * Updated samples * More fixes * Fix build errors * More fixes * More changes * More changes * More fixes * Added benchmarks * Updated SwipeGestureRecognizer * Fixes based on PR feedback * More fixes * Removed Appearing/Disappearing events * Fix merge issue * More changes * Fix build * Remove TabbedPage brush properties * New showcase sample * More changes * Moved navigation brushes properties to resources * Fix default connected animation duration fallback * CommandBar returns empty list with null * Ignore AVP1030 on PrimaryCommands/SecondaryCommands for now * Updated samples * More changes and tests * Updated samples * Updated render tests * Updated tests * Added Page AutomationPeers --------- Co-authored-by: Julien Lebosquain <julien@lebosquain.net>	7 months ago
Nikita Tsukanov	b2ecb21e03	Render target/surface housekeeping (#20811 ) * Render target/surface housekeeping * diff and build * build and stuff * really * Make more APIs private * Apply suggestions from code review Co-authored-by: Julien Lebosquain <julien@lebosquain.net> * Address review comments --------- Co-authored-by: Julien Lebosquain <julien@lebosquain.net>	7 months ago
David Federman	ea7e9a7aed	Onboard onto Central Package Management (#20704 ) * Onboard onto Central Package Management * Remove SharpDX * Add back <clear /> to NuGet.config package source mapping * Package mapping * Inline props where appropriate * Lost a space	7 months ago
Nikita Tsukanov	697cba7193	Composition system SP1 (#20497 ) * baseline benchmark * Composition system SP1 * Allow reusing the same BitmapCache object for multiple visuals * Fix warnings * Address review * API suppressions --------- Co-authored-by: Julien Lebosquain <julien@lebosquain.net>	8 months ago
Julien Lebosquain	656d691940	Move ConstructorArgumentAttribute and remove BindingMode from ReflectionBinding constructor (#20589 ) * Move ConstructorArgumentAttribute to Avalonia.Base assembly * Move IOptionalFeatureProvider to Avalonia namespace * Remove BindingMode from ReflectionBinding constructor * Update API suppressions	8 months ago
Benedikt Stebner	e8424283fe	Universal GlyphTypeface implementation (#19852 ) * Introduce a universal IGlyphTypeface implementation that does not rely on any platform implementation * Revert changes * Fix Android * Make the test happy * Fix build * Update baseline * Fix naming * Fix headless * Move interfaces to dedicated files Make GlyphTypeface.GlyphCount an integer * Fix GlyphCount * Make IGlyphTypeface NotClientImplementable * Make sure we cache platform typefaces by their desired name, style, weight and stretch * Update baseline * Only use IGlyphTypeface * Fix Android * Try to clear the buffer before we encode somethimg * Add needed test font * Add more unit tests * Reduce allocations * Remove Direct2D1 test files * More tests * More complete table implementations * More adjustments * Use batch APIs * Handle invalid timestamps * Update baseline * Introduce a CharacterToGlyphMap struct for faster access * Remove AggressiveInlining * Remove AggressiveInlining * Make the head table optional for legacy fonts * Remove Load method. Fix TextBlockTests * Fix nullables * Remove redundant folder * Update Api baseline * revert diff helper changes * revert changes * Use bare minimum font for Headless platform and introduce a test font manager that uses the Inter font for testing. * Add missing font file for Headless platform --------- Co-authored-by: Gillibald <stebner@avaloniaui.net> Co-authored-by: Julien Lebosquain <julien@lebosquain.net>	9 months ago
Youssef Victor	c9e210eb70	Update to xunit.v3 (#20331 ) * Update to xunit.v3 * Little more progress * More fixes * Keep VSTest supported * Adjust Nuke * Few fixes * Fix for xunit 2 * Fix GetData override * Adjust * Use MTP for xunit 2 * Fix test * Better fix * --no-progress * Few more fixes * no progress * Fix test * Better fix * TRX * Move to Directory.Build.props * Unify on MTP v2 * Update * Update to stable * 1.0.1 * 1.0.2 * Fix some warnings * Fix more warnings * Fix more warnings	9 months ago
Steven Kirk	f2fc8d02f4	Move `IBinding` to `BindingBase` and tidy up some binding APIs (#19589 ) * Update ncrunch config. * Tidy up reflection and multi-binding APIs: - Move `BindingBase` and `MultiBinding` into Avalonia.Base - `BindingBase` becomes a true base class for all bindings, and contains only the `Instance` method - Properties common between reflection and compiled bindings are moved into `StandardBindingBase` - `Binding` is moved to Avalonia.Base and renamed to `ReflectionBinding` - A compatibility shim for `Binding` remains in Avalonia.Markup - Remove `IBinding` and `IBinding2` - Remove `ITreeDataTemplate's usage of `InstancedBinding` - Remove `NativeMenuBarPresenter`s usage of `InstancedBinding` - Remove `InstancedBinding` as it is now unused This required an update to the DataGrid submodule: cell data validation has been temporarily removed as this used `InstancedBinding`. * `Instance()` => `CreateInstance()`. The use of "Instance" as a verb is quite unusual apparently ;) * Seal classes where appropriate. * Seal classes where appropriate. * Remove `StandardBindingBase`. Simply duplicate the members in reflection and compiled binding classes. * Delete deleted submodule directory. * Add missing attribute. Fixes compile error. * Fix reference to removed class. * Update suppressions.	9 months ago
Julien Lebosquain	ce0a1c6b3e	Update dependencies (#20139 ) * Update dependencies * Partially revert appium upgrades * Fix integration tests	10 months ago
Dan Walmsley	4ff49db604	Optimization: Add an optimized path for notifying property changes for inherited va… (#18223 ) * Add an optimized path for notifying property changes for inherited values Avoids many allocations of event args. * fix unit tests with hack to prove concept. * update docs and remove value tuple that could cause allocation. * add a benchmark for inherited property change notifications.	2 years ago
Julien Lebosquain	aac83dcd53	Layout performance improvements (#18315 )	2 years ago
Julien Lebosquain	edceb96f34	Improve AvaloniaObject.GetValue performance (#15342 ) * Improve AvaloniaProperty.GetValue performance * Remove unneeded MethodImpl on Optional.GetValueOrDefault * Use ReferenceEqualityComparer instead of TypeEqualityComparer --------- Co-authored-by: Jumar Macato <16554748+jmacato@users.noreply.github.com>	2 years ago
Max Katz	326ef7c9fd	Add .NET 8 target (Avalonia 11.1) (#14535 ) * Add .NET 8 target to packable projects * Update test projects * Update sample projects to target .NET 8 * Update main Avalonia package to target .NET 8 as well * Remove MSBuildEnableWorkloadResolver hack and some minor change * Use net8.0 in nuke build as well * Pin 8.0.0 SDK * Adjust API validation * Pin a valid version * Remove net461 target from Avalonia package * Remove unused net6.0 target from Avalonia.Designer.HostApp * Adjust API diff * Update ControlCatalog.Android.csproj * Remove MSBuildEnableWorkloadResolver * Fix Browser issues on .NET 8 * Fix .NET 8 error * Fix merge conflicts * Replace explicit TFMs in .csproj files with a centralized registry * Fix merge conflict * Fix AvsLegacyWindowsTargetFrameworks and add hacks for the tests * Update XCode project CommandLineArgument --------- Co-authored-by: Nikita Tsukanov <keks9n@gmail.com>	3 years ago
Steven Kirk	3b1eb338e6	Binding System refactor (#13970 ) * Update ncrunch config. * WIP: Benchmarks * Initial refactor of binding infrastructure. - `ExpressionObserver` has been removed and its functionality merged with `BindingExpression` - `BindingExpression` handles all types of `BindingMode` itself; doesn't require `BindingOperations.Apply` to set up a separate observable for `TwoWay/`OneWayToSource` bindings - This allows us to fix some long-standing issues with `OneWayToSource` bindings - Expression nodes have been refactored - No longer split between `Avalonia.Base` and `Avalonia.Markup` - Categorize them according to whether they use reflection or not A few tests are failing around binding warnings: this is because the next step here is to fix binding warnings. * Make default binding Source = UnsetProperty. Null is a theoretically valid value for `Source`; setting it to null shouldn't mean "use the data context". * Move logging to BindingExpression. As `BindingExpression` now has enough information to decide when it's appropriate to log an error/warning or not. Fixes #5762 Fixes #9422 * Add compatibility hack for older compiled bindings. Previously, `CompiledBindingPathBuilder` didn't have a `TemplatedParent` method and instead the XAML compiler rewrite templated parent bindings to be a `$self.TemplateParent` property binding. resulting in extraneous logs. Add a constructor with an `apiVersion` to `CompiledBindingPathBuilder` which will be used by newer versions of the XAML compiler, and if a usage is detected using an `apiVersion` of 0, then upgrade `$self.TemplatedParent` to use a `TemplatedParentPathElement`. * Log errors from property accessors. * Don't log errors for named control bindings... ...on elements which aren't yet rooted. * Log errors for failed conversions. * Use consistent wording for binding warnings. "Could not convert" instead of "Cannot convert". * Log warnings for converter exceptions. * Don't convert new TargetTypeConverters each time. * Added failing test for implicit conversion. * Support cast operators in compiled bindings. A bit of a hack as we'd ideally not be using reflection when using compiled bindings. * This shouldn't be a public API. Should only be used for tests. * Make enum/int conversion work. * Check for SetValue equality after conversion. And also use "identity equals" where value types and strings use `object.Equals` and reference types use `object.ReferenceEquals`. * Added ConverterCulture back to bindings. * Fix merge error. Removed deleted files from csproj that were re-added due to indentation changes. * Use BindingExpression directly in ValueStoe. * Introduce BindingExpressionBase. And `UntypedBindingExpressionBase`. * Make TemplateBinding a BindingExpression. * Make DynamicResource use a BindingExpression. * WIP: Start exposing a BindingExpression API. * Finish exposing a BindingExpression API. * Fix OneTimeBinding. * Remove unneeded classes/methods. * Don't call obsolete API. * Make BindingExpressionBase the public API. This matches WPF's API. * Added BindingExpressionBase.UpdateTarget. * Initial implementation of UpdateSourceTrigger. * Don't use weak references for values. If they're boxed values, they can get collected. * No need for virtual/generic methods here now. * Reintroduce support for binding anchors. Turns out these were needed by animations, just our animation system has no unit tests so I missed that fact earlier. Add a basic animation unit test that fails without anchor support, and add binding anchors back in. Currently a private API as I suspect this feature shouldn't be needed outside the framework. * Include new property in clone. And add real-life example of `UpdateSourceTrigger=LostFocus` to BindingDemo. * Fix merge error. * Updated BindingExpression tests. - Make them run for both compiled and reflection bindings (found a bunch of tests that fail with compiled bindings) - Make them not depend on converting the `BindingExpression` to an observable and instead test the end result of the binding on an `AvaloniaObject` * Fix compiled binding indexer tests. * Use data validation plugins in PropertyAccessorNode. Added a warning suppression for now: we may need a separate `DataValidators` list for AOT-friendly plugins. * Don't separate plugins by reflection. `DataAnnotationsValidationPlugin` is public and so it can't be moved. No point in moving the others if this one will be in the wrong place. * Remove unneeded methods. * Make reflection binding tests use a string. Convert the `System.Linq.Expression` to a string and then use this, as reflection bindings will always be instanced with a string path. * Added TODO12 plan for IBinding2. * Use more specific exception. * Fix nits from code review. * Make expression nodes sealed where possible. * Unsubscribe on Stop, don't re-subscribe. D'oh. * Tweak ExpressionNode lists. Saves a few K in benchmarks and it's a cleaner API. * Add a pooled option in BindingExpressionGrammar. Micro-optimization. * Avoid allocations when enumerating binding plugins. * Add IBinding2 support to observable bind overloads. In the case of `TemplateBinding`, the `IObservable<object?>` bind overload is selected by C#. Add an explicit check for an `IBinding2` here to use the more performant code-path. * Remove disposed binding from ImmediateBindingFrame. * Added TemplateBinding benchmarks. * Remove duplicate items. Seems to have been caused by a merge error. * Fix exception when closing color picker. And add tests. * Don't skip converter when binding to self. * Don't pass UnsetValue to converters. This follows WPF behavior. * Log element name if present. More useful than just logging the control hash code. * Respect binding priority. * Throw on mismatched binding priorities. We don't want to respect the binding priority in this case as it breaks `TemplateBindings` when the default `LocalValue` priority is passed. Instead make sure that the priority parameter matches that of the expression. This reverts commit `a72765d705`. * Convert to target type in TemplateBinding. * Short-circuit target type conversion for same types.	3 years ago
Max Katz	4c0873eb9a	Fix build	3 years ago
Steven Kirk	04c50f8bf2	Make style selector internals internal.	3 years ago
Steven Kirk	5b7d5a33ae	Fix benchmarks compile error,	3 years ago
Max Katz	810558140f	Introduce static AssetLoader	3 years ago
Max Katz	4aa6035ec3	Reuse text and geometry related headless mocks in Avalonia unit tests and benchmarks	3 years ago
Max Katz	a507e92b31	Fix theme-dependend markup extensions not knowing current theme context	3 years ago
Benedikt Stebner	0442c3b337	Introduce advanced render options	4 years ago
Nikita Tsukanov	312f1250e2	Don't use AvaloniaObject in IPlatformRenderInterface	4 years ago
Nikita Tsukanov	0f4ed2a4b9	Updated tests to use IDispatcherImpl	4 years ago
Nikita Tsukanov	17492be5b6	Removed RunLoop from IPlatformThreadingInterface	4 years ago
Dariusz Komosinski	d9d208b935	Add benchmark for Or selector and avoid extra virtual calls getting list count.	4 years ago
Benedikt Stebner	147db08f90	Introduce GlyphRun.Bounds	4 years ago
Nikita Tsukanov	ae1fcfed51	Refactored DrawingContext and VisualBrush, added DrawingBrush (#10419 ) Refactored DrawingContext and VisualBrush, added DrawingBrush	4 years ago
Benedikt Stebner	f5caa61051	[SKIA] UseOpacitySaveLayer feature switch (#9964 ) * [SKIA] Introduce UseOpacitySaveLayer feature switch * Only maintain _currentOpacity when OpacitySaveLayer is disabled	4 years ago
José Pedro	f8af158a38	Updated packages used by tests.	4 years ago
Julien Lebosquain	5e13c5b59a	Implemented layout and render time graph overlays	4 years ago
Nikita Tsukanov	abf8819947	Implemented pixel format transcoding for Bitmap and WriteableBitmap	4 years ago
Julien Lebosquain	5ffd961742	Perf: improved text wrapping	4 years ago
Benedikt Stebner	cb884c6e9e	Reintroduce customizable GlyphRun.BaselineOrigin Reintroduce Brush IBrush inheritance	4 years ago
Julien Lebosquain	7a1f74a3d3	Benchmarks: option to use Skia for text layout	4 years ago
Julien Lebosquain	10a3b79d12	Perf: various misc text layout optimizations	4 years ago
Benedikt Stebner	07c11c75e7	Use IGlyphRunImpl in the IDrawingContextImpl	4 years ago
Julien Lebosquain	89a78f557b	Don't keep the text layout buffers around if they're too large	4 years ago
Nikita Tsukanov	d714f37fce	Implemented interop with externally managed GPU memory	4 years ago
Julien Lebosquain	fd0720fc56	Updated text layout benchmark with trimming/wrapping	4 years ago
Julien Lebosquain	4144be11fe	Pass GlyphInfo directly to GlyphRun	4 years ago
Max Katz	a0d22499cd	Fix benchmarks build	4 years ago

1 2 3 4 5

221 Commits (themes/fluent2)