* Introduce helper function fail

* Simplify signatures for assertContentCloseEnough

* In the middle of refactoring

* Don't use nested assertions in assertCloseEnough family of functions

* Platform-aware default epsilon

* Correct one of assertion error messages

* Yet another typo in a commit message

* TextBlobBuilder: common appendRun

* TextBlobBuilder: common appendRunPosH

* TextBlobBuilder: canAppendRunWithRSXform, canAppendRunWithPosition, canAppendMultipleTimes

* drop commit (duplicates code from PR #356)

* TextBlobBuilder: nAppendRun impl

* TextBlobBuilder: appendRunPosH impl

* TextBlobBuilder: AppendRunPos impl

* TextBlobBuilder: appendRunRSXform impl

* fix js test
Co-authored-by: Oleksandr Karpovich <oleksandr.karpovich@jetbrains.com>

* Split TypefaceTest to the part that passes on all platforms and on JVM only

* Breaking for jvm in order to move faster to newer syntax

* org_jetbrains_skia_Typeface__1nGetVariationAxes returns nothing

* Introduce _nGetVariationAxesCount (for JVM only so far)

* Pass count to _nGetVariationAxesCount

* Finally passing the right way

* SOME DEBUG

* org_jetbrains_skia_Typeface__1nGetVariationAxes for native and js

* Restore axis on all platforms

* Passing isHidden boolean flag as a separate int field

* org_jetbrains_skia_Typeface__1nGetVariations on native and js

* Preparations for taming org_jetbrains_skia_Typeface__1nGetTableTags on js and native platforms

* org_jetbrains_skia_Typeface__1nGetTableTags on js and native platforms

* Preparations for commonization of org_jetbrains_skia_Typeface__1nGetUTF32Glyphs

* getUTF32Glyphs commonized

* Minor code rearrangments in preparations of getKerningPairAdjustments commonization

* Passing count to _nGetKerningPairAdjustments

* if (glyphs.size > 0) { _nGetKerningPairAdjustments(_ptr, glyphs, glyphs.size)} else null

* Refactor org_jetbrains_skia_Typeface__1nGetKerningPairAdjustments on JVM and prepare commonization

* cross-platform actual fun String.intCodePoints()

* Uncomment yet another assertion

* Preparations for conquering family names

* Passing variations count separately

* Passing raw value to Java_org_jetbrains_skia_TypefaceKt__1nMakeClone

* org_jetbrains_skia_Typeface__1nMakeClone commonized

* org_jetbrains_skia_Typeface__1nMakeClone commonized-  fix for native

* Wrap _nMakeClone with _nMakeClone rather than wrapping param

* Fix interop scope wrapping at getUTF32Glyphs

* Finally fix org_jetbrains_skia_Typeface__1nMakeClone for js

* Narrow down familyNames to JVM test only

(I want to have a separate PR about Array<String>)

* Explicitely casting so that Typeface.cc compiles on macos and win

* Better intCodePoints() implementation for js/native

* Fix test

* Removed redundant helper array that was a trace of direct jvm port

* Get rid of yet another redundant jvm-copied idiom
Co-authored-by: Nikolay Igotti <igotti@gmail.com>

Co-authored-by: Oleksandr Karpovich <oleksandr.karpovich@jetbrains.com>

* add TextBlobTest.kt: canMakeFromPos

* add TextBlobTest.kt: getIntercepts

* add TextBlobTest.kt: tightBounds

* add TextBlobTest.kt: blockBounds

* TextBlob: make _nGetBlockBounds code reusable by moving it to commonMain/cpp

* TextBlob: make _nGetTightBounds code reusable by moving it to commonMain/cpp

* TextBlob: make _nGetPositions code reusable by moving it to commonMain/cpp

* TextBlob: make _nGetGlyphs code reusable by moving it to commonMain/cpp

* TextBlob: firstBaseline and lastBaseline

* TextBlob: _nMakeFromPosH

* TextBlob: _nMakeFromRSXform

* TextBlob: _nGetClusters

* fix jni compilation for windows: add reinterpret_cast

* update github action to upload test reports for failing tests too

* update TextBlobTest.kt

* add comment

* PR review improvements

* update epsilon

* update epsilon

* update epsilon
Co-authored-by: Oleksandr Karpovich <oleksandr.karpovich@jetbrains.com>

* Prototype for common skiko events.

* Remove rednundant braces

* Introduce StruStyleTests

* StrutStyle.fontStyle commonized

* Setting fontStyle in StrutStyle

* All withResult family of functions accept InteropScope handlers

* Introduce InteropScopeContext

* Fix after rebase

* Get rid of InteropScopeContext alias

* add TextLine.cc implementations

* enable existing TextLineTest for all targets

* PR review improvements
Co-authored-by: Oleksandr Karpovich <oleksandr.karpovich@jetbrains.com>

* Fix crash when transparent window displayed on external display on linux.

* [js] Correct external definitions for HEAP*-family - otherwise they are unreachable

* Remove definedExternally part in definitions since there's just no need in them anymore

* update Shaper jni bindings to use common ShapingOptions interop

* add Shaper shapeLine native binding impl

* add ShaperTest.canShapeLine for all targets

* add ShaperTest.canShapeTextBlob for all targets

* add PR review improvements

* remove unnecessary std::unique_ptr in Shaper.cc
Co-authored-by: Oleksandr Karpovich <oleksandr.karpovich@jetbrains.com>

* add commonMain/cpp for common bindings (js,native,jvm)

* fix typo in FontRunIterator.cc #include
Co-authored-by: Oleksandr Karpovich <oleksandr.karpovich@jetbrains.com>

* use ManagedString._nMake from toInterop(String)

* add more ManagedString tests
Co-authored-by: Oleksandr Karpovich <oleksandr.karpovich@jetbrains.com>

* Implement Animation native methods, add test

* Remove dead code

* add `fun toInterop(interopPointers: Array<InteropPointer>): InteropPointer`

* add Test: toInteropForArrayOfInteropPointers
Co-authored-by: Oleksandr Karpovich <oleksandr.karpovich@jetbrains.com>

* add Pattern and Matcher actual implementation for js and native

Update FontFeature.kt and FontVariation.kt parse functions to use regex numeric groups instead of named groups because k/n doesn't support named groups

* PR review improvements
Co-authored-by: Oleksandr Karpovich <oleksandr.karpovich@jetbrains.com>

In some packagings, Skiko (+its data file) will be placed alongside
the other shared libraries used by the JVM. This has some advantages
with respect to code signing, and also avoids a .skiko directory
littering the user's home directory. That's especially important on
Windows where dotfiles/dotdirs don't work.

The system property still overrides this.
Co-authored-by: Mike Hearn <mike@hydraulic.software>

Remove no longer needed workaround for https://bugs.chromium.org/p/skia/issues/detail?id=10929 (#330)