HarfBuzz-NG does character mirroring where appropriate.
A simple solution is to unset RightToLeft shaper flag when
text gets shaped with HB-NG. Instead, move the mirroring code
right to HB-old proxy function and decrease code duplication.
Change-Id: Icdcd50b73b3e6a43da4b85addc7d8f51edf86512
Reviewed-by: Lars Knoll <lars.knoll@digia.com>
In UTF-16, SMP code point is encoded with a pair of surrogates,
thus making SMP code point twice wider than BMP code point.
Change-Id: Ifb0e78cff50cca7231d07097a730188bd98c54cb
Reviewed-by: Lars Knoll <lars.knoll@digia.com>
...and thus consume 4 bytes less per glyph and increase the performance a bit.
It seems, the only CTFontGetAdvancesForGlyphs() returns both x and y
advances, though y advances are always equal to 0 for horizontal
orientation and x advances are always equal to 0 for vertical orientation.
Also, rename `advances_x` to `advances` for consistency
and declare QGlyphLayout's data size in a single place.
Change-Id: I56b20f893f8a6feb7aa870e3edbca99dd93ba2e2
Reviewed-by: Eskil Abrahamsen Blomfeldt <eskil.abrahamsen-blomfeldt@digia.com>
Reviewed-by: Lars Knoll <lars.knoll@digia.com>
Reviewed-by: Simon Hausmann <simon.hausmann@digia.com>
This didn't work for a loooong time already;
not it is possible to re-introduce this optimization once again.
Change-Id: I35c40e68933227bda6cbc038c2d8f6709db62451
Reviewed-by: Lars Knoll <lars.knoll@digia.com>
The HB face caching mechanism introduced in 227e9a40cf
wasn't complete due that fact that HB-NG doesn't parse the entire
font table at once but rather references a table on-demand.
This incompleteness caused a crash in case the engine doesn't
get cached or when it removed from the cache and then re-used.
Task-number: QTBUG-36099
Change-Id: I7816836107655ce7cf6eb9683bb5dc7f892f9cd1
Reviewed-by: Lisandro Damián Nicanor Pérez Meyer <perezmeyer@gmail.com>
Reviewed-by: Michael Krasnyk <michael.krasnyk@gmail.com>
Reviewed-by: Lars Knoll <lars.knoll@digia.com>
Reviewed-by: Allan Sandfeld Jensen <allan.jensen@digia.com>
This reverts commit 65b12fbdb1.
QFontEngine is not always loaded from QFontDatabase, resulting in
the flag not being set.
Change-Id: I39bc5bd4a8dea153d191cfc55f4324195f75f64c
Reviewed-by: Konstantin Ritt <ritt.ks@gmail.com>
Reviewed-by: Mitch Curtis <mitch.curtis@digia.com>
Reviewed-by: Lars Knoll <lars.knoll@digia.com>
This is an enabler to fallback to native font rendering when using a
bitmap font in Qt Quick.
Task-number: QTBUG-32737
Change-Id: I6d841dd5ef54d78a00f7fab9d80e9c95ff7f7b98
Reviewed-by: Eskil Abrahamsen Blomfeldt <eskil.abrahamsen-blomfeldt@digia.com>
This is to work around a bug in GCC 4.7's link-time optimiser. Without
it, linking QtOpenGL and QtPrintSupport fails because the compiler
generates the code for the virtuals, including the virtual table, but
not the methods.
Change-Id: Idcecd51c32e6228e0eefe67f1bd630ea1f5d5da7
Reviewed-by: Konstantin Ritt <ritt.ks@gmail.com>
Move scriptRequiresOpenType() body right into QFontEngine::supportsScript(),
thus centralizing use of this performance cheat.
Change-Id: I5f494b086f8f900b631c491f41e9cb800002c0f6
Reviewed-by: Lars Knoll <lars.knoll@digia.com>
Whilst having the objectName set for each engine is
somewhat handy when debugging, deriving from QObject just for that
is a wasting of memory in all other cases.
This also broke the font engine abstraction by allowing qobject_cast()
to access some private data; the only sane way to distinguish
engines is querying their Type value.
Change-Id: Ib1d195692859eb39089f6d8d9016cb8f9dcc0400
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@digia.com>
Reviewed-by: Lars Knoll <lars.knoll@digia.com>
Instead of loading HB face in QFreetypeFace::getFace(), defer allocation
until the first FT font engine for that face gets initialized;
then, QFreetypeFace "reparents" and manages the loaded HB face.
Change-Id: I2ac8ead4c6ed25d341af9c9cf0c34dfb979f8390
Reviewed-by: Eskil Abrahamsen Blomfeldt <eskil.abrahamsen-blomfeldt@digia.com>
QFontEngineMulti::stringToCMap() stores the fallback engine index
in a glyph index'es high byte, which means the maximum fallback engine index
it can store is 255, so limit the number of tries we're doing to this value.
Otherwise we could end up with `fontEngineMulti->engine(glyph >> 24) == 0`
after successful stringToCMap() call.
Task-number: QTBUG-30412
Change-Id: I06907a39186fd207f3ce4b732a1a54e615744082
Reviewed-by: Lars Knoll <lars.knoll@digia.com>
...and operate directly on a passed array pointers,
without having to copy the input and output data.
Calling QFontEngine:: stringToCMap()/recalcAdvances() with
partially-constructed QGlyphLayout is completely safe
iff `glyphs` and `advances_?` members were initialized.
Also get rid of QGlyphLayoutInstance that was used just in a single place
and never was really needed.
Change-Id: I48fab246fd69fc869f948220a553c3574d93c772
Reviewed-by: Lars Knoll <lars.knoll@digia.com>
This caused glitches up to unreadable text with i.e.
pixelSize 256 and stretch factor 4x /* ((256*4)<<16)<<6 */.
Change-Id: Ib6a038a043d820a94bd2019c50390a815a2a8277
Reviewed-by: Konstantin Ritt <ritt.ks@gmail.com>
We're treating fontDef.pixelSize as a box height everywhere but in
harfbuzzFont initialization code the logic was opposite.
Change-Id: I4eecf7861a6084c7f2dbe1df85a71f2e251da845
Reviewed-by: Eskil Abrahamsen Blomfeldt <eskil.abrahamsen-blomfeldt@digia.com>
...by updating HB_Face::isSymbolFont value to make HB know that the font is a symbol font.
U+00AD is a Unicode soft hyphen character, which shouldn't be threated
like a default ignorable (invisible) when 0x00AD occurred in a symbol font.
Change-Id: Idd4760aded644f7f11a1a8d3c03c348abc69a3e4
Reviewed-by: Eskil Abrahamsen Blomfeldt <eskil.abrahamsen-blomfeldt@digia.com>
1. when there were some engines with ref > 1 in the cache, prior to calling
QFontDatabase::{add,remove}ApplicationFont()/removeAllApplicationFonts()
(QFontCache::clear() has never decreased engine's cache_count);
2. when the QFontEngineData's engine is not in cache i.e. the Box or Test font engine
(~QFontEngineData() didn't free engines it keeps).
Instead of using the font engine's (external) "cache_count" counter,
QFontCache now references a given font engine every time it is inserted to
the cache and dereferences exactly that number of times in clear().
Change-Id: I87677ebd24c1f4a81a53526f2e726e596b043c61
Reviewed-by: Lars Knoll <lars.knoll@digia.com>
This affects HB_Font and HB_Face.
As of now, the Hurfbuzz API usage is concentrated in qfontengine(|_ft).cpp
and qtextengine.cpp, thus it is a lot easier to switch to Hurfbuzz-NG.
Change-Id: Ie06959efd5d6080fe44c407d9f5de0a07dd1c210
Reviewed-by: Lars Knoll <lars.knoll@digia.com>
Reviewed-by: Josh Faust <jfaust@suitabletech.com>
HB_Face's supported_scripts[] expects HB_Script, so QChar::Script should
be remapped via script_to_hbscript().
Change-Id: Ib068c35ab76567fe9a61da7d8ab01133a6f58bc0
Reviewed-by: Lars Knoll <lars.knoll@digia.com>
A new glyph type is added to the glyph caches for ARGB bitmap glyphs,
and the raster and OpenGL paint engines have been modified to support
this glyph type for drawCachedGlyphs().
The CoreText font engine implements support for these glyphs through
the CTFontDrawGlyphs API, since CGContextShowGlyphsWithAdvances does
not handle color glyphs.
Change-Id: Idad9ce75a911cae130d65aebe59142772a16fc12
Reviewed-by: Eskil Abrahamsen Blomfeldt <eskil.abrahamsen-blomfeldt@digia.com>
Change 8927084d0a made a mistake in
reversing the boolean logic of the type >= TxProject logic,
causing us to try to use the font engine for perspective
transformations instead of falling back to QPainterPath.
Change-Id: Ideb59751ace23ab83f8ebd4f02dbe6c1724644a5
Reviewed-by: aavit <eirik.aavitsland@digia.com>
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@digia.com>
Some cruft had built up over time, and this is an attempt at cleaning up
the naming and use of these functions, and should not have any behavioral
effects.
The function supportsTransformations() has been renamed in QPaintEngineEx
to reflect its use, which is to decide if QPainter needs to pre-transform
the coordinates of the static text before asking the paint-engine to draw
it. The new name is requiresPretransformedGlyphPositions().
The OpenGL and CoreGraphics (Mac) paint engines keep their behavior of
not needing pre-transformed text, while the raster engine needs this
when using cached glyphs. The base-class implementation assumes that
all transforms that include a projection will need pre-transform,
which is also the case for the raster engine.
All decisions in the paint engines about whether or not to use the
glyph cache when drawing text are now deferred to the function
shouldDrawCachedGlyphs(), which has been refactored for the GL paint
engine(s) to share more logic. All implementations call the base
class implementation, which ensures that large font sizes will not
be cached. The raster engine will in addition ask the font engine
whether or not it can produce glyphs for the glyph-cache with the
given transform.
This is the only remaining instance of the supportsTransformations()
function, and will for all font engines except the CoreText engine
support affine transformations. The CoreText engine on the other hand
only supports translations (for now).
Change-Id: I8fb5e43e3de3ef62a526a79a6dfeda7f9546771d
Reviewed-by: Eskil Abrahamsen Blomfeldt <eskil.abrahamsen-blomfeldt@digia.com>
If (*nglyphs < len), stringToCMap() sets *nglyphs to len and returns
false immediately; the caller then must resize the buffer and re-try.
However, QFontEngineMulti::stringToCMap() doesn't update the nglyphs value
and thus the second call would fail, too. This is quite unexpected.
Change-Id: Id2cce7b9faf7706c382fccf023e1b7affa9a10be
Reviewed-by: Lars Knoll <lars.knoll@digia.com>
Reviewed-by: Eskil Abrahamsen Blomfeldt <eskil.abrahamsen-blomfeldt@digia.com>
In change 1582407fc7, the Q_WS_MAC
code path which disabled drawing cached glyphs for any transform
was removed, as was the comment that scaling and rotation wasn't
supported by the Mac font engines. This obviously broke transformed
text on Mac, so we need to put it back.
I put it into the font engine itself where it belongs, and I kept
the somewhat confusing naming convention which is used in the
paint engine to minimize this patch. I'll clean up these function
names in a future commit.
Task-number: QTBUG-27362
Change-Id: I4fc6a503eedd4b1ebaf3ee659d948f997f433338
Reviewed-by: Samuel Rødal <samuel.rodal@digia.com>
in case when the layout is partially initialized.
We shouldn't access any data except of indices
if GlyphIndicesOnly flag has been passed in.
Change-Id: I264689b498e0f9de8b5c040d47dbae4f6ef391c4
Reviewed-by: Eskil Abrahamsen Blomfeldt <eskil.abrahamsen-blomfeldt@digia.com>
Change copyrights and license headers from Nokia to Digia
Change-Id: If1cc974286d29fd01ec6c19dd4719a67f4c3f00e
Reviewed-by: Lars Knoll <lars.knoll@digia.com>
Reviewed-by: Sergio Ahumada <sergio.ahumada@digia.com>
When accessing a fallback engine, we always need to call
ensureEngineAt() to make sure it's loaded.
Change-Id: Ib27e34137cfe8a3dd2b358aef3b3296a4ca52478
Reviewed-by: Jiang Jiang <jiang.jiang@nokia.com>
by avoiding the glyph metrics calculation when those metrics weren't requested
(QFontEngine::GlyphIndicesOnly flag has been passed).
As a side effect, this fixes a crash in case QGlyphLayout
was initialized only partially (with the glyphs array and the size).
Change-Id: I7d67abc2a74683131361fa21f8be203f61f247bc
Reviewed-by: Konstantin Ritt <ritt.ks@gmail.com>
These flags are specific to font engine(s) and has nothing
to do with the text engine or the text layout.
Change-Id: I4bb793c3c634b3cf0ae0a8a8c23b946fad5874b6
Reviewed-by: Konstantin Ritt <ritt.ks@gmail.com>
This is the approach introduced together with the WebKit "backdoor"
that lets us use QRawFont with QTextLayout. We could use it all
over the place and share more code.
Change-Id: Ie1963679755f37ba9204d67554a163f1d1115604
Reviewed-by: Eskil Abrahamsen Blomfeldt <eskil.abrahamsen-blomfeldt@nokia.com>
Required changes for using shared graphics cache for distance field
raster glyph rendering. Most of the logic is in platform plugins.
Platform plugins should implement
QPlatformIntegration::createImagePaintEngine() to create a subclass
of QRasterEngine.
Change-Id: Icf0a396e722e43b4caa2c1849aae38753cde38f1
Reviewed-by: Eskil Abrahamsen Blomfeldt <eskil.abrahamsen-blomfeldt@nokia.com>
This logic was introduced as part of change I92dfb39289a359f49caa02c2caf8baf66098fb59
but isn't used anymore.
Change-Id: I5bcfea99a7a2993434e1e978195a70dae52d6cfa
Reviewed-by: Eskil Abrahamsen Blomfeldt <eskil.abrahamsen-blomfeldt@nokia.com>
The mechanism in fontconfig which determines if a certain character
is available (FcCharSetHasChar()) may give false positives, in which
case we would load and unload those fonts per every char for which
FC gave us a false positive. This was a major performance regression.
Specifically the false positives happened when looking at e.g.
italic variants of certain multilingual fonts, since we only check
the charset of the font family as a whole and not of the specific variant,
which may only support a subset of the chars.
To optimize this, we remove the deletion of the font engines after
loading them, but also wait with loading the opentype tables until
they are actually needed. This means that for the false positives,
we will load the font, but the cached data for each unused font will
be much smaller.
Change-Id: Idfc794401a2080da5946bf65204eb947aeb635ed
Reviewed-by: Lars Knoll <lars.knoll@nokia.com>
QFontEngine::grayPalette() mistakenly returns an
empty vector, so even if there's a user (the class
it exported, after all), it cannot correctly use it.
So, just remove it.
Change-Id: Id5f70139e5f6ed9a2a3f28400cd0750d14cb7dc0
Reviewed-by: Olivier Goffart <ogoffart@woboq.com>
Reviewed-by: Jiang Jiang <jiang.jiang@nokia.com>
When we request glyphs from fallback fonts, we would potentially
load all fonts on the system into memory. This is especially true
for glyphs that are not supported by any font (or by the last in
the list) in any "Common" script (which e.g. includes CJK).
This would make any application which tried to display unsupported
glyphs use huge amounts of memory for keeping unused fonts cached,
only limited by the number of fonts on the system.
The patch contains two solutions: First, before loading the font,
the multi font engine will be asked whether it needs to be tried
for the given character. By default, this will always be true, so
all fonts will be tried, but with the new font config multi engine
in the platform plugin, it will ask FontConfig whether the font
contains a glyph for the character.
Should the font be loaded and still fail to resolve the character
(which could be the case for other platforms), we will simply
delete it again immediately instead keeping it cached.
Change-Id: I92dfb39289a359f49caa02c2caf8baf66098fb59
Reviewed-by: Lars Knoll <lars.knoll@nokia.com>
As in the past, to avoid rewriting various autotests that contain
line-number information, an extra blank line has been inserted at the
end of the license text to ensure that this commit does not change the
total number of lines in the license header.
Change-Id: I311e001373776812699d6efc045b5f742890c689
Reviewed-by: Rohan McGovern <rohan.mcgovern@nokia.com>
glyphMargin() support for QTextureGlyphCache is implemented in
respective font engines, thus this function is platform dependent.
Before Qt 5 the code is guarded in macros like #ifdef Q_WS_MAC,
now we should move them into QFontEngine and its subclasses.
So far only Windows font engines support it. FreeType and Core Text
based font engines all ignore it.
Change-Id: Ia14016533d8fbfaacf848a7d3bc928f8197318f5
Reviewed-by: Eskil Abrahamsen Blomfeldt <eskil.abrahamsen-blomfeldt@nokia.com>
Move subpixelPositionForX from QTextureGlyphCache to QFontEngine,
since some font engines like QFontEngineFT may need a custom
implementation or tweak it a little bit.
In QRasterPaintEngine::drawCachedGlyphs, do not add aliasedCoodinate
to x offset as that will break subpixel positioning.
Change-Id: Idbcec617509459b80965220ceb07b17737649bbf
Reviewed-by: Eskil Abrahamsen Blomfeldt <eskil.abrahamsen-blomfeldt@nokia.com>
Frederik Gladhorn
Konstantin Ritt
Yoann Lopes
Gunnar Sletta
Thiago Macieira
Giuseppe D'Angelo
Tor Arne Vestbø
Sergio Ahumada
Eskil Abrahamsen Blomfeldt
Eskil Abrahamsen Blomfeldt
Iikka Eklund
Pierre Rossi
Jiang Jiang
Marc Mutz
Jason McDonald
Bradley T. Hughes