In OT kern subtable header, the format is the high byte of 'coverage'
bit mask (bits 8-15), and the low byte (bits 0-7) is the actual coverage
bits.
In AAT kern, it's the opposite: the coverage flags are the high byte,
whereas the subtable format is the low byte.
Also adjusted the test data, and set coverage to 1 for OT kern subtable
(which means the usual horizontal kerning).
- When compiling kern subtables for version=1.0 kern tables (AAT)
the subtable header was written incorrectly: there is no version,
the length is a uint32 and there's an additional uint16 for
tupleIndex
- Use the 'coverage' low byte to select subtable "format", instead
of the 'version' field, only present in OT kern subtable header.
The getkern method was failing with AttributeError on 'unknown'
subtable formats, as their class only has 'format' instead of
'version' attribute. The 'version' attribute is renamed to
'format' also to avoid confusion, but the old one is kept for
backward compatiblity. In the only implemeted subtable class,
'format' becomes a class attribute rather than instance's
(it must always be 0).
- KernTable_format_0 now takes an 'apple=False' argument, used to
know the different headers and whether to read/write tupleIndex.
- minor pep8 whitespace and indentation fixes
- A new 'tupleIndex' attribute is written out to TTX for apple
kern subtables. Old ttx files which lack that attribute will
still be read (with a warning) and will default to tupleIndex=0
when recompiled or dumped with current fonttools.
Fixes#1089
This also fixes fonttools/fonttools/issues/1030.
Although the roundtrip is generally possible when a VariationStore is built from source font data using the Superpolator model, it is possible to build region definitions that do not follow this model. Behdad cited the Skia "Q" example, where the tail of the Q is affected by two regions defined as:
min=0 peak=0.5 max=0.51 delta=+10
min=0.49 peak=0.5 max=0.51 delta=-10
Beyond composing ligatures, AAT finite-state transducers can also
execute insertion actions without rewriting existing glyphs. The
corresponding actions have almost the same structure as ligature
actions, so we can share a lot of the plumbing within fonttools.
This renaming is in preparation of a larger change to support `morx`
ables with insertions.
Before this change, we were emitting XML with numeric values for `morx`
coverage flags. Now, we emit XML that makes more sense to human readers.
XML files from previous versions of fonttools can still be parsed.
Before this change, the following glyph class:
@Vowels = [@Vowels.lc @Vowels.uc y Y];
Would be written back as:
@Vowels = [@Vowels.lc = [a e i o u]; @Vowels.uc = [A E I O U]; y Y];
Which is clearly invalid. It seems for GlyphClass.asFea() to work
correctly here we should be using GlyphClassName not GlyphClass
for the nested classes (similar to the code at the beginning of
parse_glyphclass_()).
Since the AAT ligature subtable does not encode the number of ligature
glyphs, we need to infer this from the total structure length. We pass
this around by creating a custom sub-reader that only has the substruct
as its data. There might have been easier ways to accomplish this, but
we should anyway change the XML output for MorxSubtables to use custom
flag names, similar to what we're already doing for flags of morph actions.
Having a custom converter for MorxSubtables is in preparation for that
later XML format change.
Previously, for closed paths, we were always dropping a lineTo segment
that followed moveTo, because after reversing the contour this lineTo
would become the last segment, and in the Pen protocol a closePath
always implies a line to the fist point.
This is OK when the move point and the following lineTo oncurve point
(which becomes last after reversal) don't overlap.
However, if they do, we ended up dropping the duplicate point.
This cu2qu issue exemplify the problem (cu2qu actually uses the
ReverseContourPointPen wrapped by ufoLib's converter pens, but
fontTools' ReverseContourPen does exactly the same):
https://github.com/googlei18n/cu2qu/issues/51
With this patch, the ReverseContourPen now emits the last lineTo
when it is the same as moveTo.
We don't need to cast to int when using the round function from py23,
as this is a backport of python3's built-in round and thus it returns
an int when called with a single argument.
Cleans up GDEF varstore as well.
What's left:
- In none of the varLib.merger Mergers we handle the CaretValueFormat3. That should be done,
even though no one uses this,
- GPOS/GSUB FeatureVariations are not applied. Shouldn't be hard.
- 'rvrn' should be folded into 'ccmp' or some other default feature.
