shadowfacts/fonttools
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity
1,445 Commits 1 Branch 0 Tags
Commit Graph

564 Commits

Author SHA1 Message Date
Behdad Esfahbod
9f23ee4cc8 Change cmap-based duplicate-name resolution from n^2 to linear time 
Similar to 85be2e0a9773acec3c6d14c345b1fd94ab3aa5c3, though much
harder to hit.
 2014-06-02 18:09:47 -04:00
Behdad Esfahbod
4ff0d4b192 Fix post table glyph name dedup logic for edge case 
This is a followup fix to 85be2e0a9773acec3c6d14c345b1fd94ab3aa5c3
Before this change, if some glyph names had a "#-number" suffix in
the post table, we could generate duplicate glyph names.  Fix that,
even though "#" is NOT a valid character in PS glyph names.
 2014-06-02 18:03:35 -04:00
Behdad Esfahbod
62dd7b2a0e Refactor getSearchRange() 2014-05-27 16:01:47 -04:00
Behdad Esfahbod
9fed95216a Fix rangeShift calculations for count=0 in sfnt, cmap, and kern 2014-05-27 15:54:04 -04:00
Behdad Esfahbod
ec5f5150e2 Simplify cmap subtable format 4 idDelta code 2014-05-21 19:14:22 -04:00
Behdad Esfahbod
e7fb216ad8 Fix cmap subtable format 4 compiling 
Was failing all these years, if a idDelta value happened to be
exactly -0x8000...

Test case: roundtripping MTLmr3m.ttf.
 2014-05-21 19:14:03 -04:00
Behdad Esfahbod
4c777d3465 Typo 2014-05-21 17:57:14 -04:00
Behdad Esfahbod
3012076942 Minor 2014-05-19 12:43:37 -06:00
Behdad Esfahbod
ebde5454e5 Merge pull request #103 from olivierberten/post4 
'post' format 4.0 support
 2014-05-15 15:36:03 -06:00
Behdad Esfahbod
487b15fd94 Add a few __delitem__ implementations 2014-05-14 13:51:10 -06:00
Behdad Esfahbod
37b7bad5b1 Micro-optimize hmtx/vmtx loading 2014-05-14 00:30:02 -04:00
Behdad Esfahbod
d56eebfa96 Use .extend() syntax instead of += for sets 2014-05-14 00:22:14 -04:00
Behdad Esfahbod
470d610eb2 Further micro-optimize cmap subtable format 4 loading 2014-05-14 00:12:03 -04:00
Behdad Esfahbod
2db5eca0df Micro-optimize cmap subtable format 4 loading 2014-05-14 00:08:15 -04:00
Behdad Esfahbod
0d182bfb80 Speed up cmap format 4 loading 
From quadratic to linear.  Sigh.
 2014-05-14 00:00:47 -04:00
Behdad Esfahbod
da223b8fa8 Minor optimization 2014-05-09 17:05:32 -04:00
Behdad Esfahbod
497863a190 Allow disabling 'head' table modified-timestamp recalculation 
Part of:
https://github.com/behdad/fonttools/issues/115
https://github.com/behdad/fonttools/issues/46
 2014-05-01 15:13:22 -07:00
Behdad Esfahbod
4bb028e44f Fix identifierToTag for Python 3 2014-03-28 15:18:14 -07:00
Behdad Esfahbod
50d6c7298a Minor 2014-03-28 14:32:24 -07:00
Behdad Esfahbod
d0a31f5a43 Make tag parameter to table constructor optional 
If not provided, extract from class name.
 2014-03-28 14:04:01 -07:00
Behdad Esfahbod
77ccf45eb0 Add NameRecord.isUnicode() 2014-03-12 12:32:27 -07:00
Behdad Esfahbod
0f74e80d59 Add isUnicode() and isSymbol() to cmap subtables 2014-03-12 12:17:33 -07:00
Behdad Esfahbod
319643ade2 Fix up regression in promoting lookups to extensions 
Was broken by 6ed2eb409231a7fe3e0c2a135f9a89796206f639

Fixes https://github.com/behdad/fonttools/issues/110
 2014-02-20 16:31:15 -05:00
Behdad Esfahbod
6338375bd8 Allow loading zero-length tables 
Fixes https://github.com/behdad/fonttools/issues/106
 2014-01-30 16:48:11 -05:00
Olivier Berten
ec8dccb731 More array instead of pack 2014-01-28 23:45:39 +01:00
Olivier Berten
149b9f97e3 use array instead of unpack 2014-01-28 20:59:10 +01:00
Olivier Berten
4a73f8b4e9 'post' format 4.0 support 2014-01-28 14:37:15 +01:00
Behdad Esfahbod
a36c74e8b8 Add logic for keeping PUSHB/PUSHW distinction 
And avoid merging consequent pushes.  No option to control this
externally right now, but the logic is there.

Part of https://github.com/behdad/fonttools/issues/92
 2014-01-19 17:30:19 +08:00
Behdad Esfahbod
1ae29591ef from __future__ import absolute_import 
Such that our Python 2 is closer to Python 3.

Part of https://github.com/behdad/fonttools/issues/77
 2014-01-14 15:07:50 +08:00
Behdad Esfahbod
95f795f40a Move offset overflow fixing logic where it belongs 2013-12-20 21:52:28 -05:00
Behdad Esfahbod
b76d6ff14a [merge] Remove BaseTable.__hash__ need 2013-12-20 20:24:27 -05:00
Behdad Esfahbod
398770d51c [merge] Map GSUB/GPOS feature / lookup indices properly 2013-12-19 15:30:24 -05:00
Behdad Esfahbod
9c5e2ce1b6 Add a get() method to TTFont ala dict.get() 2013-12-19 11:38:56 -05:00
Behdad Esfahbod
ee6340f475 Only track one parent per Writer 
We were not using the extra parents anyway.
 2013-12-17 06:31:37 -05:00
Behdad Esfahbod
ab0ca1bd6b Minor simplification 2013-12-17 06:19:15 -05:00
Behdad Esfahbod
6188b5b5f0 Remove unused variable 2013-12-17 06:11:41 -05:00
Behdad Esfahbod
dea08f2ba1 Minor optimization 2013-12-17 06:04:28 -05:00
Behdad Esfahbod
f6502635b9 Minor 2013-12-17 05:59:05 -05:00
Behdad Esfahbod
9e482332ba Fix subtable-split logic 
Though, the split logic isn't implemented for GPOS, so I have not
been successful in actually exercising it so far.
 2013-12-17 05:46:51 -05:00
Behdad Esfahbod
42d305a824 Further adjustment 2013-12-17 05:10:30 -05:00
Behdad Esfahbod
fcc56e637a Choose most compact representation when writing PUSH instructions 
Can use some testing...
 2013-12-17 05:06:37 -05:00
Behdad Esfahbod
965c87fb4a When disassembling, merge multiple PUSH ops 
Fonts typically have multiple consecutive PUSHB / PUSHW
operators.  Merge them in the disassembled output.

This, for now, generates worse assmebly because all PUSHB's
grouped with PUSHW's now generate as PUSHW.  Followup fix
will address that.
 2013-12-17 04:19:14 -05:00
Behdad Esfahbod
ce84d3215e Generate PUSH mnemonic for PUSHB/NPUSHB/PUSHW/NPUSHW 
Reduces XML noise while reenabling the optimization
during assembly.

Fixes https://github.com/behdad/fonttools/issues/73
 2013-12-17 04:02:10 -05:00
Behdad Esfahbod
3714c78c92 When writing PUSH instructions, obey requested opcode 
Previously fonttools was choosing between the optimal opcode
from PUSHB, NPUSHB, PUSHW, and NPUSHW.  We now respect
whatever was requested and err if the format doesn't support
the data.

Morevoer, if the number of numbers to push is more than 255,
we add multiple push instructions to push all the numbers.

Finally, add a new pseudo-opcode "PUSH" that automatically
chooses the best format.

This, by itself, reduces roundtrip noise by not optimizing
the bytecode as it was before.  In a followup commit I'll
change the bytecode disassembler to always produce PUSH
in the textual instructions instead of the four variants.
That way, we get both the optimization during assembling,
and reduced noise in XML.

Part of https://github.com/behdad/fonttools/issues/73
 2013-12-17 04:00:01 -05:00
Behdad Esfahbod
e93e29125f Don't write Format for Coverage, ClassDef, Single/Alternate/LigatureSubst 
The format values for those are automatically handled in
postRead/preWrite to choose optimal format.  As such, don't write them
in XML.  Reduces noise.

Part of https://github.com/behdad/fonttools/issues/73
 2013-12-17 03:06:10 -05:00
Behdad Esfahbod
6ed2eb4092 Write LookupType as a comment 
We compute LookupType from the type of the subtables, so write it out
as an informative comment in XML as opposed to an element.
 2013-12-17 02:57:29 -05:00
Behdad Esfahbod
5fec22bfcd Compute correct LookupType when compiling font 
Also ensures that all subtables of a lookup have the same
lookup type.
 2013-12-17 02:42:18 -05:00
Behdad Esfahbod
ea82d6dfd1 Handle NULL coverage table in ligature substitute 
Just to be in line with Single and Alternate subst.
 2013-12-17 01:58:26 -05:00
Behdad Esfahbod
dafdb2933a Further optimize table loading 
Again, both faster and more memory efficient
 2013-12-17 00:58:02 -05:00
Behdad Esfahbod
b4070bd629 Minor optimization of table loading 
Is both faster and more memory efficient.
 2013-12-17 00:44:33 -05:00