shadowfacts/fonttools
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity
1,622 Commits 1 Branch 0 Tags
Commit Graph

1622 Commits

This Branch
This Branch
All Branches
Author SHA1 Message Date
Behdad Esfahbod
4cc9dd8fd4 Make make check make all 2015-04-26 16:32:56 -04:00
Behdad Esfahbod
51dc7450d9 Add badges 2015-04-26 03:07:16 -04:00
Behdad Esfahbod
3d6932549c Fix immediate decompiling of cmap subtables 
This codepath never worked, and is never exercised during normal operations.
Found by landscape.io.
 2015-04-26 03:01:25 -04:00
Behdad Esfahbod
df2906b0ad Fix a few issues found by landscape.io 2015-04-26 02:17:13 -04:00
Behdad Esfahbod
f588727554 Fixup; ouch! 2015-04-26 02:06:36 -04:00
Behdad Esfahbod
b30e12ae00 More whitespace 2015-04-26 02:01:01 -04:00
Behdad Esfahbod
bd67253118 Some more whitespace fixes from pep8 tool 2015-04-26 01:59:01 -04:00
Behdad Esfahbod
db0dc7ba32 Fix Mac East Asian encodings 
Ouch!
 2015-04-26 01:37:18 -04:00
Behdad Esfahbod
5287b52d22 Whitespace 2015-04-26 00:49:36 -04:00
Behdad Esfahbod
19f361b97c Support running run-tests.sh from other directories 
We'll see if this works with the bare /bin/sh's of TravisCI...
 2015-04-25 15:55:45 -07:00
Behdad Esfahbod
cc072eca9f Use utf_16_be instead of utf_16be 
That's the canonical name.

https://github.com/behdad/fonttools/issues/236
 2015-04-24 12:48:37 -07:00
Behdad Esfahbod
0ac019ca2c Use native mac_romanian and mac_croatian encodings 
Apparently they are implemented in Python, just miss aliases.

https://github.com/behdad/fonttools/issues/236
http://bugs.python.org/issue24043
 2015-04-24 12:32:56 -07:00
Behdad Esfahbod
4c3fbd2600 Minor 2015-04-23 16:28:30 -07:00
Behdad Esfahbod
ffd2e35371 [subset] Add note re --desubroutinize resulting in smaller WOFF/WOFF2 2015-04-22 16:02:59 -07:00
Behdad Esfahbod
2df891f25c Handle double-encoded UTF-16BE 
Merge the logic with the marked-Mac-but-UTF-16BE logic.

Fixes https://github.com/behdad/fonttools/issues/249
 2015-04-22 02:22:11 -07:00
Behdad Esfahbod
c4fb6459b0 Handle Mac-marked actual-UTF-16BE ASCII-only name 
This seems to take care of quite a few of misencoded files I've found.
 2015-04-22 02:15:51 -07:00
Behdad Esfahbod
2d443eeac2 Tighten check for ASCII in previous commit 2015-04-22 01:59:50 -07:00
Behdad Esfahbod
7e4092aba8 Implement heuristics for decoding odd-length UTF-16BE data in name table 
Part of https://github.com/behdad/fonttools/issues/249
 2015-04-22 01:55:43 -07:00
Behdad Esfahbod
abf7dc6568 If Unicode-compatible name failed to decode, write unicode="False" 2015-04-22 01:16:34 -07:00
Behdad Esfahbod
13d53f081c Minor 2015-04-21 11:08:49 -07:00
Behdad Esfahbod
d73e707317 Always write encoding as "UTF-8" in XML output 
Fixes https://github.com/behdad/fonttools/issues/246
 2015-04-21 11:07:48 -07:00
Behdad Esfahbod
801dcf21d6 Report failing modules 2015-04-21 11:06:56 -07:00
Sascha Brawer
dfe0b85b9d Test that XMLWriter keeps capitalization of passed encoding 
Unit test for #246
 2015-04-21 11:01:01 -07:00
Behdad Esfahbod
aa74114f1d Use cannonical encoding ids 2015-04-20 18:21:46 -07:00
Behdad Esfahbod
5580c60b64 Fix build 2015-04-20 18:20:04 -07:00
Behdad Esfahbod
00c89aa9cf Preserve case and other detail of XML encoding passed in 
Part of fixing https://github.com/behdad/fonttools/issues/246

Or, should I alsways write "UTF-8" for the sake of consistency?
I think I actually like doing that.
 2015-04-20 18:04:47 -07:00
Behdad Esfahbod
9d37988235 Make makeOutputFilename's overwrite argument optional 
6d7f3bf840 (commitcomment-10807720)
 2015-04-20 11:50:37 -07:00
Behdad Esfahbod
9f8742306c Minor 2015-04-20 10:00:18 -07:00
Behdad Esfahbod
95e2db7e47 Another try to fix TravisCI 
Apparently the problem was with the first shift, not with the loop:
749eb3de80 (commitcomment-10792844)
 2015-04-19 12:04:47 -07:00
Behdad Esfahbod
ecbe88cf17 Try fixing the build 2015-04-19 11:53:35 -07:00
Behdad Esfahbod
749eb3de80 Process all arguments to run-tests.sh as filters 2015-04-19 04:56:08 -07:00
Behdad Esfahbod
fc7d3b52ed Upgrade Mac East Asian encodings to native implementation if available 
Fixes https://github.com/behdad/fonttools/issues/236
 2015-04-19 04:52:25 -07:00
Behdad Esfahbod
42fbc21069 Fix x_mac Chinese names 2015-04-19 04:46:12 -07:00
Behdad Esfahbod
11b93b8491 Use canonical name of Wansung: euc_kr 2015-04-19 04:36:52 -07:00
Behdad Esfahbod
5036a39e3f Normalize encoding name for lookup 2015-04-19 04:28:22 -07:00
Behdad Esfahbod
702b94f00a Use canonical Python encoding names 2015-04-19 04:24:55 -07:00
Behdad Esfahbod
5a18fc83e6 Add getEncoding() to cmap subtables 2015-04-19 03:50:55 -07:00
Behdad Esfahbod
e561b758c1 Move getEncoding into fontTools.misc.encodingTools 
Logic developed as part of:
https://github.com/behdad/fonttools/pull/208
 2015-04-19 03:36:20 -07:00
Behdad Esfahbod
12c3f6ecd4 Docstring 2015-04-19 03:20:02 -07:00
Cosimo Lupo
1b58b828af [bezierTools] fix bug in splitLine function 
The function was returning wrong values when the two points are
vertically aligned and the split 'isHorizontal':
Fixes https://github.com/behdad/fonttools/issues/241
 2015-04-19 03:10:59 -07:00
Behdad Esfahbod
6e01a6e350 Add Roman Croatian and Romanian encodings 
Concludes https://github.com/behdad/fonttools/issues/236
 2015-04-16 18:24:07 -07:00
schriftgestalt
6d7f3bf840 [ttx] Add -f option to overwrite output file(s) 
Fixes https://github.com/behdad/fonttools/pull/208
 2015-04-16 17:55:22 -07:00
Sascha Brawer
0ec78556cb Instruct git to ignore emacs backup files 2015-04-16 17:50:39 -07:00
Behdad Esfahbod
ba0a3b9abb Revamp name table Unicode handling some more 
Part of https://github.com/behdad/fonttools/issues/236

Now we fallback to ASCII for unknown encodings.  Not sure if this might be a bad idea.
The main user-visible difference is that if there's an ASCII-only text in an unknown
encoding, we still "decode" it and use unicode="True" instead of unicode="False".

Or is assuming that any unsupported encoding is ASCII-compatible too intrusive?
 2015-04-16 17:09:49 -07:00
Behdad Esfahbod
f419f572db Add TODO 2015-04-16 16:35:51 -07:00
Sascha Brawer
b82b50cbfb Raise UnicodeDecodeError for structurally invalid names 
This will make it impossible to process fonts with super-broken
name tables. But we do not handle arbitrary broken fonts anyway,
so this is arguably better than silently ignoring junk content.
Resolves a review comment in #235.

In the unit test, replace calls of deprecated unittest.assertEquals()
by calls of unittest.assertEqual().
 2015-04-16 16:32:30 -07:00
Behdad Esfahbod
3f36721d87 Really fix python 2.6 build this time 2015-04-16 14:12:59 -07:00
Behdad Esfahbod
c9402c1685 Fix test on Python 2.6 2015-04-16 13:40:35 -07:00
Behdad Esfahbod
eefeb258ef Fix codecs end check 
There was a bug before Python 3.4 where an extra byte was included
in e.end when the error callback was called.  That hided a bug in
the code.

Fixes build with Python 3.4+
 2015-04-16 13:31:08 -07:00
Behdad Esfahbod
4ba27843a9 Fix drawing of glyf's after numpy removal! 
Should really start adding tests for these...

Fixes https://github.com/behdad/fonttools/issues/238
 2015-04-16 03:26:06 -07:00