You can clone with
HTTPS or Subversion.
Don’t bather with shortening the name table
After we git rid of the full OFL text it only saves less than 2 KB in
WOFF and WOFF 2, not worth it.
Update copyright and license text
* Make sure copyright year is updated at build time
* Use license text from , short and concise, the full text is
available on the web or in the font tarball.
* Clean updating the version while at it.
1. adobe-fonts/source-serif-pro#3 (comment)
Build with Python 2.7
Some installations default to Python 3, but we are not there yet, so use
No need for that
Alway set a script in the tests
Otherwise we are essentially testing the behaviour of
hb_buffer_guess_segment_properties(), which is pointless.
Print more info about the failed test
Support Python 3
Drop the unneeded extra directory level
This file is gone now
Link to the Arabic page here
Remove the obsolete README.txt
The relevant bits are merged into README.md.
Fix markdown syntax
Move the plain text docs to the top directory
That is a remnant of the ‘open font template’ thing that no one else was
fool enough to follow.
Oraganise this file a bit
Improve test output a bit
Use HarfBuzz Python support
The tests are much faster now, but requires HarfBuzz from master branch.
Don’t use installed Amiri when building the manual
A bit of consistency
Amiri moved to main Google Fonts
And Piwik is long gone from Source Forge.
fix tools/runtest.py to make the tests pass again
This fix makes use of the positional [TEXT] argument of hb-shape:
`hb-shape [OPTION...] [FONT-FILE] [TEXT]`
Also it encodes the hb-shape text argument as utf-8 if it is a unicode-string.
That -e is not needed, it seems
This check is not needed anymore
Sorts Mill editor takes care of ignoring non-existent glyphs in features
Use the new generate/mergeFeatureString functions
More TODO items
Support building WOFF2 fonts
Fix bug tracker link
Thanks to Mohammed Assar.
Rename to match other make*.py scripts
Use NBSP where appropriate