Fallback code for openTypeOS2UnicodeRanges and #254
Conversation
|
Thanks Khaled! I'll review this tomorrow. Just one thing. You know that the fonttools' might be worth merging with your own work to avoid duplication. |
Setting some Unicode range seems to be required by some applications on Windows. I don’t have the exact details, but fonts that don’t set Unicode range and code pages fail on Windows. People who don’t like the fallaback, can always set openTypeOS2UnicodeRanges explicitly. This should have been in fontInfoData.py, but there is no way to access the ttFont there.
|
Dropped my Unicode ranges implementation and used fontTools’, they seem to be equivalent. |
Lib/ufo2ft/outlineCompiler.py
Outdated
| def _calcCodepageRanges(self): | ||
| codepageRanges = set() | ||
|
|
||
| unicodes = [i for i in self.unicodeToGlyphNameMapping.keys() if i is not None] |
There was a problem hiding this comment.
the keys in that mapping are never None, you don't need to check if i is not None. If we do checks like that elsewhere, we need to change it.
There was a problem hiding this comment.
Yes, I copied that from somewhere else in the same file.
Lib/ufo2ft/outlineCompiler.py
Outdated
| hasCP866 = 0 | ||
| hasCP869 = 0 | ||
| for code in unicodes: | ||
| if code == 0x00DE and hasAscii: |
There was a problem hiding this comment.
all this magic constant hex digits, would be nice to have comments showing what characters they refer to.
maybe you could use byteord('Þ') (defined in fontTools.misc.py23) so it's clear what these characters are.
Lib/ufo2ft/outlineCompiler.py
Outdated
| if hasCP708 and hasLineart: | ||
| codepageRanges.add(61) # Arabic; ASMO 708 | ||
|
|
||
| return list(codepageRanges) |
There was a problem hiding this comment.
it can stay a set. the intListToNum only checks if some integer is contained in that sequence
|
I think this would also need some accompanying test cases, especially since there're so many |
|
The code pages code is basically a translation of the code from FontForge (as mentioned in the commit message), I don’t what is the rationale behind it, but it seems to work. Alternatively, we can just do the same as Unicode ranges and if any character of the code page is present set the corresponding bit, but I don’t know if this will work. |
ok, I missed that. |
Direct translation of FontForge implementation, hopefully it is time tested…
|
I see you wrote the unicode literals in the code directly instead of their hex codes, that's good. However, I don't understand how come the tests on Python 2.7 did not fail with |
|
gonna fix it myself, thanks |
|
Thanks Cosimo! |
I don’t have the exact details, but fonts that don’t set Unicode range and code pages fail on Windows. With this code fonts built with fontmake without setting
openTypeOS2UnicodeRangesandopenTypeOS2UnicodeRangeswork on Windows.