Another round of (additional) analysis on libre fonts.
Hin-Tak Leung
Posts: 363
Hi,
Since FontVal 2.0 was released and I uploaded the subset of 2000 libre font reports on hinting [1] & the glyf tables [2], I hope some people have been busy fixing them. If not, "soon" would be good.
FontVal 2.0 doesn't differ much from FontVal 1.0 except much faster, on libre fonts. There are some differences on proprietary fonts. Most of the post-2.0 activities were in adding detection of use and mis-use of undocumented behaviors, quirks of specific font engines (yes, a "BUG-FOR-BUG mode...). Those don't really benefit libre fonts, libre font editors, or libre font engines. Here are two additions which do:
- Apple's version of truetype spec has a little extra sentence saying " Setting the loop variable to zero is an error". So this check is added. A surprising large number of libre fonts do that. ("fedora-2.0-vs-2.1pre-b64+5.4.diff.summary.txt" / details in "fedora-2.0-vs-2.1pre-b64+5.4.diff.txt" [3])
- E6019 - "Funit coordinate must be -16384 .. 16383". A surpising number of libre fonts also do 32767 or 65535 :-). I also found that FontVal 1.0 implemented that check wrongly - it was checking the input as F26.6 and so for [-1048576, 1048512] instead. The font list is in "fedora-2.0-vs-2.1pre-b66-E6019.summary.txt", details in "fedora-2.0-vs-2.1pre-b66-E6019.diff.txt",
- there was a small anomaly with mona-vlgothic/mona-VLGothic.ttf - it was a Pass with FontVal 1.0 but Application Error with FontVal 2.0. Turned out the font is inconsistent on how many glyphs it has! So neither was correct, and it is now
- "W6024 | Invalid Argument" with glyph ids, when trying to access the invalid glyphs. Somebody please fix the the inconsistency.
- Ray Larabie explained that the version of Larabie fonts carried on libre platforms is vastly out-of-date, due to not being able to agree on the license of later versions. Somebody please fix the old version, or try to agree on something new. :-).
The success or not of FontVal, as a libre project, depends on font problems found being addressed. I am not looking at making another release any time soon because the post-2.0 changes don't really benefit libre fonts/editors/engines. (The bare "basic" test run on 6000 fonts takes about 9 hours... I just got hold of another 33000 font reports...). Please feel free to donate nonetheless ( https://sourceforge.net/p/hp-pxl-jetready/donate/ ).
Hin-Tak
[1] https://lists.freedesktop.org/archives/create/2016-July/005218.html
[2] https://lists.freedesktop.org/archives/create/2016-August/005222.html
[3] all the summary files referred are in: http://htl10.users.sourceforge.net/tmp/FontVal-test-results-2016July/
Since FontVal 2.0 was released and I uploaded the subset of 2000 libre font reports on hinting [1] & the glyf tables [2], I hope some people have been busy fixing them. If not, "soon" would be good.
FontVal 2.0 doesn't differ much from FontVal 1.0 except much faster, on libre fonts. There are some differences on proprietary fonts. Most of the post-2.0 activities were in adding detection of use and mis-use of undocumented behaviors, quirks of specific font engines (yes, a "BUG-FOR-BUG mode...). Those don't really benefit libre fonts, libre font editors, or libre font engines. Here are two additions which do:
- Apple's version of truetype spec has a little extra sentence saying " Setting the loop variable to zero is an error". So this check is added. A surprising large number of libre fonts do that. ("fedora-2.0-vs-2.1pre-b64+5.4.diff.summary.txt" / details in "fedora-2.0-vs-2.1pre-b64+5.4.diff.txt" [3])
- E6019 - "Funit coordinate must be -16384 .. 16383". A surpising number of libre fonts also do 32767 or 65535 :-). I also found that FontVal 1.0 implemented that check wrongly - it was checking the input as F26.6 and so for [-1048576, 1048512] instead. The font list is in "fedora-2.0-vs-2.1pre-b66-E6019.summary.txt", details in "fedora-2.0-vs-2.1pre-b66-E6019.diff.txt",
- there was a small anomaly with mona-vlgothic/mona-VLGothic.ttf - it was a Pass with FontVal 1.0 but Application Error with FontVal 2.0. Turned out the font is inconsistent on how many glyphs it has! So neither was correct, and it is now
- "W6024 | Invalid Argument" with glyph ids, when trying to access the invalid glyphs. Somebody please fix the the inconsistency.
- Ray Larabie explained that the version of Larabie fonts carried on libre platforms is vastly out-of-date, due to not being able to agree on the license of later versions. Somebody please fix the old version, or try to agree on something new. :-).
The success or not of FontVal, as a libre project, depends on font problems found being addressed. I am not looking at making another release any time soon because the post-2.0 changes don't really benefit libre fonts/editors/engines. (The bare "basic" test run on 6000 fonts takes about 9 hours... I just got hold of another 33000 font reports...). Please feel free to donate nonetheless ( https://sourceforge.net/p/hp-pxl-jetready/donate/ ).
Hin-Tak
[1] https://lists.freedesktop.org/archives/create/2016-July/005218.html
[2] https://lists.freedesktop.org/archives/create/2016-August/005222.html
[3] all the summary files referred are in: http://htl10.users.sourceforge.net/tmp/FontVal-test-results-2016July/
Tagged:
0
Categories
- All Categories
- 43 Introductions
- 3.7K Typeface Design
- 803 Font Technology
- 1K Technique and Theory
- 622 Type Business
- 444 Type Design Critiques
- 542 Type Design Software
- 30 Punchcutting
- 136 Lettering and Calligraphy
- 83 Technique and Theory
- 53 Lettering Critiques
- 485 Typography
- 303 History of Typography
- 114 Education
- 68 Resources
- 499 Announcements
- 80 Events
- 105 Job Postings
- 148 Type Releases
- 165 Miscellaneous News
- 270 About TypeDrawers
- 53 TypeDrawers Announcements
- 116 Suggestions and Bug Reports