[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [ft-devel] Another round of (additional) analysis on libre fonts.
From: |
Hin-Tak Leung |
Subject: |
Re: [ft-devel] Another round of (additional) analysis on libre fonts. |
Date: |
Thu, 10 Nov 2016 09:37:06 +0000 (UTC) |
--------------------------------------------
On Thu, 10/11/16, address@hidden <address@hidden> wrote:
> >
- 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.
> Sorry to say, but this Apple
addition is nonsense. Allowing a zero
loop
variable is *extremely* useful for programmatically
created
bytecode. For example, hinting
created by ttfautohint uses that all
the
time. It doesn't do any harm (the loop is simply
skipped), and it
works with all hinting
engines used today, AFAIK. I strongly
recommend to disable this warning.
Apple's version of the truetype spec has that extra sentence. I assume it is
later since the Microsoft word version linked from Microsoft's web site is 2008
and in a few other ways seems briefer/older. Since I am not making a release
any time soon either way, I guess it is just a note for people who might be
using git HEAD [which is b66] to ignore it. (and there is one other thing I
like to not be in git HEAD also - some noise about twilight zone - please
ignore much of those too). b66 was pushed out to github untested. As I said, it
takes about 9 hours to run the basic single-size test... and then much human
time to compare it with the last 6000 reports... FontVal 2.0 went through about
3 or 4 rounds of such from b51 onwards to b54.
Hin-Tak