From: Frédéric WANG <fred.wang@free.fr>

Date: Wed, 23 Sep 2015 23:16:41 +0200

Message-ID: <560316B9.3060603@free.fr>

To: Murray Sargent <murrays@exchange.microsoft.com>, Mozilla Math Developers <dev-tech-mathml@lists.mozilla.org>, "www-math@w3.org" <www-math@w3.org>

Date: Wed, 23 Sep 2015 23:16:41 +0200

Message-ID: <560316B9.3060603@free.fr>

To: Murray Sargent <murrays@exchange.microsoft.com>, Mozilla Math Developers <dev-tech-mathml@lists.mozilla.org>, "www-math@w3.org" <www-math@w3.org>

Thank you for the suggestion and reference Murray. Actually, the "check whether there is a MATH table" workaround has already been used since Gecko 31, because many OpenType fonts with a MATH table didn't set the OS/2 USE_TYPO_METRICS flag properly. However, it's not really good to keep such "hacks" so this will be removed starting with Gecko 43 (to be released at the end of the year): https://bugzilla.mozilla.org/show_bug.cgi?id=1170782 From my testing of free math fonts and from what I heard (for non-free math fonts), all the fonts with a MATH table but STIX sets the OS/2 USE_TYPO_METRICS flag. That's not much a problem for STIX since the current release has bugs that prevent it to be usable for math layout anyway and it's best to use the XITS fork. The STI Pub is aware of the bugs and I'm confident that at least the OS/2 USE_TYPO_METRICS bug will be fixed in version 2. On 09/23/2015 10:58 PM, Murray Sargent wrote: > Frédéric noted that the STIX font doesn't support the OS/2 USE_TYPO_METRICS flag. > > For math fonts, a workaround is to check if there's a MATH table and use the typo metrics if so. Another work around is to compare the type ascent/descent to the TEXTMETRIC ascent/descent and assume a high font is involved if the latter is significantly bigger (1.5 × or more). Here's a post (http://blogs.msdn.com/b/murrays/archive/2009/12/01/high-fonts-and-math-fonts.aspx#10023606) on high fonts which explains a little more. > > Murray -- Frédéric Wang maths-informatique-jeux.com/blog/fredericReceived on Wednesday, 23 September 2015 21:17:18 UTC

*
This archive was generated by hypermail 2.3.1
: Wednesday, 23 September 2015 21:17:18 UTC
*