[XeTeX] List of ligatures in a font

Jonathan Kew jfkthame at googlemail.com
Wed Mar 14 22:18:17 CET 2012


On 14 Mar 2012, at 20:17, Peter Dyballa wrote:

> 
> Am 14.3.2012 um 17:32 schrieb d fulano:
> 
>> But this is exactly my question:: how do I determine what are the "standard ligatures" in a font?

I presume you mean "what ligatures will be used by default when typesetting with this font" -- i.e., they don't require explicitly activating optional features, they're implemented as part of the 'liga' feature (or other on-by-default features).

The flexibility of OpenType lookups makes this a hard question to answer. Even if you exhaustively try every pair of characters, you might miss ligatures that only occur when a given pair is found _in a particular context_. And what about a ligature of more than two characters? The old OS X (AAT) version of Zapfino, for example, had a special 7-character ligature for the word "Zapfino". No amount of checking character *pairs* would detect that.

> 
> Then see which ligatures are defined by the Unicode consortium in the recent release, version 6.

Completely irrelevant. The (few) "ligature" characters encoded in Unicode - mostly for legacy-compatibility reasons - have no bearing on what ligatures may be present in any particular OpenType font.

> You can also determine the names of the glyphs in a font outside of XeTeX and check for example for the three characters "lig" or "LIG" or "Lig", or in even more variations, in their names. Likely you'll find a few more. And some can have names like "ufb03" or "afii57718".

You can determine the glyph names within xetex, even. But as there are a number of conventions for glyph names - as well as the freedom for the designer to ignore all the glyph-naming conventions! - this won't really help. Glyph names are not a reliable indicator of which glyphs are ligatures (nor of which character pairs or sequences will be mapped to those ligatures, nor whether they'd be enabled by default - which I think was a key part of the original question).

JK




More information about the XeTeX mailing list