Clarify glyph name ranges (2.g.i)

[PeterCon]

The current description of glyph name ranges is unclear in a few critical points.

Fixes #1222

Description

This is a proposed enhancement to the feature file spec documentation. The current
wording in section 2.g.i regarding ranges is unclear. At several points, a reader can be
left scratching their heads wondering what was meant. This is intended to provide
greater clarity in that section.

Checklist:

• [X ] I have followed the Contribution Guidelines
• I have added test code and data to prove that my code functions correctly
• I have verified that new and existing tests pass locally with my changes
• I have performed a self-review of my own code
• I have commented my code, particularly in hard-to-understand areas
• I have made corresponding changes to the documentation

[frankrolf]

The new examples added imply that glyph ranges can be given without spaces (such as A-Z).
This is no longer the case. A glyph range must be specified A - Z (with spaces), while A-Z is interpreted as a glyph name.

The reason for this change was a font editor’s decision to allow glyph names such as be-cyr.

[PeterCon]

No examples were added.

[frankrolf]

No?

[PeterCon]

Those are not (intended as) examples of usage or syntax specification but rather are simply describing the characters that are significant for this mechanism. The syntax is stated after this paragraph, and is clear about requiring spaces.

[PeterCon]

Revised the change to avoid potential confusion as suggested by @frankrolf .

[frankrolf]

Thanks!

[khaledhosny]

A-Z is interpreted as a glyph name only if the font has a glyph with that name, otherwise it is interpreted as a glyph range.

[frankrolf]

I tested my statement with a feature file that has

sub A-Z by A.sc-Z.sc;

and received the complaint “font does not contain a glyph named A-Z.“
At any rate, I think it’s better to stay consistent?

EDIT: just tried again:

makeotfexe [ERROR] <Test> Glyph "A-Z" not in font. [/Users/fgriessh/Desktop/test/features.fea 2]
makeotfexe [ERROR] <Test> Glyph "A.sc-Z.sc" not in font. [/Users/fgriessh/Desktop/test/features.fea 2]
makeotfexe [FATAL] <Test> aborting because of errors

[khaledhosny]

What I said is described in https://adobe-type-tools.github.io/afdko/OpenTypeFeatureFileSpecification.html#2fi-glyph-name:

In order disambiguate whether a text token is a range argument or a glyph name which contains a hyphen, the implementation should first assume that the token is a glyph name. If there is no glyph in the source data with a name that matches the token, then the implementation should interpret the token as a range argument.

If makeotf does not actually implement this, then text should be removed from the spec (this would also help simplifying FontTools’ feaLib implementation, which requires a font or glyph list just to disambiguate glyph names as described above).

[khaledhosny]

I tested my statement with a feature file that has

sub A-Z by A.sc-Z.sc;

and received the complaint “font does not contain a glyph named A-Z.“

Ah, yes. There is no ambiguity here as a range can’t possibly work. The range needs to be in a glyph class:

sub [A-Z] by [A.sc-Z.sc];

[frankrolf]

Thanks for the clarification! I agree allowing both notation formats creates ambiguity.

[PeterConstable]

Just noting: the issue you're discussing is orthogonal to the one my PR addresses.

[frankrolf]

That’s correct, however I think the change you made is good (thanks again!).
I will leave the acceptance of the PR up to higher powers, since I don’t want to overrule whatever concerns others may have.
(Since it’s labor day in the US today, we probably won’t get another reaction until tomorrow).
Thanks for your patience.

[josh-hadley]

👍 Thank you!