DigestPointPen questions

[justvanrossum]

I'm looking into rewriting DigestPointPen by subclassing RecordingPointPen and have some questions.

I have fixed #8, and see that Adobe uses the getDigestPointsOnly() method. That simply returns a tuple of all points, optionally sorted, without any additional info about point types of contour structure.
https://github.com/adobe-type-tools/afdko/pull/230/files#r159970100

How do other people use DigestPointPen?

I noticed that since robofab, it was enhanced with an identifier argument to beginPath(), and that is added to the digest. However, any additional keyword arguments to addPoint() are ignored, including the identifier keyword arg. The identifier argument to addComponent() is put into the digest.

• Is the inconsistent handling of identifier on purpose, or is it an oversight?
• Are identifier values even at all relevant to "digests"?
• Should the ignoreSmoothAndName flag influence the recording of identifier?

I would really love to come to some consensus about what it is that needs to be digested and for what purposes.

One last thing: the DigestPointStructurePen class doesn't do anything about the identifier argument in beginPath(), so if there is a unigue identifier, then comparing outline structures this way will fail because of that.

[typemytype]

identifier must be unique (for each glyph...) so its maybe weird to compare digests/structures with identifiers... I guess identifier could be ignore in the digests

[anthrotype]

actually I don't see the point of that getDigestPointsOnly. What's the use of a list of coordinates (even sorted by default) without information about point types and contour structure? Two completely different glyphs could well get the same digest.

And yes I agree identifiers by definition should be unique so it doesn't make sense to include in the digest.

@miguelsousa pinged me on #7 to publish fontPens on PyPI. @justvanrossum do you think the current issue #12 is a blocker for the release?

[justvanrossum]

No, definitely not a blocker.

(I don't see the point of getDigestPointsOnly either.)