Zulip Chat Archive

Stream: mathlib4

Topic: Bundled docs


Pedro Sánchez Terraf (Jan 09 2024 at 19:40):

I was checking docs#CategoryTheory.Bundled and noticed that the instance type is rendered as c ↑s, but it would be more reasonable to display it as c α, like the sources. Is this a known issue? Or is there a reason why the current way is better?

Junyan Xu (Jan 09 2024 at 20:07):

Looks like some pretty printing / delaborator bug ... what is even s?

Kyle Miller (Jan 09 2024 at 20:33):

It seems to be that s is the name of the structure variable when the documentation is being generated? (I didn't check docgen to verify.)

The attribute [coe] α is what causes the arrow notation to be used. That's there because the α projection is the CoeSort instance, and so you'd want s.α to pretty print as ↑s.

Junyan Xu (Jan 09 2024 at 20:36):

So you think we should open an issue at https://github.com/leanprover/doc-gen4 ?

Kyle Miller (Jan 09 2024 at 20:41):

I wonder what it should do instead -- use self rather than s?

Henrik Böving (Jan 09 2024 at 20:42):

https://github.com/leanprover/doc-gen4/blob/main/DocGen4/Process/StructureInfo.lean#L23

Winston Yin (尹維晨) (Jan 09 2024 at 22:50):

Kyle Miller said:

I wonder what it should do instead -- use self rather than s?

I’ve had this confusion before. Given that s is pretty commonly used, I second this suggestion. Or colour s in a special way?

Henrik Böving (Jan 10 2024 at 08:28):

Changed it to be self

Eric Wieser (Jan 10 2024 at 11:42):

From the same page: why is the ( linked here:

image.png

Henrik Böving (Jan 10 2024 at 11:56):

Eric Wieser said:

From the same page: why is the ( linked here:

image.png

sorry, what part is linked that you don't think should be?

Henrik Böving (Jan 10 2024 at 11:56):

Oh now i see it nvm

Henrik Böving (Jan 10 2024 at 11:57):

That's not intended, probably a bug either in the delaborator or my analysis of the delaborator output.


Last updated: May 02 2025 at 03:31 UTC