Turkish small caps return wrong glyph

[alerque]

When the language is set to Turkish, the small caps feature should return a dotted-i glyph for i inputs. See this example were ElstobD (v2.201) fails to do so and returns an undotted-i shape instead as in English. The second example line here is Brill, which is correct:

\begin[papersize=a7]{document}
\nofolios
\font[family=ElstobD,language=en,features=+smcp]{First Input: }
\font[family=ElstobD,language=tr,features=+smcp]{Bir İlk Ilık}

\font[family=Brill,language=en,features=+smcp]{First Input: }
\font[family=Brill,language=tr,features=+smcp]{Bir İlk Ilık}
\end{document}

[alerque]

Come to think of it, the lowercase undotted-i is also wrong. It's correctly dotless but it is not a capital shape at all, that should be mapped to the same capital a small caps i would give in English.

[psb1558]

Come to think of it, the lowercase undotted-i is also wrong. It's correctly dotless but it is not a capital shape at all, that should be mapped to the same capital a small caps i would give in English.

You mean when smcp is active? If so, I see I got the second case wrong not only here, but also in Junicode.

[alerque]

Yes, lowercase dotless-i (ı) when rendered with +smcp is not returning a small caps glyph, it is just the same one as it was. Note the 3rd letter in Ilık. Elstob is showing the regular lower case shape. The shape that should be shown is the same glyph as used in English for a capital I without a dot, but small cap sized of course.

[psb1558]

Thanks. That's what I thought, but I wanted to make sure it was straight in my head. Fixing now.

[psb1558]

I've uploaded just the variable fonts. Before I make a release, would you mind grabbing them and making sure I've got them right?

[psb1558]

I've uploaded new copies of the variable fonts after checking out the capitals with c2sc and fixing (or trying to fix) the fi and ffi ligatures. If you could check those things out, I'd be very grateful. In particular, a dotted i following f keeps its dot, and the narrow form of f is substituted. I'm pretty sure this is right:

But what should happen with dotless i? Should it be like this

or more like this (looking less like a ligature)?

[alerque]

Using the variable font build from cea8879cd05e2ded99a1fe9d4e404694467c50ff the results look correct. Updated test case using a local font file directly:

\begin[papersize=a7]{document}
\nofolios
\font[filename=./Elstob.ttf,language=en,features=+smcp]{First Input: }
\font[filename=./Elstob.ttf,language=tr,features=+smcp]{Bir İlk Ilık}

\font[family=Brill,language=en,features=+smcp]{First Input: }
\font[family=Brill,language=tr,features=+smcp]{Bir İlk Ilık}
\end{document}

[psb1558]

Okay, thanks for testing. I'll update the repository and prepare a release sometime today.