[Fira Sans] Greek Cypriot consonants ςͅ ζͅ ξͅ ψͅ unacceptable in italics

[disk0x]

Hello. Fira Sans already provides surprisingly good support for the combining diacritics used in literary/academic typesettings of Cypriot Greek. (Fira Mono sadly does not work with those characters at all)

The italics can be improved as it regards the combinations of U+0345 COMBINING GREEK YPOGEGRAMMENI (subscript iota) with ς, ζ, ψ, and ξ.

This is how the characters render in Fira Sans 4.202. Green is acceptable (even if not perfect), Red is unacceptable.

Here's a reference in Serif and Sans

The research that proposes those characters

[carrois]

Hi, thanks for reporting. We’ll fix it, looks like a simple anchor issue. 4.3 and a new multiscript extension is close to release.

Green is acceptable (even if not perfect), Red is unacceptable.

Do you mean the /xi (03BE) or anything else?

[disk0x]

Thanks for looking into it!

Do you mean the /xi (03BE) or anything else?

A few very minor issues:

• ξ (03BE) and Ξ (039E) combined with breve (0306) or caron (030C) look slightly off-centre to me
• All characters with a descenting part (ς ζ ψ ξ) plus subscript-iota (0345) should have a shortened descenting part (see image from Carlito Enalea), so that in total, they have the same height as the characters without the iota-subscript. This is to avoid line-spacing issues such as: where ζͅ infringes into ξ (and makes it look like ξ̆ (xi + breve). Or clash with the acute accent

The second one might be solved in a different way than Carlito Enalea if it doesn't match the design philosophy of Fira Sans.

[carrois]

Hi.

Thanks for the detailed feedback. We will care about that ASAP and then get back to you if still questions occur.

As we launch our new companies website these days we are not able to update anything quite soon. We hope for your understanding. Keep on reporting!

Cheers your bBox Type team.

Am 12.02.2018 um 09:51 schrieb disk0x notifications@github.com:

Thanks for looking into it!

Do you mean the /xi (03BE) or anything else?

A few very minor issues:

ξ (03BE) and Ξ (039E) combined with breve (0306) or caron (030C) look slightly off-centre to me All characters with a descenting part (ς ζ ψ ξ) plus subscript-iota (0345) should have a shortened descenting part (see image from Carlito Enalea), so that in total, they have the same height as the characters without the iota-subscript. This is to avoid line-spacing issues such as: https://camo.githubusercontent.com/74564ea20eb44f15df0eb2f5bb1103dd1c6b9b65/687474703a2f2f692e696d6775722e636f6d2f775876706377642e706e67 where ζͅ infringes into ξ (and makes it look like ξ̆ (xi + breve). The second one might be solved in a different way than Carlito Enalea if it doesn't match the design philosophy of Fira Sans.

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/carrois/Fira/issues/24#issuecomment-364860640, or mute the thread https://github.com/notifications/unsubscribe-auth/AHcTKsXj7MzllDeM-zE6pmSykbp2F-yXks5tT_t-gaJpZM4SA-gq.

[carrois]

Should be fixed with 4.3 / resp. FiraGO. Can you prove?

[disk0x]

I purged my OS-provided Fira 4.2, copied /Fira_Sans_OTF_4300/ in my local font file, and forcibly regenerated fontconfig's cache with fc-cache -r. The renderings below are from LibreOffice 5 and Firefox 58.

Fira Sans Normal seems to have gotten worse

Fira Sans Condensed and Compressed are working

The scenario of clashing descenders and accents still exists even in Cond. and Compr.

The Fira Sans Normal likes particularly weird for me, so give me some time to also try it on other computers and font rendering backends and report back.

[carrois]

Hi. Thanks for investigating. This is strange, we simply added anchors in concerned glyphs and it should work out quite fine. Also that it only affects the Normal width in your environment is odd because all the fonts come from one master file and we added the same anchors to all masters. So at least problems should occur in all widths. Looks like it's really something like cache etc., let’s say local reasons?!

In my text-Editor it looks ok:

Please report if you have further information.

For the clashing issue we didn’t go into detail. This affects a whole lot of other languages as well (the ones using subscript and superscript accents at the same time) – and it would be a huge effort to optimize every single accented descender glyph. Hope you understand!

[disk0x]

Hello, sorry for the delay.

I can officially confirm that all problems are fixed, including in the Fira Sans Normal variant.

My package manager lied to me about deleting 4.2, it was still in /usr/share/fonts/ but the package was removed, that's why I had problems with Normal.

[carrois]

yey