Cascadia Code Upright bugfixes

[aaronbell]

Summary of the Pull Request

A large set of bug fixes identified while working on the Italic, but solving Github reported issues.

PR Checklist

• Closes Add Bitcoin #422 - Bitcoin added
• Closes Add replacement character FFFD #427 - FFFD glyph added
• Closes Bug: Top bar of Serbian Cyrillic capital Dje Ђ in bold is too thin #418 - top bar corrected
• Closes Bug Report - Close square bracket misformed #433 - hinting corrected to ensure alignment
• Closes Handling of short combing arrow ligatures such as >>= should be reconsidered #435 - adds consistent ligature form for =>> <<= =<< and >>= (the infinite arrows still work with addition of more equals)
• Closes Feature Request: Remove the ligratures for *) symbol #443 - ligature now ignores (*) scenario
• Closes ===> shows an artifact #454 - adds ignore to prevent equal_equal ligature from showing up
• Closes VS Code doesn't recognize "Code PL" edition of the font but it recognize the "Cascadia Code". #467 - Not specifically sure of the problem here, but suspect that it will be fixed with this - update.
• Closes 2105.24 - Regular - bar_bar_equal_end.seq interpolation issue #477 - fixed
• Closes 2105.24 - Regular - dollar.BRACKET.600 interpolation issue #478 - fixed
• Closes 2105.24 - Regular - less_less_hyphen_middle.seq interpolation issue #479 - fixed
• Closes 2105.24 - Regular - less_slash_greater.liga interpolation issue #480 - fixed
• Closes No weigths in JetBrains 2021 typography settings panel #481 - JetBrains enumerates fonts weird. I've modified the internal naming so that it will register Cascadia Code correctly. Also aligned postscript naming with Google's recommendation, so will show up as "Regular" instead of "Roman". Still works in Word!

Other Cascadia Code fixes:

• General improvement of weight balancing
• Weight of lowercase rounds reduced in the Bold weight in Cascadia Code.
• Weight of Capital stems increased in Extralight weight in Cascadia Code.
• Tweaked weight of ogonek in ExtraLight.
• Added a localized form for ij and IJ should a user chose to use those codepoints and want an accented version.
• Split fraction bar at heavier weights to improve clarity of fractions.
• Adjusted standard box drawing characters to align with GDI metrics, and added a complete set of DWrite-specific ones that align with sTypo (using rclt).
• Ironed out some tiny inconsistencies in the <$ $&gt; &lt;$&gt; ligatures which I'm sure no one will ever notice.
• Fixed centering of braces and some hyphens.
• Fixed inconsistency between semicolon/colon and period weight in bold. Also fixed slight differences in hyphen-like glyphs in bold. You're as surprised as I am.
• Increased weight of underscore in bold.
• Adjusted weighting of Ɫ.
• Changed design of commaaccent, commaaccentmod commaturnedabove and commaabove to be more distinguishable (following design of quotes).
• Increased height of βδθλ to align with the ascender height. They were too low before.
• Fixed descents of various greek lowercase glyphs that were inconsistent.
• Modified ξ weighting.
• Felt ligated, might edit later.
• Tweaked ªºⁿʷʸᶿᶻ⁰¹²³⁴⁵⁶⁷⁸⁹ in imperceptable ways.
• Corrected some additional interpolation bugs

I do enjoy giving Dustin presents.

[aaronbell]

Also, remember friends, there are more fixes to come! So If you don't see yours yet, don't worry. I've got everything on the backlog :)

[DHowett]

I do enjoy giving Dustin presents

😄 and I do very much appreciate the work you do here

[DHowett]

Closes #477 - Not specifically sure of the problem here, but suspect that it will be fixed with this - update.
Closes #477 - fixed

were these both supposed to be the same number?

[aaronbell]

Nope. :)

[DHowett]

Felt ligated, might edit later.

lol

[DHowett]

Couple followup questions (sorry: I am out of my depth, and if my questions are annoying please let me know!)

• This introduces another background directory -- should we delete or keep for now?
• The changelog says it introduces the localized ij form, but it seems to be commenting out lock_latin_2 (locl.fea) which I thought would remove it
• The changelog says it introduces rclt/stypo forms for the box drawing glyphs, but that code is also getting commented out in rclt.fea
  • It also deletes the stypo versions of the box glyphs ("boxD_oubleD_ownA_ndH_orizontal.stypo.glif
    deleted file mode 100644")

[DHowett]

UGH, Okay, ignore me. Only the background one is correct.

I was looking at the diff the wrong way. 19 years of software engineering and this is what I've done to myself.

Only the background question still stands. I realized that I thought your diff deleted them, but then that clued me in that it was backwards and it is in fact adding them.

[aaronbell]

👍 It happens! The background can be deleted. I usually clear them out of the source on my end but must have re-added some by accident.

[DHowett]

In that case, it does look like the .fea and .glif files are growing support for Hebrew and Arabic (there's a lot of them added in this PR)

[aaronbell]

Yup! All of that stuff is disabled for now. But once we have the rest of the bits ready I’ll turn it all back on.

[DHowett]

Oh this is great! Will the glyphs not even be included in the final font? This is so cool

[aaronbell]

Nope. They’ve been set as non-exporting.