[Webkit-unassigned] [Bug 258944] New: Optical Size (opsz) axis is handled very inconsistently in variable fonts

[bugzilla-daemon at webkit.org]

https://bugs.webkit.org/show_bug.cgi?id=258944

            Bug ID: 258944
           Summary: Optical Size (opsz) axis is handled very
                    inconsistently in variable fonts
           Product: WebKit
           Version: Safari Technology Preview
          Hardware: Mac (Apple Silicon)
                OS: Unspecified
            Status: NEW
          Severity: Major
          Priority: P2
         Component: Text
          Assignee: webkit-unassigned at lists.webkit.org
          Reporter: stephen at arrowtype.com
                CC: mmaxfield at apple.com

Created attachment 466956

  --> https://bugs.webkit.org/attachment.cgi?id=466956&action=review

Reproduction repo

In Safari Technology Preview (Release 173 (Safari 17.0, WebKit 18616.1.20.2)):
- In a variable font with an Optical Size (`opsz`) axis and Weight (`wght`) axis
- If the `wght` and `opsz` axes are at their defaults (often but not always with `wght=400` and `opsz` somewhere between about `8` and `20`)

...The `opsz` that gets *rendered* is the font’s maximum `opsz`. 

This is tested in Roboto Flex, Roboto Serif, Newsreader, Fraunces, and other variable fonts with `wght` & `opsz` axes.

Reproduction repo: https://github.com/arrowtype/safari-opsz-test

Reproduction site: https://arrowtype.github.io/safari-opsz-test

![Safari Optical Size issue, Fraunces](images/safari-issue-opsz-fraunces.png)
![Safari Optical Size issue, Roboto Flex](images/safari-issue-opsz-robotoflex.png)
![Safari Optical Size issue, Newsreader](images/safari-issue-opsz-newsreader.png)

Another simple way to test this is to load such font into [Wakamai Fondue](https://wakamaifondue.com/beta/). It has type testers that are set to the font’s defaults. Slide the sliders around a little, and notice how the rendered font will “jump” in style at certain values.

Please note: in the current main Safari, there are other `opsz` inconsistencies. For example, if you go to a font’s max `opsz`, the rendered font weight will suddenly be the min `wght`. So, please test for that while fixing this, and don’t e.g. simply revert changes to opsz handling (not that I think you would do so, but there have been problems here for awhile, so please be careful but also quick, if possible, in addressing it).

Thanks so much for all your work here!

-- 
You are receiving this mail because:
You are the assignee for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.webkit.org/pipermail/webkit-unassigned/attachments/20230706/91d4a10d/attachment.htm>