1187682 - Use ScriptPercentScaleDown and ScriptScriptPercentScaleDown constants instead of scriptsizemultiplier

Status: RESOLVED FIXED

(Core :: MathML, defect)

Description

[Frédéric Wang (:fredw)]

This might be a bit controversial, but I open this for the record.

The MathML spec says that the font-size scaling factor when incrementing the script level is scriptsizemultiplier, with default value 0.71

The OpenType MATH table contains two scaling values for script levels 1 and 2, that we currently do not use:
- ScriptPercentScaleDown Percentage of scaling down for script level 1. Suggested value: 80%.
- ScriptScriptPercentScaleDown Percentage of scaling down for script level 2 (ScriptScript). Suggested value: 60%.

Currently, we always use scriptsizemultiplier but I wonder whether in some situation we should prefer the OpenType MATH constants as they are closer the the font designer's intention.

For Latin Modern Math, we have ScriptPercentScaleDown = 70% and ScriptScriptPercentScaleDown = 50%. The spec does not seem to say whether these percentages are relative to the base size or the current size. If this is the former then the MathML values .71 and .71^2=0.5041 should already be close to the values from the OpenType MATH table, so the difference will not be significant.

Comment 1

[Frédéric Wang (:fredw)]

(Just mentioning: I opened this after Wikipedia's issue https://phabricator.wikimedia.org/T106890#1483191)

Khaled: Do you know more about the interpretation of ScriptScriptPercentScaleDown (i.e. is it relative to scriptlevel=0 size or to the scriptlevel=1 size?)

Comment 2

[Khaled Hosny]

I think both values should be relative to the base size, that is how I interpreted them and it seems to be what MS implementation is doing IIRC.

Comment 3

[Frédéric Wang (:fredw)]

(In reply to Khaled Hosny from comment #2)
> I think both values should be relative to the base size, that is how I
> interpreted them and it seems to be what MS implementation is doing IIRC.

Thanks. In that case for Latin Modern Math the values 70% and 50% should be essentially the same as .71 and .71^2=0.5041 given by scriptsizemultiplier

Comment 4

[Frédéric Wang (:fredw)]

Putting this under MathML Core, since it's now defined here:

https://mathml-refresh.github.io/mathml-core/#the-math-script-level-property

Also adding "see also 1548471", several existing tests use an integer scriptsizemultiplier to avoid rounding issues with script level scaling. Probably they can be replaced with tests using font with fixed ScriptPercentScaleDown & ScriptScriptPercentScaleDown, or better converting them to use testharness.js with rounding errors rather than reftests. https://w3c-test.org/css/css-fonts/math-script-level-and-math-style contains various such tests.

Comment 5

[Frédéric Wang (:fredw)]

Attachment: Bug 1187682 - Use ScriptPercentScaleDown and ScriptScriptPercentScaleDown constants. r=emilio

Comment 6

[Pulsebot]

Pushed by [e-mail]:
https://hg.mozilla.org/integration/autoland/rev/20fadec3e9dc
Use ScriptPercentScaleDown and ScriptScriptPercentScaleDown constants. r=emilio

Comment 7

[Web Platform Test Sync Bot (Matrix: #interop:mozilla.org)]

Created web-platform-tests PR https://github.com/web-platform-tests/wpt/pull/25948 for changes under testing/web-platform/tests

Comment 8

[Razvan Maries]

Backed out for perma failures.

Push with failures: https://treeherder.mozilla.org/#/jobs?repo=autoland&selectedTaskRun=AKrFit4UTVGSU_Qh252BUw.1&resultStatus=testfailed%2Cbusted%2Cexception&searchStr=reftest&revision=62112d7e4fd9e6931ca7c246ef6ef3c4d1dd4727&failure_classification_id=2

Log: https://treeherder.mozilla.org/logviewer.html#/jobs?job_id=317398963&repo=autoland&lineNumber=16538

Backout: https://hg.mozilla.org/integration/autoland/rev/1ac1579b1471067fdeb105b8f1fed563326f5451

Comment 9

[Web Platform Test Sync Bot (Matrix: #interop:mozilla.org)]

Upstream PR was closed without merging

Comment 10

[Frédéric Wang (:fredw)]

not going back to this in the short term

Comment 11

[BugBot [:suhaib / :marco/ :calixte]]

The bug assignee is inactive on Bugzilla, so the assignee is being reset.

Comment 12

[Suhaib Mujahid [:suhaib]]

Sorry, there was a problem with the detection of inactive users. I'm reverting the change.

Comment 13

[Pulsebot]

Pushed by [e-mail]:
https://hg.mozilla.org/integration/autoland/rev/317a9610fc6d
Use ScriptPercentScaleDown and ScriptScriptPercentScaleDown constants. r=emilio

Comment 14

[Andreea Pavel [:apavel]]

https://hg.mozilla.org/mozilla-central/rev/317a9610fc6d

Comment 15

[Web Platform Test Sync Bot (Matrix: #interop:mozilla.org)]

Upstream PR merged by moz-wptsync-bot