Betterment "Auto-deposit" text is shown with a weird flag instead of a hyphen
Categories
(Core :: Layout: Text and Fonts, defect, P2)
Tracking
()
Tracking | Status | |
---|---|---|
firefox-esr78 | --- | unaffected |
firefox82 | --- | unaffected |
firefox83 | + | fixed |
firefox84 | --- | fixed |
People
(Reporter: dholbert, Assigned: jfkthame)
References
(Regression)
Details
(Keywords: regression)
Attachments
(4 files)
STR:
- In an account at Betterment.com, view one of your "Goals" and look at the "Auto-Deposit" text
ACTUAL RESULTS:
There's a flag glyph instead of a hyphen, in the "Auto-Deposit" text. See attached screenshot.
EXPECTED RESULTS:
Hyphen.
Reduced testcase coming up.
Mozregression gives me this range:
https://hg.mozilla.org/integration/autoland/pushloghtml?fromchange=97770f1880a215b81e1fe806f9d6e486246bfc69&tochange=6122ff2a25ce5c45a14d3b0cf0931f85aa9e4053
which points to this being regressed by bug 1670581.
Reporter | ||
Comment 1•4 years ago
|
||
Reporter | ||
Comment 2•4 years ago
|
||
Reporter | ||
Comment 3•4 years ago
|
||
[Tracking Requested - why for this release]: newly-introduced regression
I'm using Nightly 84.0a1 (2020-10-23) (64-bit) on Ubuntu 20.04, for what it's worth.
Assignee | ||
Comment 4•4 years ago
|
||
Huh, that's really weird. But it reproduces for me on macOS too. Will investigate.
Assignee | ||
Comment 5•4 years ago
|
||
Test file that loads the webfont from the site, and displays various dash/hyphen characters.
Here, we can see that U+2010 and U+2011, which should (per bug 1670581) fall back to rendering U+002D when they're not available in the chosen font, are instead rendering the flag glyph.
Assignee | ||
Comment 6•4 years ago
|
||
Ah, I get it.... when it's a CFF font, we use the built-in harfbuzz font functions rather than our callbacks, and that means the hyphen fallback mappings don't get applied; the weird flag must be that font's .notdef glyph.
So we need to ensure we override the get_nominal_glyph function with our own version (where the fallback is implemented) for all fonts, including CFF.
Assignee | ||
Comment 7•4 years ago
|
||
Updated•4 years ago
|
Pushed by jkew@mozilla.com: https://hg.mozilla.org/integration/autoland/rev/0a6c18c6e9d5 Ensure we use our implementation of the get_nominal_glyph callback when shaping a CFF font, so that hyphen fallbacks can take effect. r=lsalzman
Updated•4 years ago
|
Comment 9•4 years ago
•
|
||
Backed out changeset 0a6c18c6e9d5 (Bug 1673098) for causing webrender failures.
Backout link: https://hg.mozilla.org/integration/autoland/rev/4847c70f7faa440dd17636336f18bca7ba8a4b14
Failure log: https://treeherder.mozilla.org/logviewer.html#/jobs?job_id=319746146&repo=autoland&lineNumber=19399
Comment 10•4 years ago
|
||
Pushed by malexandru@mozilla.com: https://hg.mozilla.org/integration/autoland/rev/d91eaf128119 Ensure we use our implementation of the get_nominal_glyph callback when shaping a CFF font, so that hyphen fallbacks can take effect. r=lsalzman
Comment 11•4 years ago
|
||
Re-landed the changes that were backed out since the wpt-sync that followed looks to have fixed them somehow.
Re-land link: https://hg.mozilla.org/integration/autoland/rev/d91eaf1281199ecf04e36eea1d169e1bb39c371c
Comment 12•4 years ago
|
||
bugherder |
Comment 13•4 years ago
|
||
Jonathan, should we uplift this to Beta as this is an S2 bug and the fix seems to be small?
Assignee | ||
Comment 14•4 years ago
|
||
Yes, I think it'd be good to uplift; the fix is simple and should be very safe.
Assignee | ||
Comment 15•4 years ago
|
||
Comment on attachment 9183621 [details]
Bug 1673098 - Ensure we use our implementation of the get_nominal_glyph callback when shaping a CFF font, so that hyphen fallbacks can take effect. r=lsalzman
Beta/Release Uplift Approval Request
- User impact if declined: Unicode hyphens may be mis-rendered by some webfonts
- Is this code covered by automated tests?: Yes
- Has the fix been verified in Nightly?: Yes
- Needs manual test from QE?: No
- If yes, steps to reproduce:
- List of other uplifts needed: None
- Risk to taking this patch: Low
- Why is the change risky/not risky? (and alternatives if risky): Simple fix to ensure we apply a fallback character mapping to CFF as well as TTF fonts; should have no effect on any other characters than the Unicode hyphens we're targeting here.
- String changes made/needed:
Comment 16•4 years ago
|
||
Comment on attachment 9183621 [details]
Bug 1673098 - Ensure we use our implementation of the get_nominal_glyph callback when shaping a CFF font, so that hyphen fallbacks can take effect. r=lsalzman
Approved for 83 beta 6, thanks.
Comment 17•4 years ago
|
||
bugherder uplift |
Updated•4 years ago
|
Updated•4 years ago
|
Description
•