Open Bug 1101550 Opened 11 years ago Updated 4 years ago

Incorrect font selected for unified Han character in untagged content

Categories

(Core :: Graphics: Text, defect)

x86_64
Linux
defect

Tracking

()

People

(Reporter: SimonSapin, Unassigned)

References

()

Details

Attachments

(1 file)

On my Archlinux machine, in http://tangorin.com/kanji/%E6%99%B4 (a Japanese dictionary website) as well as in the address bar, the character 晴 U+6674 is rendered with 円 instead of 月 as the bottom right part, which my Japanese teacher says is wrong. Devtools show that the Baekmuk Dotum font is used. It seems to be a Korean font. I suppose this appearance is correct in Korean, and this issue is a result of Unicode’s Han unification. Adding lang="ja" to the markup fixes the issue (by using a different font). Given Han unification, that’s clearly the correct solution, so feel free to close this bug. Still, I’m curious. What do we do to pick a font for unified characters when the language is not tagged? What part of the code is responsible for this choice? Blink seems to be doing something different, since Chromium gives a result that’s locally better for the same untagged content.
Severity: normal → S3
You need to log in before you can comment on or make changes to this bug.

Attachment

General

Created:
Updated:
Size: