335359 - cannot see caret easily if caret is between some Chinese characters

Status: VERIFIED FIXED

(Core :: Layout, defect)

Description

[Masayuki Nakano [:masayuki] (he/him)(JST, +0900)]

Now, bug 287813 and bug 334635 were fixed, we can see the new caret that is same color as forground color, and that is not ineverted. Therefore, when the caret is between some Chenese characters, we cannot see the caret to easy. See screenshot.

Comment 1

[Masayuki Nakano [:masayuki] (he/him)(JST, +0900)]

Attachment: screenshot

Comment 2

[Masayuki Nakano [:masayuki] (he/him)(JST, +0900)]

I think that we have two way for fix.

1. the caret should be drawn non-background-color and non-foreground color.
2. we should ensure that the caret drawn area doesn't have any glyphs.

I think that (2) is not good for some languages. Because current implement breaks some script if the text has letter-spacing.

But I don't know what color is best for the caret in (1).

Comment 3

[Robert O'Callahan (:roc) (email my personal email if necessary)]

what if we make the caret wider?

Comment 4

[Masayuki Nakano [:masayuki] (he/him)(JST, +0900)]

(In reply to comment #3)
> what if we make the caret wider?

Yeah, it's cool.

Comment 5

[Jesse Ruderman]

Just for Chinese text?  (A wide caret for English text, like Mozilla used to have, is kinda ugly and non-standard.)

Comment 6

[Masayuki Nakano [:masayuki] (he/him)(JST, +0900)]

(In reply to comment #5)
> Just for Chinese text?  (A wide caret for English text, like Mozilla used to
> have, is kinda ugly and non-standard.)

No, for all scripts. I'll work on this ASAP if nobody work on this.

Comment 7

[Robert O'Callahan (:roc) (email my personal email if necessary)]

Making the caret wider based on a test of the Unicode character before/after it should be not too hard to do.

Comment 8

[Benjamin Smedberg]

-> masayuki per his comment

Comment 9

[Masayuki Nakano [:masayuki] (he/him)(JST, +0900)]

Attachment: screenshot of wide caret

wide caret(2px) is good for Chinese characters, but for latain text, I think that this seems to be too thick. Is it O.K.?

Comment 10

[Robert O'Callahan (:roc) (email my personal email if necessary)]

How about making it 2px wide if it's between two CJK characters, otherwise 1px?

Comment 11

[Masayuki Nakano [:masayuki] (he/him)(JST, +0900)]

Um... I don't have an idea for it. But I try it.

Comment 12

[Masatoshi Kimura [:emk]]

*** Bug 336086 has been marked as a duplicate of this bug. ***

Comment 13

[Masayuki Nakano [:masayuki] (he/him)(JST, +0900)]

We have a dup(bug 336086) that said the 't' has same problem. This is difficult, because this bug depends on the font.

Comment 14

[Ryan Jones-Ward [:sciguyryan]]

(In reply to comment #13)
> We have a dup(bug 336086) that said the 't' has same problem. This is
> difficult, because this bug depends on the font.
> 

I was the reporter of that one... its interesting to note that it does not seem to occur with any numbers at all but does with some characters.

Comment 15

[Robert O'Callahan (:roc) (email my personal email if necessary)]

Attachment: fix

This turned out to be incredibly simple. The patch makes the caret an extra CSS pixel thicker when it's positioned at a Kanji character. Seems to work great.

Masayuki, please try the patch and tell me if you like it.

Comment 16

[Blake Kaplan (:mrbkap) (inactive)]

Comment on attachment 316946 [details] [diff] [review]
fix

Is it possible to replace the magic numbers with a meaningful identifier or a pointer to where you got them from? r+sr=mrbkap either way.

Comment 17

[Robert O'Callahan (:roc) (email my personal email if necessary)]

Sure, will do.

Comment 18

[Robert O'Callahan (:roc) (email my personal email if necessary)]

Comment on attachment 316946 [details] [diff] [review]
fix

Simple, low-risk fix for a CJK usability regression.

Comment 19

[Damon Sicore (:damons)]

Comment on attachment 316946 [details] [diff] [review]
fix

a1.9+=damons

Comment 20

[Robert O'Callahan (:roc) (email my personal email if necessary)]

checked in.

Comment 21

[Masayuki Nakano [:masayuki] (he/him)(JST, +0900)]

-> v.

But the flexible caret width behavior is pretty strange. If I found some better idea, I'll file a new bug. But now this is ok. The accessibility issue is gone. Thank you, roc.