96535 - Font sizes too large for Japanese UI

Status: VERIFIED FIXED

(Core :: Internationalization, defect)

Description

[Jon Rubin]

The client font size for Japanese localized builds is too large.  This leads to
several problems in localized builds, the most severe of which involves
window-size problems that cause characters to be overlapped by other characters,
or the window border itself.  A smaller font size would hopefully alleviate this
situation.

Comment 1

[Allan Masri]

This problem is most severe with Red Hat linux 7.1

Comment 2

[Jon Rubin]

RedHat 7.1-JA, to be specific.  We have not yet checked 7.1-US.

Comment 3

[rchen]

It will be nice if we can do something for emojo release.

Comment 4

[Jaime Rodriguez, Jr.]

Ftang/Bstell - What are your thoughts? Does this look doable for M0.9.4?

Comment 5

[kill this account]

The minimum Japanese font size is contolled by 

    pref("font.min-size.variable.ja", 16);
    pref("font.min-size.fixed.ja", 16);

I had set these to 10 but this lead to Japanese text overlapping, bug 89520,
so I was force to put them back to 16.

This may be fixed by bug 95721 but until then there is not good solution.

Comment 6

[Yuying Long]

Just want to make clear:
This problem is not only on Linux RH7.1-Ja, but also on Linux RH6.2-Ja.

Comment 7

[rbs]

Yep, this problem will be solved as a fall trough from bug 95721. See my
comments on bug 30910 which is the main tracker bug for the font's min-size that
will allow to bump the content whithout affecting the chrome. 

Comment 8

[kill this account]

Bug 95721, "get actual string height" (aka: bug 74186, bug 96609, bug 99010),
will not be addressed in the 0.9.4 time frame.

Other options:

 1) A separate Redhat 7.1 release with normally sized dialogs and
    a separate Redhat 6.2 release with big dialogs.

 2) Include some Japanese fonts in the release so we can use them
    in a normally sized dialogs.

Comment 9

[msanz]

marking nsbranch- since it won't be fixed in this release

Comment 10

[kill this account]

merge cc: list from bug 99412

Comment 11

[kill this account]

*** Bug 99412 has been marked as a duplicate of this bug. ***

Comment 12

[kill this account]

Attachment: patch; adjusted Japanese font sizes

Comment 13

[kill this account]

Ying has been using these settings for a while now.

Could I get some other people to also try this patch?

(Katakai?)

Thanks

Comment 14

[kill this account]

For those that try/test these settings:

In addition to lowering the minimum size; 

  Attachment 51120 [details] [diff] sets the prefered font family and size. 

This is intended to change the *default* values.

People that would prefer something different *should* go to the font
prefs dialog (Edit->Preferences->Fonts) and change the settings.

One of the hoped for test results is to understand what will happen to
typical users if/when this becomes part of the standard release.

Comment 15

[Masaki Katakai]

I'll try, also will ask japanese-users to test the setting on mailing-list.

Comment 16

[kill this account]

Katakai-san: ?????????????? (thanks.)

Comment 17

[kill this account]

(pasting Japanese did not work)

Comment 18

[Masaki Katakai]

I'm seeing one problem on RH6.2J,

  Changing font face to -misc-fixed-jis and there is only one
  bold font exists for 10 point size and the font is now used
  for bold style. Of course, it's smaller than the normal fonts.

But I'm not sure the issue will be serious. Because RH6.2J is
still main target for Mozilla? I believe RH7.1 does not have
this issue.

I have some questions about your comments in bug 99412.

> pref("font.size.variable.ja", 18); 

>On both RH 6.2J and 7.0J a "font.size.variable.ja" of 16 looks
>okay on many pages but leads to jammed chars on 
>http://www.yahoo.co.jp. Moving up to 18 seems to spread it out
>enough.

I think this 18 is too big. Can you take the snapshot where
you're seeing the "jammed chars" on yahoo.co.jp?

> pref("font.scale.bitmap.min.ja", 40); 

What does this changing mean?

Comment 19

[kill this account]

> > pref("font.scale.bitmap.min.ja", 40); 
> What does this changing mean?

It means "do not use bitmap scaling" when the requested size is less than 40 px

> I think this 18 is too big. Can you take the snapshot where
> you're seeing the "jammed chars" on yahoo.co.jp?

Ying: would you kindly do this?

Comment 20

[Ying-Lin Xia]

Change font.size.variable.ja to 16 does make Yahoo page even worse. Please see 
screenshots attached below.

Comment 21

[Ying-Lin Xia]

Attachment: font.size.variable.ja=18, Linux 7.0J

Comment 22

[Ying-Lin Xia]

Attachment: font.size.variable.ja=16, Linux 7.0J

Comment 23

[kill this account]

Now that bug 99010 is in we need to reevaluate this.

We also need to consider bug 95721.

Comment 24

[Masaki Katakai]

Attachment: unix.js of 0.9.4 with the patch, can be used on 0.9.4 with official JLP0.9.4

Comment 25

[Masaki Katakai]

Attachment: unix.js of nightly, can be used on nightly but no JLP

Comment 26

[kill this account]

--> ftang

Comment 27

[Frank Tang]

bulk move NEW FUTURE bug to ASSIGN

Comment 28

[kill this account]

I believe this has been fixed some time ago.

If it is still an issue please reopen

Comment 29

[Yuying Long]

Yes, we have fixed it on N6.2 and N6.2.1 JA.