74889 - Default Linux Chinese font seems too big and there are some character display as "?"

Status: VERIFIED FIXED

(Core :: Internationalization, defect, P2)

Description

[Yuying Long]

Build: 04-05-2001 Linux trunk

1. Launch the browser and go: http://www.china.com and visit both Traditional
and Simplified Chinese page, you will see those Chinese characters inside the
bottons on the right top of the page are too big, it's not a good UI.(see
attached screeshot)
2. There are some "?" display at the middle of the page.

Note: This is not the case for Windows and Mac.  And there is similar bug 60530
before, I guess maybe we should change the default fonts?

Comment 1

[Yuying Long]

Attachment: Screen shot for big font characters and "?"

Comment 2

[nhottanscp]

What is the font face and size setting in your pref for Chinese?

Comment 3

[Yuying Long]

For traditional Chinese:
1. Serif: Greyed out menu item "No font avalible for Traditional Chinese", and 
there is no way that you can change this font, and size set to "16".
2. Sans Serif: There are only different kinds of iso-88591 font avalible. 
3. Monospace: Only different kinds of iso-88591/8859-15 avalible, size set to 
"13" by default.  

For Simplified Chinese:
Serif: isas-fangsongti-gb2312, Sans Serif and Monospace: isas-songti-gb2312, all 
size set to "16".

The "?" only existing in Traditional Chinese, that's problem the font problem, 
but for Simplified even I have font set up, characters still display too large.

Comment 4

[nhottanscp]

Is this a regression?
Reassign to bstell.

Comment 5

[James "Dexter" Akula]

Is this a dupe of bug 76897?  Perhapse it is all languages.

Comment 6

[Jaime Rodriguez, Jr.]

Brain/Nhotta - Please accept and plus this one. 

Per i18n triage, we are assigning a P2 and a M0.9.1

Comment 7

[Jaime Rodriguez, Jr.]

Suggest we move this to M0.9.2

Comment 8

[kill this account]

I believe that Yuying Long's comments of 2001-04-05 13:31 are because
of a bug in the font preferences code. That bug is now fixed.

> > For traditional Chinese:
> 1. Serif: Greyed out menu item "No font avalible for Traditional Chinese", 
>    and there is no way that you can change this font, and size set to "16".
> 2. Sans Serif: There are only different kinds of iso-88591 font avalible. 
> 3. Monospace: Only different kinds of iso-88591/8859-15 avalible, size set to 
>    "13" by default.  

Yuying: would you come by and show me what is happening here?

Thanks

Comment 9

[Jaime Rodriguez, Jr.]

changing to nsbeta1-. great, if this problem has already been addressed, but 
this one does not meet beta stopper guidelines. i wouldn't put too much effort 
into this one.

Comment 10

[Phil Peterson]

TM to 0.9.2 per PDT triage (it's OK to check it in by Friday or after 0.9.2 branch).

Comment 11

[kill this account]

Attachment: barebone html showing the problem

Comment 12

[kill this account]

set NS_FONT_DEBUG=5 (ie: 4 + 1) to see font loading

the page says load:

 .text2 {color:#000000; font-size: 12px; line-height: 130%; font-family:
"ËÎÌå";  text-decoration: none}

but we loaded:

 loaded -isas-song ti-medium-r-normal--16-160-72-72-c-160-gb2312.1980-0

Comment 13

[kill this account]

changing the pref("font.min-size.variable.zh-CN", 10); had no effect
(also tried zh-TW, ja, ko)

Comment 14

[kill this account]

two separate factors apply here:

1) in modules/libpref/src/unix/unix.js
     pref("font.min-size.variable.zh-CN", 10);
     pref("font.min-size.fixed.zh-CN", 10);

2) in gfx/src/gtk/nsFontMetricsGTK.cpp the code is trying to avoid
   using scaled fonts because they look so poor in some cases. Thus
   the code says if an available unscaled font is between 0.8 and 1.8
   of the desired size use it. In this case the 12px font * 1.8 = 21px
   so the (Redhat stock) 16px unscaled font is used.

   2266   /*
   2267    * XXX Maybe revisit this. Upper limit deliberately set high (1.8)
in     2268    * order to avoid scaling Japanese fonts (ugly).
   2269    */
   2270   if ((ratio > 1.8) || (ratio < 0.8)) {
   2271     scalable = 1;
   2272   }

Comment 15

[Jaime Rodriguez, Jr.]

Marking as nsCatFood.

Comment 16

[Frank Tang]

I think this is a critical one.

Comment 17

[kill this account]

Attachment: image showing without/with font scaling control

Comment 18

[kill this account]

Attachment: image showing overlapping characters

Comment 19

[kill this account]

The Unix client does not do a good job choosing between bitmap and scaled fonts
and can end up using vastly oversized bitmaps when scaled fonts are available.

In attachment http://bugzilla.mozilla.org/showattachment.cgi?attach_id=37750
note how much larger the bitmapped fonts are compared to the scaled fonts.

In attachment http://bugzilla.mozilla.org/showattachment.cgi?attach_id=37751
note the overlapping characters.

Comment 20

[kill this account]

Attachment: patch adding font scaling control

Comment 21

[Frank Tang]

OK, I also attach two screen shot, using the same machine, one with 4.x which 
disply reasonable result and one with 6.x which disply bad result. 

Comment 22

[Frank Tang]

Attachment: 4.x result which is reasonable.

Comment 23

[Frank Tang]

Attachment: 6.x result which is not reasonable.

Comment 24

[Frank Tang]

pdt+ base on 6/11 pdt meeting.

Comment 25

[kill this account]

Attachment: smaller patch with only the non-critical parts

Comment 26

[Frank Tang]

r=ftang for Created an attachment (id=38347)
The idea is bstell will break his origional patch into two patch. The first one 
(38347) include name changes without new functionality . and the sencond patch 
include the real functionality changes. We break into two patch so the review
will be easier. 

Comment 27

[Christopher Blizzard (:blizzard)]

sr=blizzard

Comment 28

[Christopher Blizzard (:blizzard)]

a=blizzard on behalf of drivers for the trunk

Comment 29

[kill this account]

attachment 38347 [details] [diff] [review] checked into trunk

Comment 30

[Frank Tang]

clear " part 1 r=ftang, sr=blizzard, need a=" for status since bstell check in 
the first part .

Comment 31

[kill this account]

Attachment: patch with the parts that control font scaling

Comment 32

[kill this account]

attachment 39018 [details] [diff] [review] has the code that actually controls scaling

Comment 33

[Frank Tang]

r=ftang for 06/18/01 17:23 
http://bugzilla.mozilla.org/showattachment.cgi?attach_id=39018

Comment 34

[Frank Tang]

change status to "[PDT+] r=ftang, ask blizzard to sr= (6/19 9:45) a="

Comment 35

[Frank Tang]

Ask alecf to sr= 12:32

Comment 36

[Christopher Blizzard (:blizzard)]

sr=blizzard

Comment 37

[Asa Dotzler [:asa]]

a= [e-mail] for checkin to the trunk.
(on behalf of drivers)

Comment 38

[kill this account]

checked in

Comment 39

[kill this account]

*** Bug 83416 has been marked as a duplicate of this bug. ***

Comment 40

[kill this account]

*** Bug 27551 has been marked as a duplicate of this bug. ***

Comment 41

[kill this account]

Attachment: patch; pref to not distinguish outline scaled fonts from bitmap scaled fonts

Comment 42

[Yuying Long]

Brian:
I still see some "?" shows in Chinese page(see attached file later).
Also the buttons that I original reported still appears too big.

Comment 43

[Yuying Long]

Attachment: See the "?" in this page

Comment 44

[kill this account]

Yuying: since this bug has so much size related discussion
would you kindly open a new bug to address the '?'.

In that new bug could you note the encoding and the missing character's value
(in the encoding or in unicode)?

Comment 45

[Yuying Long]

OK, the "?" issue is a seperater problem.

However seems the Simp. Chinese font still appreas too big - see my screenshot
later.

Comment 46

[Yuying Long]

Attachment: 07-13 branch build.

Comment 47

[Yuying Long]

I'll mark this one as verified for now cause the "?" problem has been fixed.
For the font still showing big, I'll open another bug for that, it's too late to
re-open this PDT bug at this point.