128338 - Incorrect fonts used for certain "lang=" attributes

Status: VERIFIED FIXED

(Core :: Internationalization, defect)

Description

[Per Kristian Gjermshus]

When setting NS_FONT_DEBUG=3 and loading the page in the URL field I get the 
output below. The "pre" section with the lang="nb" lang="foobar" uses the comic
font and the rest are correct. The comic font is the first with iso8859-3
encodint on my system, so that must be why it is selected. I first thought that
the problem was that the language nb mapped to the wrong character set. That
does not explain why "foobar" renders wrong while "foo-bar" seems to be correct. 

As far as I can tell from looking at the relevant standards both "no",
"no-bokmal" and "nb" should be correct identifiers for the same language. 


GetFontNames -adobe-courier-*-*-*-*-*-*-*-*-*-*-iso8859-1
loaded -adobe-courier-medium-r-normal--12-120-75-75-m-70-iso8859-1
FindFont(0061)[monospace, ]
returns -adobe-courier-medium-r-normal--12-120-75-75-m-70-iso8859-1
GetFontNames -*-*-*-*-*-*-*-*-*-*-*-*-cns11643.1986-2
GetFontNames -*-*-*-*-*-*-*-*-*-*-*-*-cns11643.1992.2-0
GetFontNames -*-*-*-*-*-*-*-*-*-*-*-*-cns11643.1992-2
GetFontNames -*-*-*-*-*-*-*-*-*-*-*-*-cns11643.1992-3
GetFontNames -*-*-*-*-*-*-*-*-*-*-*-*-cns11643.1992.3-0
GetFontNames -*-*-*-*-*-*-*-*-*-*-*-*-cns11643.1992.4-0
GetFontNames -*-*-*-*-*-*-*-*-*-*-*-*-cns11643.1992-4
GetFontNames -*-*-*-*-*-*-*-*-*-*-*-*-cns11643.1992.5-0
GetFontNames -*-*-*-*-*-*-*-*-*-*-*-*-cns11643.1992-5
GetFontNames -*-*-*-*-*-*-*-*-*-*-*-*-cns11643.1992.6-0
GetFontNames -*-*-*-*-*-*-*-*-*-*-*-*-cns11643.1992-6
GetFontNames -*-*-*-*-*-*-*-*-*-*-*-*-cns11643.1992.7-0
GetFontNames -*-*-*-*-*-*-*-*-*-*-*-*-cns11643.1992-7
GetFontNames -*-*-*-*-*-*-*-*-*-*-*-*-cns11643-2
GetFontNames -*-*-*-*-*-*-*-*-*-*-*-*-cns11643-3
GetFontNames -*-*-*-*-*-*-*-*-*-*-*-*-cns11643-4
GetFontNames -*-*-*-*-*-*-*-*-*-*-*-*-cns11643-5
GetFontNames -*-*-*-*-*-*-*-*-*-*-*-*-cns11643-6
GetFontNames -*-*-*-*-*-*-*-*-*-*-*-*-cns11643-7
GetFontNames -*-*-*-*-*-*-*-*-*-*-*-*-iso8859-3
loaded -microsoft-comic sans ms-medium-r-normal--12-*-0-0-p-*-iso8859-3
FindFont(0061)[monospace, ]
returns -microsoft-comic sans ms-medium-r-normal--12-*-0-0-p-*-iso8859-3
FindFont(0061)[monospace, ]
returns -adobe-courier-medium-r-normal--12-120-75-75-m-70-iso8859-1
FindFont(0061)[monospace, ]
returns -microsoft-comic sans ms-medium-r-normal--12-*-0-0-p-*-iso8859-3
FindFont(0061)[sans-serif, ]
returns -monotype-arial-medium-r-normal--16-*-0-0-p-*-iso8859-1

Comment 1

[Per Kristian Gjermshus]

Attachment: Screenshot showing what this looks like on my system

Comment 2

[Roy Yokoyama]

I am clueless in Linux.  
assigning to shanjian and cc'ing bstell

Comment 3

[R.K.Aa.]

"no" and "nb" are the same language. Why does moz suddenly try to get a unicode
font? It should be pulling iso8859-1.

The same walk through cns11643* is performed in a debug here, but moz picks
another font - adobe-courier - since that is the first font on my system that
maps to the cns "fallback" (?) iso8859-3

Resolving as new - this looks odd.

Comment 4

[Per Kristian Gjermshus]

This might be a duplicate of #60328

Comment 5

[Shanjian Li]

Attachment: patch

Comment 6

[Shanjian Li]

the problem is in our language group look up. "no" is specified in
langGroups.properties, so we have no problem identify its lang group. For rest
of the language specification, we do not have such item in the property file. A
bug in the function cause language name contains "-" default to x-western, but
language name without "-" will default to nothing. The patch I proposed fix this
problem. 

Comment 7

[R.K.Aa.]

This is a goodie.. built with the patch on Linux: Things look a lot better now.

Comment 8

[Shanjian Li]

*** Bug 129371 has been marked as a duplicate of this bug. ***

Comment 9

[Shanjian Li]

nominate this one as nsbeta1

Comment 10

[Shanjian Li]

*** Bug 119284 has been marked as a duplicate of this bug. ***

Comment 11

[Simon Montagu :smontagu]

Comment on attachment 72154 [details] [diff] [review]
patch

r=smontagu

Comment 12

[Christopher Blizzard (:blizzard)]

Comment on attachment 72154 [details] [diff] [review]
patch

sr=blizzard

Comment 13

[Ronald Clifford Buckman]

Confirming this problem on Windows 98.  I get an Algerian font on my computer,
which is the first Western, fantasy font.  The fcc's FM query has this problem.
 <a
href="http://www.fcc.gov/fcc-bin/fmq?state=WA&call=&arn=&serv=&city=Longview&freq=0.0&fre2=107.9&facid=&dkt=&list=2&dist=&dlat2=&mlat2=&slat2=&NS=N&dlon2=&mlon2=&slon2=&EW=W">This

 link</a> demonstrates the problem.  The lang specified is "english".

Comment 14

[Frank Tang]

I think we should also add
nb=x-western
to 
/intl/locale/src/langGroups.properties


nsbeta1+ and ask nhotta to check in for shanjian

without this patch, some webpage will show very bad type face. 

Comment 15

[Scott Collins]

Comment on attachment 72154 [details] [diff] [review]
patch

a=scc

Comment 16

[nhottanscp]

checked in to the trunk

Comment 17

[Yuying Long]

Verified fixed on 03-25 trunk build / RH7.2-JA.