-
Bug
-
Resolution: Fixed
-
P2
-
1.1.6
-
1.1.6
-
sparc
-
solaris_2.5.1
-
Not verified
Issue | Fix Version | Assignee | Priority | Status | Resolution | Resolved In Build |
---|---|---|---|---|---|---|
JDK-2019425 | 1.2.0 | Brian Beck | P2 | Resolved | Fixed | 1.2beta4 |
Name: bb33257 Date: 03/09/98
font.properties.ar has the following bugs :
- dialog.1, dialoginput.1, serif.1, sansserif.1, monospaced.1
all have the fontname "David".
This is a Hebrew font and does not exist on Arabic Windows.
I replaced David with the same name as the English font
(Arial,Courier New, etc) on my Arabic NT machine and
got Arabic characters.
- All instances of
fontcharset.dialog.1=sun.io.CharToByteCp1255
should be changed to
fontcharset.dialog.1=sun.io.CharToByteCp1256
(1255 is the Hebrew Windows codepage, the correct value for
Arabic is 1256)
- The exclusion range info needs to be tuned. At a minimum,
the space codepoint should be added to each line, i.e. :
exclusion.dialog.0=0032,0100-ffff ... etc.
When 0032 is not included, the space character acts like
a LTR character even if it is between Arabic, for example :
STORAGE: MY NAME
is displayed as YM EMAN
when 0032 is not in the exclusion range
while the correct display would be EMAN YM.
This might imply that either :
- all neutral characters must be included in the exclusion
range so they are not picked up from font.0 which is
labelled as ANSI
OR
- ARABIC_CHARSET should be specified on font.0, rather
than font.1
I am not familiar enough with the effect of font.properties
to determine which is better.
======================================================================
font.properties.ar has the following bugs :
- dialog.1, dialoginput.1, serif.1, sansserif.1, monospaced.1
all have the fontname "David".
This is a Hebrew font and does not exist on Arabic Windows.
I replaced David with the same name as the English font
(Arial,Courier New, etc) on my Arabic NT machine and
got Arabic characters.
- All instances of
fontcharset.dialog.1=sun.io.CharToByteCp1255
should be changed to
fontcharset.dialog.1=sun.io.CharToByteCp1256
(1255 is the Hebrew Windows codepage, the correct value for
Arabic is 1256)
- The exclusion range info needs to be tuned. At a minimum,
the space codepoint should be added to each line, i.e. :
exclusion.dialog.0=0032,0100-ffff ... etc.
When 0032 is not included, the space character acts like
a LTR character even if it is between Arabic, for example :
STORAGE: MY NAME
is displayed as YM EMAN
when 0032 is not in the exclusion range
while the correct display would be EMAN YM.
This might imply that either :
- all neutral characters must be included in the exclusion
range so they are not picked up from font.0 which is
labelled as ANSI
OR
- ARABIC_CHARSET should be specified on font.0, rather
than font.1
I am not familiar enough with the effect of font.properties
to determine which is better.
======================================================================
- backported by
-
JDK-2019425 Arabic Font Property Problems
-
- Resolved
-