Submitted On 13-DEC-2000
ttwang
Usages of java -Dfile.encoding=SJIS should be supported.
Otherwise how can we regression test NLS bug fixes? Only
one or two machines in our lab have foreign locales loaded.
-Thomas Wang from HP ejl lab
Submitted On 15-AUG-2001
jhawksley
I agree with Thomas. Being able to set this property is
much better than having to hard-code every IO routine to
explicitly state the character encoding.
Submitted On 24-AUG-2001
malamut
Another good reason to include especially the possibility
to set file.encoding on all JVMs is this: The algorithm
which automatically sets this property on JVM startup is
undocumented and sometimes leads to undesired results. We
just experienced this problem on our systems while using a
change root environment for a server program on a Linux
machine.
Within the change root environment, file.encoding defaulted
to some strange ANSI 1968 coding (probably just US-ASCII),
whereas, otherwise, the default was iso-8859-1 on the same
machine. For obvious reasons, we need both settings to be
consistent for our system to work, but this seems to be
impossible to achieve unless we remove the change root
environment, thus creating a security leak... :-(
Submitted On 12-SEP-2002
luke_hillary
Evaluation: "It's too late for 1.2"
It's too late for 1.4 now as well, maybe version 1.5 ?
Submitted On 08-JAN-2003
weic
why is sun so arrogant on this? file.encoding setting behaves
differently from paltform to platform, which is against java's
own advocate. the rumor is that .net can do better:-)
Submitted On 04-JUL-2003
le0nard
2 luke_hillary:
now it is too late for 1.5 ;-(
PLEASE NOTE: JDK6 is formerly known as Project Mustang
|