When I try to do native2ascii the attached sample file(encoding=EUC_TW),
the converted file contains extra comma at some lines.
This problem appears when using SUNWj3rt bundled in s10_20
(VERSION=1.4.1,REV=2002.09.06.05.56)
and does not appear in the previous one bundled in s10_17.

% native2ascii -encoding EUC_TW test.txt
{ "MSG269", "\u5df2\u65b0\u589e\u8edf\u9ad4\uff1a" },
{ "MSG270", "\u5df2\u79fb\u9664\u8edf\u9ad4\uff1a" },
:
{"MSG270", "\u5df2\u79fb\u9664\u8edf\u9ad4\uff1a" },
{ "MSG270", , "\u5df2\u79fb\u9664\u8edf\u9ad4\uff1a" },
            ^^^ extra comma

This extra comma causes the error when javac is executed for the
converted file.

We have used native2ascii and javac in building localized java files.
But, we cannot use the latest version unless this bug is fixed.

This problem happens on both sparc and x86.