Community
Participate
Working Groups
See bug 228936 for some history. Currently all console output from a profiled Java application is converted to the system's default encoding. So for example if a Java application outputs Japanese characters, then on an English Windows system the characters will show up as ?. Supporting different character sets can be achieved by setting the file.encoding system property. This is the approach taken by the JDT launch implementation, storing the encoding in the DebugPlugin.ATTR_CONSOLE_ENCODING attribute of the launch config. JDT chooses a sensible default based on the encoding of the file where main() resides. For parity with the JDT launch, we should pass -Dfile.encoding=<value of ATTR_CONSOLE_ENCODING>. By doing so we can support the display of any character set supported by the workbench's JVM.
Created attachment 167565 [details] patch Attached patch implements the needed changes.
Created attachment 168722 [details] patch2 Updated patch to remove mention of wrong bugzille #
Jon, Kathy, Can this go into 4.7.0 or should we wait for 4.7.1? This is an enhancement I found while working on 308169, so as far as I know it is not affecting anybody.
Notes on further testing: - With an External Java launch on OS/390 the behaviour is unchanged. i.e. adding -Dfile.encoding seems to have no effect on the output. So there is no improvement, but does not seem to have any negative impact. - External Java launch on Linux platform does seem to benefit from the improvement; if encoding is correctly specified then proper national language characters can be output in profiling mode.
Deferring to TPTP 4.7.1.
Created attachment 182147 [details] Updated patch against 4.7.1
Missed this in 4.7.1 cycle. Checked into HEAD and resolving for 4.7.2 cycle.
Verified in TPTP-4.7.2-201101121900.
Closing in TPTP 4.7.2 on behalf of the originator/owner.
Closing.
