This Bugzilla instance is a read-only archive of historic NetBeans bug reports. To report a bug in NetBeans please follow the project's instructions for reporting issues.
[010814] Only on W98
Created attachment 2285 [details] StackTrace
This bug occures sometimes (it means, let say once in ten attempts). Also the files are compiled, Exception is written to the Output window, insofar that user thinks, that files were not compiled.
I can't say for sure that this is a fastjavac bug. This exception is caused by win32 CreateProcess call which returns false (execution failed) with error number set in 0 (ERROR_SUCCESS). Looks like we've faced MS puzzle.
Has been forwarded to the Sun's fastjavac team.
So, I do not think this is problem of fastjavac, the same happens to me in CPPLite when compiling/linking on W98 (native utility is GNU gcc by cygwin). I think this is either win or jre problem. I was trying to investigate it a bit, but without success.
Really seems like Win platform bug. Temporarily closing, if some solution is found, it will be fixed.
Bug is probably given by bug #4086045 (java bugparade). Thread which runs external process has to have NORM_PRIORITY on W98.
Created attachment 3644 [details] Suggested patch
Created attachment 3645 [details] Source and compiled class
To psuchomel: next time please use context diff (cvs diff -c) or unified diff (cvs diff -u) when producing a patch!!!
I don't see that much problem with the workaround but I don't like the patch. I'll find a better way. Should we try to integrate this fix into 3.3.0? AFAIK fastjavac is not the default compiler. How often does the problem occur on Win98?
Created attachment 3661 [details] a better patch to fix the bug
I refined the patch (see attachment) and committed it into release33 branch
merged into release330 branch
VERIFIED in [200207080100]
Resolved for 3.4.x or earlier, no new info since then -> closing.