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.
There seems to be a race condition causing an intermittent failure initializing a remote compiler collection with an existing userdir. To duplicate: 1. Remove your userdir 2. Run the IDE and define a remote host and do some action (build or debug). 3. Exit the IDE and restart. At this point I get a botched compiler collection about 25% of the time. If you open the T->O panel and select the remote host, it shows 0 compiler sets. If I run with "-J-Dcnd.remote.logger=FINE" I see the CompilerSetManager.initRemoteCompilerSets() has been called twice for and that there was an IOException in CompilerSetScriptManager.runScript(). At this point, the remove host is unusable (although I can remove the node and re-add it). Here is the cnd.remote.logger output (from messages.log): > FINE [cnd.remote.logger]: CSM.init: initializing remote compiler set for: gordonp@pucci > WARNING [cnd.remote.logger]: CSM.initRemoteCompilerSets for gordonp@pucci [PENDING] 1st call to initRemoteCompilerSets > FINE [cnd.remote.logger]: RSR.validate2: Validating gordonp@pucci > FINE [cnd.remote.logger]: RemoteCommandSupport<Init>: Running [...] on gordonp@pucci > FINE [cnd.remote.logger]: RSS.needsSetupOrUpdate: GET_SCRIPT_INFO returned 0 > FINE [cnd.remote.logger]: RemoteCommandSupport<Init>: Running [...] on gordonp@pucci > FINE [cnd.remote.logger]: RSS.getBinaryUpdates: GET_LIB_INFO returned 0 > FINE [cnd.remote.logger]: CSM.init: initializing remote compiler set for: gordonp@pucci > WARNING [cnd.remote.logger]: CSM.initRemoteCompilerSets for gordonp@pucci [PENDING] 2nd call... > FINE [cnd.remote.logger]: CSSM.runScript: Reading input from getCompilerSets.bash > FINE [cnd.remote.logger]: platform [SunOS i86pc] > FINE [cnd.remote.logger]: CSSM.runScript: Reading input from getCompilerSets.bash > FINE [cnd.remote.logger]: platform [SunOS i86pc] > FINE [cnd.remote.logger]: line [GNU;/opt/sfw/bin;gcc;g++;gdb;gmake] > FINE [cnd.remote.logger]: line [SunStudio_10;/opt/ss10/SUNWspro/bin;cc;CC;dmake;gdb=/usr/local/bin/gdb] > FINE [cnd.remote.logger]: line [SunStudio_11;/opt/ss11/SUNWspro/bin;cc;CC;dmake;gdb=/usr/local/bin/gdb] > FINE [cnd.remote.logger]: line [SunStudio_12;/opt/ss12/SUNWspro/bin;cc;CC;dmake;gdb=/usr/local/bin/gdb] > WARNING [cnd.remote.logger]: CSSM.runScript: IOException [Stream closed] Whenever initRemoveCompilerSets is called twice, I get this exception and the CSM doesn't get initialized. > FINE [cnd.remote.logger]: CSM.initRemoteCompileSets: platform = 1 > FINE [cnd.remote.logger]: CSM.initRemoteCompileSets: platform = 1 > FINE [cnd.remote.logger]: CSM.initRemoteCompileSets: line = [GNU;/opt/sfw/bin;gcc;g++;gdb;gmake] > FINE [cnd.remote.logger]: CSM.initRemoteCompileSets: line = [SunStudio_10;...] > FINE [cnd.remote.logger]: CSM.initRemoteCompileSets: line = [SunStudio_11;...] > FINE [cnd.remote.logger]: CSM.initRemoteCompileSets: line = [SunStudio_12;...] > FINE [cnd.remote.logger]: CSM.initRemoteCompilerSets: Found 2 compiler sets > FINE [cnd.remote.logger]: RemoteCommandSupport<Init>: Running [PATH=/bin:/usr/bin; env] on gordonp@pucci > FINE [cnd.remote.logger]: RemoteCommandSupport<Init>: Running [/...] on gordonp@pucci > FINE [cnd.remote.logger]: RemoteCommandSupport<Init>: Running [...] on gordonp@pucci > FINE [cnd.remote.logger]: RemoteCommandSupport<Init>: Running [...] on gordonp@pucci > FINE [cnd.remote.logger]: CSM.initRemoteCompilerSets: Found 2 compiler sets > FINE [cnd.remote.logger]: RemoteCommandSupport<Init>: Running [/usr/bin/test -d "" -o -f ""] on gordonp@pucci
fixed by changing threading model