Java version:
'Open JDK 11.0.10'
Command Run:
runcluster.pl -a km ../../research/outfile_2a.dtrace ../../research/outfile_2a.decls
Command line Output:
cmd_output.txt
Input Files:
outfile_2a.decls.txt
outfile_2a.dtrace.txt
Use Case:
I have a series of system traces (originally a csv) containing variable values that lack any explicit evidence of correlation between each message. I was hoping to utilize daikon to discover conditional invariants for this. An example of an invariant I want to discover is 'if (x > threshold) then y = 0'. This can be confirmed in the original trace as a series of messages would have an increasing value of x with y = 1, until x crosses some threshold and y = 0 becomes true (the variables are not actually called x and y in the trace file). However, I have numerous similar invariants hidden in the trace and I plan to use daikon to automatically detect them. My understanding is that daikon would require a splitter info file for this and runcluster.pl would be able to create one from a .dtrace file. When I attempted to call runcluster.pl, I received the command line output pasted above with a notices saying that a java command failed to execute. What could I be doing wrong?
Side note:
On a different note, how should I format my .dtrace files such that I can discover the invariant above? I naively parsed the original trace into .dtrace, .decls files where the function correlating to a message depends on who the source is. However, the x and y variables I'm mentioning above belong to two different sources. Would combining the two into a new 'function' before calling runcluster.pl and daikon generate my desired invariant?
Found Issue:
I noticed in runcluster.pl that it attempts to call the daikon jar file with the option '--var_omit_pattern="class" ' which resulted in a 'unrecognized option' error. I'm not sure if this is necessary or not but removing it produced the output above.
Java version:
'Open JDK 11.0.10'
Command Run:
runcluster.pl -a km ../../research/outfile_2a.dtrace ../../research/outfile_2a.decls
Command line Output:
cmd_output.txt
Input Files:
outfile_2a.decls.txt
outfile_2a.dtrace.txt
Use Case:
I have a series of system traces (originally a csv) containing variable values that lack any explicit evidence of correlation between each message. I was hoping to utilize daikon to discover conditional invariants for this. An example of an invariant I want to discover is 'if (x > threshold) then y = 0'. This can be confirmed in the original trace as a series of messages would have an increasing value of x with y = 1, until x crosses some threshold and y = 0 becomes true (the variables are not actually called x and y in the trace file). However, I have numerous similar invariants hidden in the trace and I plan to use daikon to automatically detect them. My understanding is that daikon would require a splitter info file for this and runcluster.pl would be able to create one from a .dtrace file. When I attempted to call runcluster.pl, I received the command line output pasted above with a notices saying that a java command failed to execute. What could I be doing wrong?
Side note:
On a different note, how should I format my .dtrace files such that I can discover the invariant above? I naively parsed the original trace into .dtrace, .decls files where the function correlating to a message depends on who the source is. However, the x and y variables I'm mentioning above belong to two different sources. Would combining the two into a new 'function' before calling runcluster.pl and daikon generate my desired invariant?
Found Issue:
I noticed in runcluster.pl that it attempts to call the daikon jar file with the option '--var_omit_pattern="class" ' which resulted in a 'unrecognized option' error. I'm not sure if this is necessary or not but removing it produced the output above.