Though, I set variable NYTPROF to "forkdepth:1", When i try to nytprofmerge the profiler outputs for my forked children along with it's parent, I get a message that says:
Reading nytprof.13232.out.21321 Option 'forkdepth' has value '-2' in nytprof.13232.out.21321 which differs from the previous value '-1'; this implies inconsistent profiles and thus garbage results.
Does anyone know a fix or why this is happening?
I'd take a guess that that's a bug in nytprofmerge and can be ignored. Feel free to file a bug report. (A small script that reproduces the problem would be super helpful).