search

mfrey / baltimore

baltimore is an analysis script for the libARA framework
GNU General Public License v3.0
0 stars 0 forks source link

Existing result files when running simulations #24

Closed fgrosse closed 11 years ago

fgrosse commented 11 years ago

It recently noticed that we have a potential issue when we run and evaluate simulations via the baltimore script. Lets assume I just ran 30 replications of my favorite scenario. Now I decide I make a quick change to the code and then I only want to run 5 replications of the same scenario.

I'm not sure what happens to the existing result files (I think they are just overwritten without any comment by OMNeT++) but the main issue I see is that after I ran my 5 replications I still have 30 result files (5 new ones and 25 from the old run). The script will fail to recognize the difference and accumulate everything into a single replication which is completely wrong.

I kind of solved this in the plain run.sh by deleting all result files first that matched the expected file name pattern.

mfrey commented 11 years ago

I've added methods for scanning the directory for existing files. If such a file exists it asks the user if the files should be removed. However, I can't push the changes (?) will try it first in the morning.

fgrosse commented 11 years ago

Oh sorry this might really be my bad. I forced an updated because I thought hopefully you are not just now working on it. Would be the best if you contact me tomorrow morning about your git issue or you wait till I come in on the afternoon