Core dump on (controller end-simulation)
I saw a post about a similar problem way back with version 2.4, but it still seems to be an issue in 2.7.2. When I call end-simulation it results in a core dump. This may be because memory is low, but I don't see why that should keep me from exiting smoothly. Is this the only way to exit breve when executed from the command line?
-Jacob
which platform and frontend?
Which platform and frontend are you using? This will help me to debug the problem.
- jon
Ubuntu
I can't believe I forgot to mention that. However, it's a modified version of Ubuntu used only at my University, so the information probably doesn't help much. If I come up with any more useful information that will help address the problem I'll pass it on.
-Jacob
CLI or IDE? - jon
CLI or IDE?
- jon
CLI
CLI
will post new betas this weekend
I found and fixed a somewhat odd bug which *might* be causing this problem and potentially the bad alloc problem you posted earlier -- but it's a little hard to tell.
The bug is simply in freeing of Python-to-steve bridge objects, but what triggers it is Python bridge objects being created unknowingly when a steve class of the same name isn't found (possibly because the file hasn't been included).
So, for example, if you say "new Mobile" in your simulation, and you don't have the "@use Mobile" to find the steve object, the breve engine will find and instantiate a Python version of Mobile! What's more amusing is that the simulation will run exactly as expected with the foreign object.
- jon
new betas
New betas posted at http://spiderland.org/breve_betas/ -- please let me know if this does or doesn't fix your problem.
It's also possible that the new build will trigger a simulation error where there wasn't one before in the event that a Python object is being instantiate unintentionally. If this happens, it can be fixed by making sure all steve classes are properly included.
- jon
Still having problems
I wasn't having any problems with python. In fact, even before the new beta my code worked on a machine that didn't have python installed (I always got a message from breve about failing to initialize the python front end).
The error message I got, which resulted in a core dump, was:
/.../breve_2.7.5b1/bin/breve: line 15: 13335 Segmentation fault (core dumped) $DIRECTORY/breve_ex $*
The simulation finished and generated all output I wanted, but dumped core when end-simulation was called.
-Jacob
Hook, Line and Sinker
"What's more amusing is that the simulation will run exactly as expected with the foreign object."
can you throw the new
can you throw the new version of python in with the next release as well?