VTSessionDB vs. Program object locking

[v-p-b]

I'm trying to use pyhidra to initiate a Version Tracking session, my code boils down to this:

p = GhidraProject.openProject(sys.argv[1], sys.argv[2], False)
# ...    
src_program = p.openProgram("/".join(src_path[0]), src_path[1], False)
dst_program = p.openProgram("/".join(dst_path[0]), dst_path[1], False)
# ...
db = VTSessionDB("Test", src_program, dst_program, session_ref)

On VTSessionDB instantiation I get the following exception:

java.lang.RuntimeException: java.lang.RuntimeException: ghidra.framework.store.LockException: domain object(s) are busy/locked

.isLocked() on both src and dst Program objects return False. I don't have other running Ghidra processes either.

How should I create input objects for VTSessionDB so locking won't break the process?

[ryanmkurtz]

I'm playing around with this now, and am having the same issue. I am assuming there is an issue in GhidraProject class, which may not get exercised much these days. In trying to use alternate API calls, it is clear that PyGhidra really needs some built-in support for working with projects better. I plan on pursuing this in the coming weeks.

[v-p-b]

Thank you for looking into this!

While accessing Programs from a GhidraProject looks like the "right" way to me (Programs are part of Projects) I'm also interested in accessing Program objects in alternative ways, even from Java directly.

[ryanmkurtz]

Yes, it indeed does look like the right way. I need to ask other team members about that error tho.

If you are writing a more standard Java-based GhidraScript, you can see some examples here:

• https://github.com/NationalSecurityAgency/ghidra/blob/master/Ghidra/Features/VersionTracking/ghidra_scripts/AutoVersionTrackingScript.java
• https://github.com/NationalSecurityAgency/ghidra/blob/master/Ghidra/Features/VersionTracking/ghidra_scripts/CreateAppliedExactMatchingSessionScript.java