Today I just finished all the ‘t’ sub-commands. Things were not so smooth at the beginning, But as time passed, things started getting clearer and clearer. The principle behind ‘t’ is quite simple. It has a database,
anal/types which behaves like a hash table (I guess). Links are stored with
link. prefix, structs with
struct. and pretty much everything else. You got the idea… Every command does one simple thing - query the database and compute output based on the certain conditions.
What I really expanded this week are the
t* they were easier than I expected
t* was very, very usefull it allowed us to store the types as a script so we can load/unload types. This will be very helpful when saving and loading projects.
EDIT : May 8, 2016
How will type commands be useful?
Well, they won’t be usefull on their own, and at their current state, I am not expecting anyone to be using them, but They well be very usefull for projects like radeco and for now when I add type propagation support to the analysis of r2.