ideas.txt

1.9.2015
by xorpd.

This project will allow saving a database of known/reversing functions, and then
finding similarities between archived functions and functions in new IDBs.


Main plan:


Server:
-------

- Saves a database of all functions known:
  For every function:
  - sha256 hash (To identify the function 1 to 1)
  - Catalog1 hash (To check similarities).
  - Reverser's name.
  - Function name (As given by the reverser).
  - Head comment (As given by the reverser).

- Handles the following request:
  - Given a function:
    - Search for exact sha256 match. If found, return it.
    - Return the function with the closest Catalog1 hash (If it is close enough.
      Don't return it if it is has a very low grade).


Client:
-------

- Could be run from outside IDA?

Main functions:

- Find known: Send all functions to the server (The functions themselves).
  - Wait for response. For each function, a tuple:
    {reverser_name:, func_name:, head_comment, exact_match: } is sent back

  - Every function for which a similarity was found will be given a name of the
    format: SIMILAR__{grade}__name
    The head comment will be changed to contain some more information.


- Commit reversing.
  - Send all the functions that were reversed (Given an important name) to the
    server. Every function will be sent using the format:

    {reverser_name:, func_name: , func_data: ,head_comment: }

    - Functions that begin with SIMILAR__ will not be sent.
    - Functions without a name will not be sent.