Palkeo is maintaining a more up to date for of Panoramix. Be sure to check it out:
https://github.com/palkeo/panoramix
git clone https://github.com/eveem-org/panoramix.git
pip3 install -r requirements.txt
You need python3.8 to run Panoramix. Yes, there was no way around it.
python3.8 panoramix.py address [func_name] [--verbose|--silent|--explain]
e.g.
python3.8 panoramix.py 0x06012c8cf97bead5deae237070f9587f8e7a266d
or
python3.8 panoramix.py kitties
Output goes to two places:
console
cache_pan/
directory - .pan, .json, .asm files
If you want to see how Panoramix works under the hood, try the --explain
mode:
python3.8 panoramix.py kitties paused --explain
python3.8 panoramix.py kitties pause --explain
python3.8 panoramix.py kitties tokenMetadata --explain
func_name -- name of the function to decompile (note: storage names won't be discovered in this mode) --verbose -- prints out the assembly and stack as well as regular functions, a good way to try it out is by running 'python panoramix.py kitties pause --verbose' - it's a simple function
There are more parameters as well. You can find what they do in panoramix.py.
Some contract addresses, which are good for testing, have shortcuts, e.g. you can run 'python panoramix.py kitties' instead of 'python3 panoramix.py 0x06012c8cf97bead5deae237070f9587f8e7a266d'.
See panoramix.py for the list of shortcuts, feel free to add your own.
- core - modules for doing abstract/symbolic operations
- pano - the proper decompiler
- utils - various helper modules
- tilde - the library for handling pattern matching in python3.8
- cache_code - cached bytecodes
- cache_pan - cached decompilation outputs
- cache_pabi - cached auto-generated p-abi files
- supplement.db - sqlite3 database of function definitions
- supp2.db - a lightweight variant o the above
Cache directories are split into subdirectories, so the filesystem doesn't break down with large amounts of cached contracts (important when running bulk_decompile on all 2.2M contracts on the chain)
All of the above generated after the first run.
bulk_decompile.py - batch-decompiles contracts, with multi-processing support bulk_compare.py - decompiles a set of test contracts, fetches the current decompiled from Eveem, and prepares two files, so you can diff them and see what changes were made
Panoramix uses a ton of pattern matching operations, and python doesn't support those as a language.
There are some pattern-matching libraries for older python versions, but none of them seemed good enough. Because of that, I built Tilde, which is a language extension adding a new operator.
Tilde replaces '~' pattern matching operator with a series of ':=' operators underneath. Because of that, python3.8 is a must.
Believe me, I spent a lot of time looking for some other way to make pattern matching readable. Nothing was close to this good.
But if you manage to figure out a way to do it without Tilde (and maintain readability), I'll gladly accept a PR :)
See the source code comments, starting with panoramix.py. Also, those slides[tbd].