This is an example implementation of a standalone UI for the geth-based signer.
It starts the signer binary as a subprocess, and communicates with the signer over standard input/output.
Whenever the signer needs confirmation or input from the user, it will invoke the UI.
The UI is responsible for displaying the relevant info to the user, and letting the signer know if the operation is approved or not, and also the password.
The idea is to have standalone signer , which uses it's own trusted UI.
The signer has a native command-line UI -- but that's very limited. A more advanced UI can make modifications to transactions, instead of just answering Approve/Reject. Also, people seem to favour graphical UIs over commandline UIs nowadays.
As of right now, this is highly untested code, more like a proof-of-concept than a finished thing. It currently only handles transactions, and rejects all other requests, such as
- Listing accounts
- Importing/exporting accounts
- Signing data
- etc...
It currently has a systray icon, but there's not anything connected to it. It currently has no packaging/bundling, which should be provided as part of a proper install.
Also, the code could be more pythonic and nice. All hands appreciated!
A few screenshots:
