In order to avoid disk thrashing for Accounts and HasAccount,
address->key file mappings are now cached in memory. This makes it no
longer necessary to keep the key address in the file name. The address
of each key is derived from file content instead.
There are minor user-visible changes:
- "geth account list" now reports key file paths alongside the address.
- If multiple keys are present for an address, unlocking by address is
not possible. Users are directed to remove the duplicate files
instead. Unlocking by index is still possible.
- Key files are overwritten written in place when updating the password.
The account management API was originally implemented as a thin layer
around crypto.KeyStore, on the grounds that several kinds of key stores
would be implemented later on. It turns out that this won't happen so
KeyStore is a superflous abstraction.
In this commit crypto.KeyStore and everything related to it moves to
package accounts and is unexported.
* Add initial UserAccount and AccountManager structs
* Add NewAccount, Sign and Accounts functions
* Refactor key stores to use key address as main identifier
while keeping the UUID.
* Use key address as file/dir names instead of UUID
* Remove flags field from key struct
* Change JSON struct fields from string to []byte
* Change GenerateNewKey API to take io.Reader for random source
* Remove mixing entropy source function
* Use testing Fatal in tests
* Simplify scrypt constants with const block
* Add key store constructors and make their types private
* Simplify key store and file namings to be less Java Enterprise™
* Change test error logging to use t.Error(err)
* Reduce number of naked returns (just like my ex-gf)
* Simplify file reading path code
* Add new generic key_store interface
* Add new plaintext key store storing unprotected keys on disk
* Add new encrypted key store storing encrypted keys on disk
* Add new entropy mixing function using OS and go runtime sources