console, cli, api fixes

[zelig]

This PR addresses several minor bugs, regressions, inconsistencies relating to geth UI including console, cli, rpc api

• console: bash history behaviour: lines with leading space are ommitted from history. fixes leading whitespace should not save to geth console history #1697
• console: exit processed even with whitespace around
• console: all whitespace lines (not only empty ones) are ignored
• registrar: methods now return proper error if reg addresses are not set + tests. fixes admin registrar commands always returning true #1457
• registrar: fix major regression: deploy registrars, register HashsReg and UrlHint in GlobalRegistrar. set all 3 contract addresses in code
• rpc/console: fix major [unreported] regression. Now all comms accept interactive password for personal.newAccount()
• rpc/console: fix regression with docRoot / admin API (natspec example on wiki)
• cli/accounts: unlock multiple accounts (with inline passwdfile). fixes accounts: unlock multiple accounts fails #1785
• crypto (CLI): catch AES decryption error on presale wallet import + fix error msg format. fixes Same derived address with any password when importing presale wallet? #1580
• CLI: improve error message when starting a second instance of geth. fixes improve error message when starting a second instance of geth #1564
• common/natspec: fix end to end test
• common/docserver: catch http errors from response

[robotally]

Vote	Count	Reviewers
👍	1	@bas-vk
👎	0

Updated: Wed Oct 21 15:04:14 UTC 2015

[bas-vk]

@zelig, review done.

[bas-vk]

This function doesn't do any real post processing and the returned string is not actually necessary. Consider changing it to something like func mustLogInHistory(input string) bool.

[zelig]

how about forgetting this automation entirely, pass interactivity as a parameter to modules and simply enforce interactive use on the console(s).

[bas-vk]

I think @ethers (#1697) asked for this feature not to hide passwords but to prevent dangerous commands not to end up in the history from which they can accidentally be executed again when the users browses through history. That means that besides password interaction also transaction confirmation from the console should be enabled since transactions will have irreversible consequences.

[zelig]

@bas-vk I agree with all points, thanks for the thorough review

[codecov-io]

Current coverage is 48.20%

Merging #1840 into develop will decrease coverage by -0.11% as of 4fa4336

Powered by Codecov. Updated on successful CI builds.

[bas-vk]

This still doesn't work.

ethdb.NewLDBDatabase will return a syscall.Errno error for which you can check. In backend.go these errors are converted to an *errors.errorString instance.

chainDb, err := newdb(filepath.Join(config.DataDir, "chaindata"))
if err != nil {
    return nil, fmt.Errorf("blockchain db err: %v", err)
}

The check in main.go for an errno wil then obviously fail.

Second problem is that the check is only added to the run subcommand. Users who will start the console (or any other subcommand) will still get the old error message when another instance already claimed the database datadir.

[bas-vk]

What did you end up doing for #1697?

[bas-vk]

This is probably a debug line and can be removed?

[bas-vk]

Apart from the minor issue I'm ok 👍