Use https:// for XMR explorer API endpoints, except if localhost or Tor

[wiz]

• If Tor *.onion hostname, use HTTP with Tor proxy
• If 127.0.0.1 or localhost, use HTTP without Tor proxy
• If LAN address or *.local FQDN, use HTTP without Tor proxy
• If any other FQDN hostname, use HTTPS with Tor proxy

[chimp1984]

utACK

[ghost]

As discussed, suggested change to simplify: https:// prefix included in URL and then only need to detect localhost so as to disable socks5 proxy. Regex changed to allow optional port specifier. Defaults needs changing to include the https. @Emzy service to use https also?

    httpClient.setBaseUrl(model.getServiceAddress());
    if (model.getServiceAddress().matches(".*127.0.0.1.*|.*.local.*|.*localhost.*")) {
        log.info("Ignoring Socks5 proxy for local net address: {}", model.getServiceAddress());
        httpClient.setIgnoreSocks5Proxy(true);
    }

[wiz]

@jmacxx after thinking about it a bit more, that still isn't nearly enough to cover all the possible cases we need to support. let me work on some new regex.

[Emzy]

utACK
I'm not a JAVA developer. The regex as far as I can parse them by eye look OK, also the tests.

Is there not a simpler way to detect a private IP address space (LAN address)?

[ghost]

Problem: each time the user edits a url then restarts bisq, modify the url again the prefix is prepended, ends up like https://https://node77.monero.wiz.biz. I suppose any existing prefix has to be removed before validating the url.

[chimp1984]

maybe easiest to just add the https to the textfield at input.

[wiz]

@jmacxx good catch, I forgot to strip the protocol prefix out, just pushed a commit that should fix that issue. I should have tested more, but I was trying to rush this out.

@chimp1984 maybe, and there is also the decision to route the connection through Tor or not, which would require its own slider switch - using regex we can strictly enforce https or Tor onion except for localhost or LAN ip address. IMO the goal is to let user be able to add xmrchain.net or whatever public explorers they want to trust, but without letting them accidentally put http:// and expose them to MITM attacks.

I'm also going to tweak the strings a bit with @m52go, since "Service addresses" is a big vague and user might be confused what exactly to enter. To be honest this feature feels like it needs a lot more testing and polish before we ship it in a public release to users...

[ghost]

ACK 👍 Tested out OK in both regtest and production.

[chimp1984]

I will review tomorrow, was out the whole day....

[wiz]

@jmacxx thanks for testing with your http:// node on LAN with custom port number, this is exactly what we needed. I suppose a good test is to paste in something like this and watch how it auto-parses it:

http://monero3bec7m26vx6si6qo7q7imlaoz45ot5m2b5z2ppgoooo6jx2rqd.onion, http://wizxmr4hbdxdszqm5rfyqvceyca5jq62ppvtuznasnk66wvhhvgm3uyd.onion, https://xmrchain.net, http://192.168.1.1:8000, https://xmrchain.net:443, http://raspberrypi.local:8081

[wiz]

@chimp1984 after this PR, what do you think about making the list of Monero Explorers into a multi-line combobox like the ones for altcoins and fiat below? we could have a list of hard-coded ones enabled by default, and allow the user to enter their own into the list, would look nicer than a comma delimitated text input field

[sqrrm]

Code looks fine to me. Can't say the same for the regexes, but they might well be correct, I can't tell. If I can get another ack for regexes I'm fine with this PR. @Emzy that seems to be your forte :)

[chimp1984]

@wiz Which one are candidates? We could add one or 2 by default for more security. Only potential issue is that more services we use the higher the chance that one has a network issue and then auto-conf fails. Beside that, are those offering onion addresses and if not are they supporting Tor connections (no captcha)?

I think a UI list is a bit too much for that feature as it would also overload more the settings UI as it is already...

[wiz]

I agree, and Devin is going to setup a 3rd node instance for us so we should be fine for now. But that is not related to this PR i suppose, and we can think about it again later.

[Emzy]

utACK

Code looks fine to me. Can't say the same for the regexes, but they might well be correct, I can't tell. If I can get another ack for regexes I'm fine with this PR. @Emzy that seems to be your forte :)

The regexes and the networks to be local addresses look right. But I'm not a JAVA dev.
I haven't run the tests, but they also look good.

[Emzy]

ACK
The test passed.

[chimp1984]

utACK