JDC is not logging any error on wrong TP configuration

[GitGab19]

JDC is not generating errors when a wrong TP configuration is used.
To reproduce it, you can simply change the port number randomly in tp_address parameter, and run it.
A discussion about this here: https://discord.com/channels/950687892169195530/1012371967992672321/1229368703737663508

[Sjors]

Worse, it's not logging any config error (it took me half an hour to realise I had to rename authority_public_key to authority_pubkey).

It does log when the config file doesn't exist.

[GitGab19]

I would label this as a good-first-issue.
Maybe I don't remember correctly, but didn't we do something in the past to fix these type of behaviours @plebhash ?

[plebhash]

didn't we do something in the past to fix these type of behaviours?

can you be more specific? are you referring to wrong TP config, or authority_public_key vs authority_pubkey?

[GitGab19]

didn't we do something in the past to fix these type of behaviours?

can you be more specific? are you referring to wrong TP config, or authority_public_key vs authority_pubkey?

I don't remember specific details tbh. I just remember that we did something to better handle configurations in general, but maybe I'm wrong

[plebhash]

we had this issue: #695
which was closed via #859

[jbesraa]

Happy to pick this up

@GitGab19 From what I understand from the discord discussion the user used 18442 instead of 8442 in one of the configs, but that is not strictly prohibited right? I wonder if this requires validation across configs or we should just validate its 8442 across all of configs?

[GitGab19]

Happy to pick this up

@GitGab19 From what I understand from the discord discussion the user used 18442 instead of 8442 in one of the configs, but that is not strictly prohibited right? I wonder if this requires validation across configs or we should just validate its 8442 across all of configs?

Hey I'm sorry for being late.
So, imo the important thing is to validate the config in order to avoid cases like the one described here: #863 (comment).
The 8442 port is not strictly mandatory, a user could change it in both TP and roles config, so I would not require it to be always 8442

[Sjors]

Indeed, you could check that the port is a valid integer between 0 and 65535. But just checking that it's an integer should be enough.

[jbesraa]

Because we are using serde the ser and de errors are generated also by serde which is fine. so I didnt need to do much beside handling this error https://github.com/stratum-mining/stratum/pull/987/files#diff-49408d7d429f7afceda135baa9cdd786c20024ec6ffd2a05353efe1b0c5c2e75R110

[rrybarczyk]

See my comment here.

I think we should address just the JDC config issue as a patch if you think it is urgent, or wait until #997 which I just opened.

[GitGab19]

See my comment here.

I think we should address just the JDC config issue as a patch if you think it is urgent, or wait until #997 which I just opened.

I would lean towards aggregating them into #997, since it seems a better wrapper for it. But @jbesraa feel free to let us know what you think about it (for me it's also ok to merge your contribution now)

[GitGab19]

Closed by #987