Only write options after onboarding

[johnny9]

[hebasto]

Needs rebase.

[johnny9]

update from 45e3262 to 08be85b:

• rebased over main

[thebagmaster]

ACK 08be85b
I tested these on Ubuntu 22.04.2 and it does what it says on the tin.
Understood that we may need to buffer the settings writing to file in on-boarding; especially if the storage location is not set and may change.
Maybe we should do our best to save what we can of what the user configured?
Are there situations where we buffer settings and different settings that we save immediately?

[hebasto]

Please rebase.

https://discord.com/channels/903125802726596648/903126039084015638/1146922061449666602

[johnny9]

Update from 08be85b to 51df97d

• rebased with main

[jarolrod]

Concept ACK

[jarolrod]

potentially, a bool indicating if we're onboarded or not could be valuable to other models. Since right now we don't have such a use case, this is fine.

[D33r-Gee]

tACK on WSL Ubuntu 22.04 on signet

Works as expected, the settings file gets created after the onboarding has finished.

settings.json:

{ "listen": true, "natpmp": true, "prune": 2861 }

Do we want the “Listen” and “natpmp” options to be true by default or was it just for testing this PR?

[D33r-Gee]

tACK on Android armv7 everything worked as expected

The settings.json was created after the onboarding with the following values:

{ "prune": 2861 }

[D33r-Gee]

ACK 51df97d

[pablomartin4btc]

Concept ACK, I'll test soon. It looks that there are some unaddressed feedback.

[pablomartin4btc]

tACK 51df97d

Tested in both master(/ main) and this PRs branch: changing all settings touched by this branch's code change and verify the changes were saved accordingly.

Things I noticed during testing:

In main:

• If the user doesn't change anything setting.json file will look like:

{
    "prune": 1907
}

• If the user disables any settings setting.json file will look like:

{
    "prune": 0
    "listen": false,
    "natpmp": false,
    "server": false,
    "upnp": false

}

In branch johnny9:write_after_onboard:

• If the user doesn't change anything setting.json file will look like:

{
    "dbcache": 16000,
    "listen": true,
    "natpmp": true,
    "prune": 1907,
    "server": true,
    "upnp": true
}

• If the user disables any settings setting.json file will look like:

{
    "dbcache": 16000,
}

After the user makes any changes on the settings mentioned above next time the user starts `bitcoin-qt` up again it looks like the UI doesn't match the configuration file (`settings.json`), is this expected?

• Using main:

• Using this branch:

[johnny9]

Update 51df97d to 61f8e6c

• Rebased with master
• Initialized all uninitialized OptionsQmlModel properties

[johnny9]

In branch johnny9:write_after_onboard:

• If the user doesn't change anything setting.json file will look like:

{
    "dbcache": 16000,
    "listen": true,
    "natpmp": true,
    "prune": 1907,
    "server": true,
    "upnp": true
}

• If the user disables any settings setting.json file will look like:

{
    "dbcache": 16000,
}

After the user makes any changes on the settings mentioned above next time the user starts bitcoin-qt up again it looks like the UI doesn't match the configuration file (settings.json), is this expected?

@pablomartin4btc I wasn't able to reproduce the output you were seeing during onboarding but I think I fixed the settings not applying at next startup with initializing the values in the constructor.

[D33r-Gee]

tACK 61f8e6c

On Ubuntu 22.04 and Android (armv7)

Works as expected, settings.json shows defaults:

{
    "listen": true,
    "prune": 1907
}

[pablomartin4btc]

tACK 61f8e6c

It behaves as expected.

dbcache setting is not persisted anymore ("dbcache": 16000,), is this correct?

[johnny9]

dbcache setting is not persisted anymore ("dbcache": 16000,), is this correct?

It should only persist if you set it to a value other than nDefaultDbCache (450)

[pablomartin4btc]

dbcache setting is not persisted anymore ("dbcache": 16000,), is this correct?

It should only persist if you set it to a value other than nDefaultDbCache (450)

Perfect, it works.

[johnny9]

Update from 61f8e6c to 4f557b0

• rebased with main

[johnny9]

Update from 4f557b0 to 9e56907

• fixed build issue with previous rebase conflict (forgot to remove <<<<<<<)

[D33r-Gee]

tACK 9e56907

lgtm, works as expected on both Ubuntu 22.04 and Android (armv7)

played around with different settings and they registered properly both in the settings.json and in the debug.log

[pablomartin4btc]

re-ACK 9e56907

[MarnixCroes]

9e56907
writing of the settings works, but it always additionally creates a settings.json.bak file?

[pablomartin4btc]

Do we need this?

Suggested change

#include <qt/optionsmodel.h>

[johnny9]

It is needed for PruneMiBtoGB i believe

[pablomartin4btc]

nit: accidentally removed the last line break?

Suggested change

	}
	}

[johnny9]

already has a newline character at the end of the file. don't need an additional blank line.

[D33r-Gee]

re ACK 9e56907 on Ubuntu 22.04 works as expected