fix: assert condition command timeout parsing

[rwtnb]

Proposed changes

Removing private access modifier from timeout field

Issues fixed

Private modifier on timeout field is preventing the field to be populated on deserialization.

[amanjeetsingh150]

I'm not too sure how changing scope of variable fixes the issue?

Can you write a test for it as well, so that it not breaks again?

[felipeduartea]

no sure how this test would look, since it happens only on Robin. I can add a the test there two if you want

[herval]

I'm not too sure how changing scope of variable fixes the issue?

Can you write a test for it as well, so that it not breaks again?

private variables don't get filled out by json parsing. If this is being parsed (it seems to be the case), it'll always be null

[amanjeetsingh150]

I feel I might be missing something, the deserialization of Yaml commands to Maestro commands works through same deserializer here in YamlCommandReader for both Robin and CLI.

If it's working on CLI and not on Robin, it sounds like we must be doing something wrong on Robin. Not sure it's clear from here what.

[felipeduartea]

it seams ot me that this is due to some library version that doesn't parse the json of private variables anymore, as the code implementations is the same in both codes. Still figuring out which one is it. Should I proceed with the release or wait till this is figured out?

I think it probably makes sense to release it that way, since the condition variable already is not private, btw do you have context on that? @amanjeetsingh150

[amanjeetsingh150]

You can go ahead with the release no worries!

But since that can be an issue with private variables. I think more importantly let's ensure there are no other private variables in maestro commands that could also face this issue.

On the version thing probably it makes sense to confirm that we are using same version of serialzers for robin for jackson or some dependency is not missing?

[felipeduartea]

But since that can be an issue with private variables. I think more importantly let's ensure there are no other private variables in maestro commands that could also face this issue.

Checked this, it seems that this was the last private val that we had on the MaestroCommand

On the version thing probably it makes sense to confirm that we are using same version of serialzers for robin for jackson or some dependency is not missing?

Also think that it's something related to this, will take a look as soon as I free up

[Fishbowler]

Rebased

[Fishbowler]

How do you reproduce the problem?

[Fishbowler]

Locally, I tried (using the release version):

appId: com.example.example
---
- launchApp
- extendedWaitUntil:
    visible: "2"
    timeout: 60000

Which times out after 60s, as it should.

Is the serialisation problem only a problem for Robin?

[felipeduartea]

Is the serialisation problem only a problem for Robin?

Yeah, this issue only happens in Robin serialization, since the timeout var is private. I created 3 .yaml files and tested them on Robin. They were all failling before and not recognizing the timeout, but with the change this is not happening anymore.

[herval]

we should probably review all the commands to see which ones have private vals and make them public, since this might be leading to other issues we're not seeing yet