[NADE][Bugfix] Use resolved app warehouse for regionless query

[sfc-gh-bgoel]

Pre-review checklist

• I've confirmed that instructions included in README.md are still correct after my changes in the codebase.
• I've added or updated automated unit tests to verify correctness of my new code.
• I've added or updated integration tests to verify correctness of my new code.
• I've confirmed that my changes are working by executing CLI's commands manually on MacOS.
• I've confirmed that my changes are working by executing CLI's commands manually on Windows.
• I've confirmed that my changes are up-to-date with the target branch.
• I've described my changes in the release notes.
• I've described my changes in the section below.

Changes description

As part of fetching the URL where the Native App is live, the CLI needs to execute a SQL query which requires the use of a warehouse. The application.warehouse property in the manager class resolves this to either be the app warehouse if found, or the one derived from the connection. Since this URL is pertaining to the app instance, we have decided to execute this query by only activating app warehouse.

[sfc-gh-cgorrie]

Can you also audit other potential gaps we might have missed -- like post-deploy scripts or package scripts? We should use this new use_warehouse primitive everywhere it makes sense.

[sfc-gh-bgoel]

This error message is now different because instead of calling use warehouse directly which uses our generic_error_handler, it calls the method self.use provided by the CLI which has its own error handling.

[sfc-gh-melnacouzi]

this logic is a bit too complicated.

1. Why do we need to stay with prev_wh if new_wh is not defined? is it a valid use case that new_wh is None (especially when dealing with snow app run)? Shouldn't we just error?

2. too many combinations. I suggest to simplify as follows if possible:

a) if new_wh is None, error out.
b) proceed to switch to new_wh if different from prev_wh.
c) yield
d) if prev_wh was None, do nothing, otherwise, switch back to it.

[sfc-gh-bgoel]

Some concerns raised in the previous comments:

1. It is possible that prev_wh is None, especially in the case when no default warehouse is assigned to a user, and hence current_warehouse() may return None. I do not have a way of verifying this as I do not have permissions to set up a new user in an account, but I do believe this is a case we need to account for since the bug ticket was created for a scenario like this when there is no fallback.
2. It is possible that new_wh is None. In our case, package_warehouse or application_warehouse could be None if user has not provided that warehouse in their snowflake.yml, config.toml and there is no default warehouse. In that case, the connection context would return a None for this property.
3. Given the two above, we do need to tackle their nullity accordingly.
   But I am open to hearing if the assumptions in 1 and/or 2 are incorrect and therefore do not need to be handled.

[sfc-gh-cgorrie]

I'm with Michel, I think we can simplify it as suggested. Sent you a way to determine if warehouse is not set.

[sfc-gh-cgorrie]

My requests: account for None in the project model warehouse properties, and simplify the use warehouse context manager as suggested by Michel.

[sfc-gh-cgorrie]

I'm with Michel, I think we can simplify it as suggested. Sent you a way to determine if warehouse is not set.

[sfc-gh-cgorrie]

LGTM