Add boolean prql.addTargetDialectToSqlFilenames setting for the generated SQL filenames

[RandomFractals]

to enable users targeting different DBMS systems to create .sql files with the target db/platform filename suffix.

For example, sql file generated from some invoices-total.prql for the duckdb target will be named invoices-total.duckdb.sql.

The same user can then change prql.target to SQLite, create and save somewhat different invoices-total.sqlite.sql file and add it to their project version control system.

In both scenarios we just add sql dialect suffix before .sql file extension name when generating sql files via #42 feature.

To keep it simple, we'll default this setting to false, and advanced users can change it in their workspace settings.

Note: we'll do this for all the supported SQL system dialects, and skip adding this sql filename suffix when the current workspace prql.target setting is Generic or None recently added in #98.

[RandomFractals]

Example of using this new setting:

[max-sixty]

My guess is that this won't be used that much — folks generally have a single DB they're working with. But no harm in adding

(does VSCode share any telemetry so we can see who is using what settings?)

[RandomFractals]

I can see folks using some transactional db and duckdb for analytics use this feature in some form or the other.

We can set up telemetry later. I don't think we have enough active users to start logging what features are used, or how many times sql preview is shown.

VSCode does have standard telemetry api, but that might require setting up Azure service, or do it with google analytics. i.e. much larger want/ask that should be written down as a separate ticket for us to consider adding it.