Accessing to package objects via #[]

[flash-gordon]

Ok, here we go :)

Keeping my (outdated) promise from #31 I refactored package.rb.

Best regards,
NS.

[jgebal]

Nice separation of responsibility. I like the way you got rid of the If/else

[jgebal]

@javornikolov, will you be able to review/approve?

[javornikolov]

Thank you @flash-gordon for the refactoring and thanks @jgebal for your review!

I can take a bit more detailed look by the end of the week, at first glimpse it looks good to me.

[flash-gordon]

@jgebal is it better now? Or did I miss something?

[jgebal]

@flash-gordon Thanks, I'll review it over the weekend - too tired to think clear now.

[jgebal]

Except the 2 comments added, the PR is really OK from my perspective.

[javornikolov]

OK from my side too.
I started wondering about the cases when we have low-case letters in the database object names (e.g. "MyPackage"), but that's a separate topic and is anyway an obscure practice.

[jgebal]

I'd give it a go with this pull request.
Case-sensitivity in Oracle is in general a bad practice I think. Just because you can use it, doesn't man that you should.
Since Ruby uses low case for variables, and Oracle, well - depends on company standards :), I'd allow the 'whatever case' approach for sting-based.
On the other hand, the framework is using symbols, not strings wherever possible. That seems a better approach.
@flash-gordon - any particular reason why you want to reference package objects by using the below syntax?
[code]
package_name['string_VAR_name']
[code]
would it be enough to use:
[code]
package_name[:symbol_var_name]
[code]

[flash-gordon]

@jgebal no, I'm fine with case-insensitive approach, just updated the PR. Also we can add string/symbol semantics for case-sensitivity (something like Sequel's implementation for column names).

[jgebal]

@javornikolov All tests are passing, Can you can merge and close this one?

[javornikolov]

Great, I'm merging. Thank you @flash-gordon for that contribution and thank you @jgebal for reviewing it!