Fix :insert_mode => :button bug

[jeffcarbs]

This is a fix for #23 (cc @bmichotte)

Setting :insert_mode to :button was causing a crash. Here was the issue:

This gem was monkey patching sectionOptions to return the Ruby object's options instance variable. However, sectionOptions was being called in the Obj-C section initializer (XLFormSectionDescriptor.formSectionWithTitle) BEFORE we were able to set options on the ruby object.

Why was this only an issue with the :button option? Well in that Obj-C initializer there's a call to canInsertUsingButton:

• https://github.com/xmartlabs/XLForm/blob/master/XLForm/XL/Descriptors/XLFormSectionDescriptor.m#L85
• https://github.com/xmartlabs/XLForm/blob/master/XLForm/XL/Descriptors/XLFormSectionDescriptor.m#L322-L325

That method is a two-part conditional. The first is is the insert mode set to button? and the second is do the section options indicate that I can insert?. The second part was where the crash was occurring but it was never getting there because the first part was evaluating to false.

The changes here:

• Stop monkey-patching sectionOptions. I couldn't see any reason this gem needed to be doing that.
• Took out the setter/getter for section#options. The custom setter was unnecessary since we're already parsing the options in the only spot where we're calling that setter. And then if we don't need a custom setter, both the setter/getter are redundant because options is an attr_accessor.

[bmichotte]

I'm ok for everything, except the sectionOptions. I added this override mainly for reasons like #3 where we need a way to change the options elsewhere (look at the exemple with the gif) and I didn't think using sectionOptions with the objc values is a good thing.

Maybe, we can keep the options as an alias for sectionOptions with something like

 def options=(value)        
   self.sectionOptions = self.class.parse_section_options(value)        
end     
alias :options :sectionOptions

What do you think ?

[jeffcarbs]

Ah I didn't consider modifying the options on-the-fly. I think I've got something working based on your suggestion. Going to clean it up and push and update.

[jeffcarbs]

@bmichotte - check the updated code. I've tested locally that it works with the example you gave in #3 and it still fixes the main bug 👍

[bmichotte]

Thanks @jcarbo !