Fixes bug in string case serialization not taking numbers into account

[matt035343]

As described in #376, the camelCase serialization does not remove underscores in front of numbers. I believe it should - let me know what you think.

[george-zubrienko]

You can run into issues with this when having fields named a_1_1 vs a_11 - both will have the same name in Camel. I think underscores should be replaced by any a-Z char instead OR you should check if camel will produce duplicate fields when applied, and runtime fail the serialisation.

[matt035343]

You can run into issues with this when having fields named a_1_1 vs a_11 - both will have the same name in Camel. I think underscores should be replaced by any a-Z char instead OR you should check if camel will produce duplicate fields when applied, and runtime fail the serialisation.

imo, I think adding arbitrary chars is worse than underscores as it may also cause collisions with the original name. Then I think it is a better to raise an exception. Also, it is possible to provide custom field names in the metadata to manually avoid the conflicts.

[george-zubrienko]

a better to raise an exception

Agreed :)

[george-zubrienko]

@matt035343 any chance you have time to add exception in case duplicate is produced and we merge? :)

[matt035343]

@matt035343 any chance you have time to add exception in case duplicate is produced and we merge? :)

Exception is now added, please take a look

[george-zubrienko]

Awesome!