Problem when using camelcase fieldnames in fields filter on CollectionSerializer

[MRigal]

Dear @onyg ,

First of all, thanks for the fantastic work you made by writing this serializer. It is really cool and would deserve to be much more known !

Unfortunately, I found a bug in the DjangoCollectionSerializer.

When you try to use it in combination with the fields argument and that you pass some camelCase fieldnames in the fields list, these won't appear in DjangoCollectionSerializer output.

Tracking down a bit the error, passing the underscored fields we could find that self._fields equals to [u'kid', u'is_visible_in_charlottenburg'] inside init(), but later, inside dump() it gets equal to [u'kid', u'isVisibleInCharlottenburg']. This is the name the field should have in the Json output, but unfortunately it is this list which is passed to the real (item) Serializer, which gets called from the _generate() inside the dump() method.
Therefore the serializer can't map the camelCase field name to the actual name defined in the serializer.

Thanks in advance for fixing it and continue writing good code !
Matt

EDIT: I precise that I'm using the DjangoCollectionSerializer, giving the request as parameter

[onyg]

Hi @MRigal ,
the DjangoCollectionSerializer with the DjangoRequestMixin overrides the fields when a fields argument is passthru the GET querystring. I will changes it very soon.
Thanks

[MRigal]

Oh thanks ! Indeed, it was taking the fields from the get request which were camelCase !

I could just stop to pass the request to the DjangoCollectionSerializer and this was avoiding the overwriting of the fields !

Thanks !