Feature/support list element validation for all constraints

[pelletier197]

Basically, every directive now has to specify if it supports list elements. If yes, instead of receiving the list of elements as an arguments, it receives each element individually. Which makes the validations easier to write, and gives better errors (specific index instead of global list error).

The only thing with that solution is that two constraints didn't work well like that. @NotEmpty and @Size could work for both strings and lists. In order to avoid confusion and make it possible to support both constraints on an input, I added two new constraints.

• @ContainerSize
• @ContainerNotEmpty
  which will only perform the validations on list inputs, while
• @Size
• @NotEmpty
  were modified to work only with strings, and IDs

I also simplified a little bit the directive code to reduce dupplication.

Resolves: #31

[pelletier197]

@bbakerman if you want to have a look at this. Sorry about it, it's quite large, but I think this is a good addition 😉.

[bbakerman]

Prove to me why the null value hanlding code (which returns true) is sensible to remove.

You have removed it all over the place. Why is that ok?

[bbakerman]

I thank you for this PR - its extensive to say the least. Smaller PRs usually have more success in open source terms.

I am musing on the breaking change here where @SiZe no longer does lists. I still need to be convinced this is a good idea.

Also the null handling as outlined above

[pelletier197]

I agree, I rarely do big PRs in open source. I tried to do too much at once. If it's better for you, I can probably split it in 2 or 3 PRs.

Regarding the change of size, here's the issue I wanted to fix. When you have a schema like this.

field( arg : [ String ] @Size(min: 2)) : ID

It would have been possible to implement only the list should have a size of min 2 or only the strings in the list should have a size of min 2, but not both at the same time. And the user can't really choose which one he wants.

Now, what I proposed is

field( arg : [ String ] @Size(min: 2) @ContainerSize(min : 5)) : ID

Which would say only lists of at least 5 strings, and the strings must have at least a size of 2.

Now maybe this change was not required. It may be an edge case, but it felt important to handle something like this. Or maybe there is another way I haven't seen?

[pelletier197]

@bbakerman to come back on this, what to do you think about my justification? If you don't like it, what do you propose?

And do you wish me to split this PR in smaller chunks?

[pelletier197]

I have had no update with this for a while... I'm currently facing the issue that I have to validate both the container size and the string size. I created another constraint just for that as a work-around, but it would be really great is something like this could be integrated in the library...

Does anyone have a better solution if the proposed one is not okay with you?

[bbakerman]

Sorry for the delay on this PR - I have not been able to put the time towards this repo that it deserves.

I am happy to accept this PR and have the change in behavior

Thank you for such an extensive PR with great quality

[pelletier197]

Thanks a lot!