Made json encoding/decoding configurable + Aeson compatible encoding/decoding #12
Conversation
Needs more testing & cleanup before being ready for a PR
Tests for existing functionality passing. Tests for aeson encoding missing - coming soon.
|
@hdgarrood could you take a look at this perhaps, if you have time? You're definitely more familiar with Aeson than me. 😄 |
|
|
||
|
|
||
| allConstructorsNullary :: Array DataConstructor -> Boolean | ||
| allConstructorsNullary constrSigns = foldr (&&) true <<< map (\c -> null c.sigValues) $ constrSigns |
There was a problem hiding this comment.
Is foldr (&&) true the same as Data.Foldable.all?
There was a problem hiding this comment.
Ok, yeah, I think this could be written allConstructorsNullary = Data.Foldable.all (null <<< _.sigValues)
There was a problem hiding this comment.
Yeah - I missed Data.Foldable.all during my Pursuit search. Thanks!
|
I tried to make it backwards compatible. So gEncodeJson became a simple wrapper around genericEncodeJson which takes an Options parameter resulting in the same format previously used. |
|
Btw. guys - you are amazing. Sending PRs to purescript projects is so much more fun, than for nixos. Answers in seconds, instead of weeks/months. I don't know how you do that, but it is really motivating. |
|
This looks good to me. Regarding Aeson specifically, I'd just say, if it works, go ahead :) |
|
I do it by happening to be on github at the same time as you, and not having many other commitments right now 😄 it is nice to know that it's appreciated, though. Incidentally, that (not having many commitments) probably won't remain the case for very much longer for me. |
|
@hdgarrood congrats? 😉 @eskimor I'm in the lucky position to work with PureScript full time now, and especially for libraries like this that we (SlamData) use it means I try to stay on top of things. I don't always manage it though! |
|
Code looks fine to me too, 👍 but I'd appreciate it if we could have at least a few tests for the Aeson side of things before merging, perhaps just a basic roundtrip encode/decode test for each sig case? |
|
@garyb It's maybe a little early for that, but cheers anyway :) |
|
@garyb : Done :-) I also checked the output - it looks good and aeson compatible. |
|
It depends how good the generators are 😄 but yep, it's wonderful for this kind of testing! Thanks 👍 |
|
If you don't mind I'll hold off merging this until we do the big psc 0.8 library update (should be really soon now!), as technically this is a breaking change (dropping the primed exports) so will mean bumping all the dependants too. We can add the |
|
Ok - no problem, I will use my local version in the meantime. |
|
Glad you have not merged it yet - I added two more tests and indeed discovered that it was not quite right yet. Ouch! Ouch! Ouch! It was too good to be true. |
|
Resolved issue #5. |
|
Bumped psc version in package.json to 0.8.0 - incorporated the unsafeCrashWith stuff. |
|
Hi there! Can this be merged now? Or is something still missing? |
|
We haven't finished updating the core libraries for 0.8 yet. I was intending to wait until then, as we'll need to bump all the dependency versions again when that's finished, which will be classed as another breaking change. |
|
Ah I get it now, you would release a new version of purescript-argonaut-codecs too when merging? So master and the latest version always stay the same? I see, ping me if anymore changes are needed for merging. |
Support for special cases
Tests pass.
|
Resolved issue #14 |
|
Huh - what's wrong with the test suite? |
Despite what the defaultOptions in aeson suggest, it works as if unwrapUnaryRecords was set to true.
eskimor
changed the title
|
Creating standalone package. |
in order to provide an aeson compatible encoding/decoding option. Also resolving issue #5.