feat(codec): molecule friendly error message

[IronLu233]

#396

[vercel]

The latest updates on your projects. Learn more about Vercel for Git ↗︎

Name	Status	Preview	Updated
lumos-website	✅ Ready (Inspect)	Visit Preview	Sep 14, 2022 at 9:59AM (UTC)

[codecov]

Codecov Report

Merging #403 (5d1cbe0) into develop (cc69d90) will increase coverage by 0.14%.
The diff coverage is 98.50%.

@@             Coverage Diff             @@
##           develop     #403      +/-   ##
===========================================
+ Coverage    81.98%   82.13%   +0.14%     
===========================================
  Files           93       94       +1     
  Lines        19105    19269     +164     
  Branches      1794     1820      +26     
===========================================
+ Hits         15664    15826     +162     
- Misses        3400     3401       +1     
- Partials        41       42       +1     

Impacted Files	Coverage Δ
packages/codec/src/error.ts	94.64% <94.64%> (ø)
packages/codec/src/high-order/nested.ts	99.17% <100.00%> (-0.83%)	⬇️
packages/codec/src/molecule/layout.ts	99.74% <100.00%> (+0.65%)	⬆️
packages/codec/src/number/uint.ts	99.23% <100.00%> (+0.10%)	⬆️
packages/codec/src/utils.ts	96.07% <100.00%> (+1.01%)	⬆️

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update cc69d90...5d1cbe0. Read the comment docs.

[PainterPuppets]

Can you add description and purpose about this pull request?

[PainterPuppets]

I'm wondering if it's possible to use the higher-order function to do non-intrusive modifications instead of directly changing the codec?

maybe example like that

withReadableError(array)(xxxx)

[IronLu233]

I'm wondering if it's possible to use the higher-order function to do non-intrusive modifications instead of directly changing the codec?

maybe example like that

withReadableError(array)(xxxx)

It is difficult.

We want track the property access path in codec pack input, but these key only appears in Array.prototype.map/reduce
methods.

When code run out of these high order function, we can not get the current key anymore. so I think we have to modify some code in codec pack logic.

[homura]

When code run out of these high order function, we can not get the current key anymore. so I think we have to modify some code in codec pack logic.

vote for this. but i want to know it is possible to refactor createObjectCodec and createArrayCodec to provide the friendly message? because it's not just molecule that can be affected by nested error. for struct/table/array/vector will call the createObject|ArrayCodec to report the friendly message

[IronLu233]

When code run out of these high order function, we can not get the current key anymore. so I think we have to modify some code in codec pack logic.

vote for this. but i want to know it is possible to refactor createObjectCodec and createArrayCodec to provide the friendly message? because it's not just molecule that can be affected by nested error. for struct/table/array/vector will call the createObject|ArrayCodec to report the friendly message

Maybe ES6 proxy class can implement this feature.

[homura]

Maybe ES6 proxy class can implement this feature.

it can be simpler

function struct(shape, fields) {
  const codec = createObjectCodec(...)

  return {
    pack(packable) {
      const packedFields = fields.map(field => codec[field].pack(packable[item]))
      return concat(...packedFields) 
    }
    unpack() {
      // ...
    }
  }
}

[IronLu233]

Maybe ES6 proxy class can implement this feature.

it can be simpler

function struct(shape, fields) {
  const codec = createObjectCodec(...)

  return {
    pack(packable) {
      const packedFields = fields.map(field => codec[field].pack(packable[item]))
      return concat(...packedFields) 
    }
    unpack() {
      // ...
    }
  }
}

I found a way to track layout struct(array, table) error without modify it's codec implementations.

This can be in another PR in #408

[IronLu233]

Seems everything is OK. Can you please help review again? @homura