Openapi parser #22
Openapi parser #22
Conversation
| @@ -43,6 +43,7 @@ type Op string | |||
| const ( | |||
| MinLengthOp Op = "minLength" | |||
| MaxLengthOp Op = "maxLength" | |||
| MultipleOfOp Op = "multipleOf" | |||
There was a problem hiding this comment.
Do we want to support this operator? 🤔
There was a problem hiding this comment.
I tried to support the things that exist in open-api. I think that we aren't using that in the cue schemas.
| @@ -97,7 +97,8 @@ func (jenny RawTypes) formatEnumDef(def ast.Object) string { | |||
|
|
|||
| buffer.WriteString("const (\n") | |||
| for _, val := range enumType.Values { | |||
| buffer.WriteString(fmt.Sprintf("\t%s %s = %#v\n", tools.UpperCamelCase(val.Name), enumName, val.Value)) | |||
| name := tools.CleanupNames(tools.UpperCamelCase(val.Name)) | |||
There was a problem hiding this comment.
I'm surprised to see "enum value name" cleaning/formatting here.
It feels like this should be done somewhere else since we're likely to do it for any target language: in a compiler pass such as this one maybe? (or another one)
There was a problem hiding this comment.
I found this issue when we have integers. It was setting things like:
0 Cursor = 0
1 Cursor = 1I set the enum name as prefix to avoid invalid values but fon any reason, after do some sync it wasn't necessary anymore 😅.
This cleanup is because I was setting the values of the enum as names and there were some with punctuation marks generating invalid names.
There was a problem hiding this comment.
Makes sense!
If we want to keep this kind of cleanup, I still think it should be done as a compiler pass though, instead of directly within the jennies :)
|
|
||
| func getArgs(v *float64, t string) []any { | ||
| args := []any{v} | ||
| if t == openapi3.TypeInteger { |
There was a problem hiding this comment.
Does that mean that if t != openapi3.TypeInterger, then args will contain a pointer to a float?
If so, I think we should dereference the pointer first: args := []any{*v}
There was a problem hiding this comment.
Yes, if it isn't an integer, we want to retrieve the float value. These values are *float4 by default and if we don't map them to integers when the type is integer, the builder will compare integers with floats.
|
|
||
| func (g *generator) walkAnyOf(schema *openapi3.Schema) (ast.Type, error) { | ||
| return g.walkDisjunctions(schema.AnyOf) | ||
| } |
There was a problem hiding this comment.
allOf, oneOf and anyOf have different semantics, we can't treat them as just a disjunction. See: https://swagger.io/docs/specification/data-models/oneof-anyof-allof-not/
I think that for the time being, it's okay if we only support oneOf and explicitly return an error if a anyOf or allOf is found.
But even for oneOf, we still should read and extract the discriminator data that might be there: https://github.com/OAI/OpenAPI-Specification/blob/main/versions/3.0.3.md#discriminator-object
There was a problem hiding this comment.
I put them as the same way from the beginning but its something that I wanted to do it later 😅. But yep, its wrong
|
Tests? 😇 |
|
Let's merge it! |
Contributes to: #4
First approach for openapi parser