Array Query Parameters Prevent Mocking
Originally posted by user brad.samuels to the Stoplight Community on 10/26/2022 at 19:59 ET. Hi, for some reason whenever I add a query parameter as an array, I cannot successfully submit a mock request. It appears with the error of "Cannot use param value {value}. JSON array expected". In principle, I understand what this means, but the Mock only provides a dropdown (which is already unusual for an array) to select one response. Is there a way I can fix the mock request for: More accurate validation (being a query parameter, it wouldn't make sense to be a JSON array as opposed to ?propertyTypes=house,unit Allow multiple values to be inputted to the field (since it is an array)Nested Masks don't seem to work
Originally posted by user MarcoDiSarno to the Stoplight Community on 03/07/2024 at 04:40 ET. I have the following use case for masking: Model A is referencing model B. (Garage.cars: array[Car] in the example) Then, a mask on the model A hides some fields and also needs to reference a mask on model B, since that also has some fields to mask. (GaragePublic.cars: array[CarPublic]) Stoplight lets me do that and I thought, "man, nice!" but when publishing the API, all these "nested masks" don't compile. Error: "Invalid JSON pointer: stoplight://resources/overrides/..." Plus, when I export the API with bundled references, it also just shows the "stoplight://resources/overrides/..." refs, which are useless outside of stoplight. The mock server also doesn't work for that project, I suspect for the same reason. Interestingly enough, it seems to work fine in the editor, just outside it doesn't.
Prism takes too long to start up
It takes quite a long time (8-10 minutes) to initialize Prism mock server. All that is logged during this initialization is this message. > [CLI] … awaiting Starting Prism… No log verbosity level setting helps to understand which specific stage of setup takes that long to perform. A swagger.yaml file is quite large, having 380+ type definitions described on 10k+ strings and 290+ paths on 14k+ strings. The mock server from this file is not even starting on cloud Stoplight Workspace I created. The questions are Is it a usual case it takes so long to startup a mock server? Are there any hacks that can help speed up the process? Is there a way to cache the state of the server? Can it be run concurrently? The only approach to speed it up I can think of right now is to split the file on several subdomains (e. g. by paths' first term) and run several prism instances and aggregate them on a single port by using docker compose, as described in the documentation. Have anyone stumbled on a similar issue? Are there better approaches? Thanks!
