flowtype - Typing conditional properties with union -


i have following method defined on custom component es6 class takes object component property. if property instance of component assigns ref else creates new instance el , opts properties :

setref({ id, component, el, opts = {}, props = {} }: refconstructortype | refinstancetype): promise<component> {     let ref: component;      if (component instanceof component) {           ref = component     } else {         ref = new component(el, opts);     } }

my type definitions refconstructortype , refinstancetype are: 

type refinstancetype = {|     component: component,     id: string,     props?: {} |};  type refconstructortype = {|     component: typeof component,     id: string,     el: element,     opts ?: {[option_ke: string]: string},     props ?: {} |};

anyway flowtype complaining:

86:     setref({ id, component, el, opts = {}, props = {} }: refconstructortype | refinstancetype): promise<component> {                     ^^^^^^^^^ component. type incompatible 86:     setref({ id, component, el, opts = {}, props = {} }: refconstructortype | refinstancetype): promise<component> {                     ^^^^^^^^^ class type: component  src/index.js:86 86:     setref({ id, component, el, opts = {}, props = {} }: refconstructortype | refinstancetype): promise<component> {                     ^^^^^^^^^ class type: component. type incompatible 86:     setref({ id, component, el, opts = {}, props = {} }: refconstructortype | refinstancetype): promise<component> {                     ^^^^^^^^^ component  src/index.js:86 86:     setref({ id, component, el, opts = {}, props = {} }: refconstructortype | refinstancetype): promise<component> {                                 ^^ property `el`. property not found in 86:     setref({ id, component, el, opts = {}, props = {} }: refconstructortype | refinstancetype): promise<component> {                                                                                 ^^^^^^^^^^^^^^^ object type  src/index.js:86 86:     setref({ id, component, el, opts = {}, props = {} }: refconstructortype | refinstancetype): promise<component> {                                     ^^^^ property `opts`. property not found in 86:     setref({ id, component, el, opts = {}, props = {} }: refconstructortype | refinstancetype): promise<component> {                                                                                 ^^^^^^^^^^^^^^^ object type

any hint on how solve this?

update 26/07/2017

following suggested pattern ended following refined code:

setref(refcfg: refconstructortype | refinstancetype): promise<component> {      let ref: component;      if (!isplainobject(refcfg)) {         throw new error('invalid reference configuration');     }      if (refcfg.component instanceof component) {         ref = refcfg.component;     } else if (typeof refcfg.component === 'function' && refcfg.el) {         const { el, opts, component } = refcfg;         ref = new component(el, opts); //eslint-disable-line new-cap     } else {         throw new error('invalid reference configuration');     }     // .... }

this pretty complex situation, safe way around not de-structure.

you're putting flow in position you're asking create variable component type component | typeof component , asking if union type valid against refconstructortype.component , valid against refinstancetype.component:

  • component == component | typeof component?
  • typeof component == component | typeof component?

basically, destructuring, implicitly fails "one in, out": https://flow.org/en/docs/types/unions/#toc-union-types-requires-one-in-but-all-out.

that's why you're seeing flow complain twice.

once fix that, other errors going solved making sure flow can accurately discriminate between different conditions in if/else.


Comments

Post a Comment

Popular posts from this blog

python - Best design pattern for collection of objects -

i have following architecture: +-----------------+ | collection | | | | | | | | | +-----------------+ | ^ | | create() | | | | v | getitem(b) +------------+ | | item | | | | | | +------+ | | +------------+ a class managing collection of items, creates items. , these items may need other items collection. the actual code in python, , collection passes parameter created item . see it, bad practice. improvement can see pass few functions of collection needed item , instead of whole instance. for example: class collection: getitem(self, id): ... createitem(self, id): item = item(id, self) # <-- pass self ... class item: __init__(self, id, col): ... col.getite...
Read more

go - serving up pdfs using golang -

Image
just wondering if guys can code. it works serving images not pdf brochure. in advance func main(){ http.handlefunc("/", func(w http.responsewriter, r *http.request){ filename := "/var/www/filedipenser/brochure.pdf" streampdfbytes, err := ioutil.readfile( filename ) if err != nil { fmt.println(err) os.exit(1) } b := bytes.newbuffer(streampdfbytes) w.header().set("content-type", "application/pdf") if _, err := b.writeto(w); err != nil { fmt.fprintf(w, "%s", err) } w.write([]byte("pdf generated")) }) err := http.listenandserve(":4111", nil) if err != nil { log.fatal("listenandserve: ", err) fmt.println(err) } } the code bit inefficient appears working. however, pdf viewers might sensitive extraneous output adding end of http response stream...
Read more

r - Using `bbmle:mle2` with vector parameters (already works using `optim`) -

i having trouble using bbmle:mle2 function when trying regression. illustrate problem, have come toy example. we define minus log-likelihood poisson distribution (or custom distribution): ll <- function(beta, z, x){ -sum(stats::dpois(x, lambda = exp(z %*% beta), log = true)) } in code above, beta parameter vector estimate, z model/design matrix , x variable of interest. i generate random data work with: set.seed(2) age <- round(exp(rnorm(5000, mean = 2.37, sd = 0.78) - 1)) claim <- rpois(5000, lambda = 0.07 i can use optim regression. here intercept model: z1 <- model.matrix(claim ~ 1) optim(par = 0, fn = ll, z = z1, x = claim) here intercept + age model: z2 <- model.matrix(claim ~ age) optim(par = c(0, 0), fn = ll, z = z2, x = claim) the way large number of different models can assessed quite easy, 1 has specify model matrix. how can made work mle2 function bbmle package? i can it, if beta one-dimensional: mle2(minuslogl = functio...
Read more