Re: Stubbing with instrument and generators
Thanks for taking a look. Here's a ticket:
https://dev.clojure.org/jira/browse/CLJ-2197
> On 2017-06-30, at 11:52, Alex Miller wrote:
>
> I don't think this is the same case as CLJ-2095 as the instrumented var
> should have the opportunity to use the gen overrides you've included. In
> particular, it seems like stest/instrument-choose-fn is currently passing the
> :gen overrides map when getting the gen for the stubbed var. I'm wondering if
> there is a bug there where the spec has already been resolved and thus the
> gen key (::y) is not getting matched. Feel free to file a ticket for this.
>
> On Friday, June 30, 2017 at 9:37:42 AM UTC-5, Michael Glaesemann wrote:
> Using spec instrument to stub functions is really helpful. I'm using stubbing
> to test Stuart Sierra-style components with some success. However, I've been
> surprised that generator override doesn't work as I would expect it to.
> Here's an example:
>
> ;; [org.clojure/spec.alpha "0.1.123"]
>
> (require '[clojure.spec.alpha :as s])
> (require '[clojure.spec.gen.alpha :as gen])
> (require '[clojure.spec.test.alpha :as stest])
>
> (defprotocol Y
> (-do-y [r]))
>
> (def y? (partial satisfies? Y))
> (s/def ::y y?)
>
> ;; Protocol methods can't be spec'd, so wrap it in a function.
>
> (defn do-y [r]
> (-do-y r))
>
> (s/fdef do-y :args (s/cat :y-er ::y))
>
> ;; Example of the protocol implementation that we're going to stub.
>
> (defrecord BadYer []
> Y
> (-do-y [_] (throw (Exception. "can't make me!"
>
>
> ;; Confirm BadYer instances are valid with respect to the protol spec.
>
> (s/valid? ::y (->BadYer))
> ;; => true
>
> ;; And confirm BadYer instances will throw when called.
>
> (try
> (do-y (->BadYer))
> (catch Exception e
> (.getMessage e)))
> ;; => "can't make me!"
>
>
> (def y-gen (gen/return (->BadYer)))
>
> ;; Confirm generator works as expected:
>
> (gen/sample y-gen 1)
> ;; => (#spec_ex.core.BadYer{})
>
> ;; We want to stub `do-y`, providing y-gen as a generator for `::y`
>
> (try
> (stest/instrument `do-y {:stub #{`do-y}
>:gen {::y (fn [] y-gen)}})
> (catch Exception e
> (ex-data e)))
> ;; => #:clojure.spec.alpha{:path [:y-er], :form :spec-ex.core/y, :failure
> :no-gen}
>
> ;; However, we *can* stub `do-y` if we replace its spec.
>
> (stest/instrument `do-y
> {:stub #{`do-y}
>:spec {`do-y (s/fspec
> :args (s/cat :y-er (s/with-gen ::y
>(fn [] y-gen}})
> ;; => [spec-ex.core/do-y]
>
> There is a ticket open[ (CLJ-2095[1]) regarding using overrides s/gen with
> custom generators. Is this a case where this applies? I can imagine that it
> could be. Not overriding something that isn't there would result in the thing
> still not being there, thus the :no-gen failure. And it is something that can
> be worked around. It would be decidedly more succinct if the gen override
> worked rather than the spec override.
>
> Best,
>
> Michael Glaesemann
> grzm seespotcode net
>
>
> [1]: https://dev.clojure.org/jira/browse/CLJ-2095
>
>
> --
> You received this message because you are subscribed to the Google
> Groups "Clojure" group.
> To post to this group, send email to clojure@googlegroups.com
> Note that posts from new members are moderated - please be patient with your
> first post.
> To unsubscribe from this group, send email to
> clojure+unsubscr...@googlegroups.com
> For more options, visit this group at
> http://groups.google.com/group/clojure?hl=en
> ---
> You received this message because you are subscribed to the Google Groups
> "Clojure" group.
> To unsubscribe from this group and stop receiving emails from it, send an
> email to clojure+unsubscr...@googlegroups.com.
> For more options, visit https://groups.google.com/d/optout.
Michael Glaesemann
grzm seespotcode net
--
You received this message because you are subscribed to the Google
Groups "Clojure" group.
To post to this group, send email to clojure@googlegroups.com
Note that posts from new members are moderated - please be patient with your
first post.
To unsubscribe from this group, send email to
clojure+unsubscr...@googlegroups.com
For more options, visit this group at
http://groups.google.com/group/clojure?hl=en
---
You received this message because you are subscribed to the Google Groups
"Clojure" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to clojure+unsubscr...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.
Re: Stubbing with instrument and generators
I don't think this is the same case as CLJ-2095 as the instrumented var
should have the opportunity to use the gen overrides you've included. In
particular, it seems like stest/instrument-choose-fn is currently passing
the :gen overrides map when getting the gen for the stubbed var. I'm
wondering if there is a bug there where the spec has already been resolved
and thus the gen key (::y) is not getting matched. Feel free to file a
ticket for this.
On Friday, June 30, 2017 at 9:37:42 AM UTC-5, Michael Glaesemann wrote:
>
> Using spec instrument to stub functions is really helpful. I'm using
> stubbing to test Stuart Sierra-style components with some success. However,
> I've been surprised that generator override doesn't work as I would expect
> it to. Here's an example:
>
> ;; [org.clojure/spec.alpha "0.1.123"]
>
> (require '[clojure.spec.alpha :as s])
> (require '[clojure.spec.gen.alpha :as gen])
> (require '[clojure.spec.test.alpha :as stest])
>
> (defprotocol Y
> (-do-y [r]))
>
> (def y? (partial satisfies? Y))
> (s/def ::y y?)
>
> ;; Protocol methods can't be spec'd, so wrap it in a function.
>
> (defn do-y [r]
> (-do-y r))
>
> (s/fdef do-y :args (s/cat :y-er ::y))
>
> ;; Example of the protocol implementation that we're going to stub.
>
> (defrecord BadYer []
> Y
> (-do-y [_] (throw (Exception. "can't make me!"
>
>
> ;; Confirm BadYer instances are valid with respect to the protol spec.
>
> (s/valid? ::y (->BadYer))
> ;; => true
>
> ;; And confirm BadYer instances will throw when called.
>
> (try
> (do-y (->BadYer))
> (catch Exception e
> (.getMessage e)))
> ;; => "can't make me!"
>
>
> (def y-gen (gen/return (->BadYer)))
>
> ;; Confirm generator works as expected:
>
> (gen/sample y-gen 1)
> ;; => (#spec_ex.core.BadYer{})
>
> ;; We want to stub `do-y`, providing y-gen as a generator for `::y`
>
> (try
> (stest/instrument `do-y {:stub #{`do-y}
>:gen {::y (fn [] y-gen)}})
> (catch Exception e
> (ex-data e)))
> ;; => #:clojure.spec.alpha{:path [:y-er], :form :spec-ex.core/y, :failure
> :no-gen}
>
> ;; However, we *can* stub `do-y` if we replace its spec.
>
> (stest/instrument `do-y
> {:stub #{`do-y}
>:spec {`do-y (s/fspec
> :args (s/cat :y-er (s/with-gen ::y
>(fn [] y-gen}})
> ;; => [spec-ex.core/do-y]
>
> There is a ticket open[ (CLJ-2095[1]) regarding using overrides s/gen with
> custom generators. Is this a case where this applies? I can imagine that it
> could be. Not overriding something that isn't there would result in the
> thing still not being there, thus the :no-gen failure. And it is something
> that can be worked around. It would be decidedly more succinct if the gen
> override worked rather than the spec override.
>
> Best,
>
> Michael Glaesemann
> grzm seespotcode net
>
>
> [1]: https://dev.clojure.org/jira/browse/CLJ-2095
>
>
--
You received this message because you are subscribed to the Google
Groups "Clojure" group.
To post to this group, send email to clojure@googlegroups.com
Note that posts from new members are moderated - please be patient with your
first post.
To unsubscribe from this group, send email to
clojure+unsubscr...@googlegroups.com
For more options, visit this group at
http://groups.google.com/group/clojure?hl=en
---
You received this message because you are subscribed to the Google Groups
"Clojure" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to clojure+unsubscr...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.
Stubbing with instrument and generators
Using spec instrument to stub functions is really helpful. I'm using stubbing
to test Stuart Sierra-style components with some success. However, I've been
surprised that generator override doesn't work as I would expect it to. Here's
an example:
;; [org.clojure/spec.alpha "0.1.123"]
(require '[clojure.spec.alpha :as s])
(require '[clojure.spec.gen.alpha :as gen])
(require '[clojure.spec.test.alpha :as stest])
(defprotocol Y
(-do-y [r]))
(def y? (partial satisfies? Y))
(s/def ::y y?)
;; Protocol methods can't be spec'd, so wrap it in a function.
(defn do-y [r]
(-do-y r))
(s/fdef do-y :args (s/cat :y-er ::y))
;; Example of the protocol implementation that we're going to stub.
(defrecord BadYer []
Y
(-do-y [_] (throw (Exception. "can't make me!"
;; Confirm BadYer instances are valid with respect to the protol spec.
(s/valid? ::y (->BadYer))
;; => true
;; And confirm BadYer instances will throw when called.
(try
(do-y (->BadYer))
(catch Exception e
(.getMessage e)))
;; => "can't make me!"
(def y-gen (gen/return (->BadYer)))
;; Confirm generator works as expected:
(gen/sample y-gen 1)
;; => (#spec_ex.core.BadYer{})
;; We want to stub `do-y`, providing y-gen as a generator for `::y`
(try
(stest/instrument `do-y {:stub #{`do-y}
:gen {::y (fn [] y-gen)}})
(catch Exception e
(ex-data e)))
;; => #:clojure.spec.alpha{:path [:y-er], :form :spec-ex.core/y, :failure
:no-gen}
;; However, we *can* stub `do-y` if we replace its spec.
(stest/instrument `do-y
{:stub #{`do-y}
:spec {`do-y (s/fspec
:args (s/cat :y-er (s/with-gen ::y
(fn [] y-gen}})
;; => [spec-ex.core/do-y]
There is a ticket open[ (CLJ-2095[1]) regarding using overrides s/gen with
custom generators. Is this a case where this applies? I can imagine that it
could be. Not overriding something that isn't there would result in the thing
still not being there, thus the :no-gen failure. And it is something that can
be worked around. It would be decidedly more succinct if the gen override
worked rather than the spec override.
Best,
Michael Glaesemann
grzm seespotcode net
[1]: https://dev.clojure.org/jira/browse/CLJ-2095
--
You received this message because you are subscribed to the Google
Groups "Clojure" group.
To post to this group, send email to clojure@googlegroups.com
Note that posts from new members are moderated - please be patient with your
first post.
To unsubscribe from this group, send email to
clojure+unsubscr...@googlegroups.com
For more options, visit this group at
http://groups.google.com/group/clojure?hl=en
---
You received this message because you are subscribed to the Google Groups
"Clojure" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to clojure+unsubscr...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.
