re-frame/docs/Debugging-Event-Handlers.md

8.2 KiB

Debugging Event Handlers

This page describes techniques for debugging re-frame's event handlers.

Event handlers are quite central to a re-frame app. Only event handlers can update app-db, to "step" an application "forward" from one state to the next.

Table Of Contents

The debug Interceptor

You might wonder: is my event handler making the right changes to app-db?

During development, the built-in debug interceptor can help. It writes to console.log:

  1. the event being processed, for example: [:attempt-world-record true]
  2. the changes made to db by the handler in processing the event

debug uses clojure.data/diff to compare app-db before and after the handler ran, showing what changed.

clojure.data/diff returns a triple , the first two entries of which debug will display in console.log (the 3rd says what hasn't changed and isn't interesting).

The output produced by clojure.data/diff can take some getting used to, but you should stick with it -- your effort will be rewarded.

Using debug

So, you will add this Interceptor like this:

(re-frame.core/reg-event-db
   :some-id
   [debug]         ;;  <----  added here! 
   some-handler-fn)

Except, of course, we need to be more deft - we only want debug in development builds. We don't want the overhead of those clojure.data/diff calculations in production. So, this is better:

(re-frame.core/reg-event-db
   :some-id
   [(when ^boolean goog.DEBUG debug)]   ;;  <----  conditional! 
   some-handler-fn)

goog.DEBUG is a compile time constant provided by the Google Closure Compiler. It will be true when the build within project.clj is :optimization :none and false otherwise.

Ha! I see a problem, you say. In production, that when is going to leave a nil in the interceptor vector. So the Interceptor vector will be [nil].
Surely that's a problem?

Well, actually, no it isn't. re-frame filters out nil.

Too Much Repetition - Part 1

Each event handler has its own interceptor stack.

That might be all very flexible, but does that mean we have to put this debug business on every single handler? That would be very repetitive.

Yes, you will have to put it on each handler. And, yes, that could be repetitive, unless you take some steps.

One thing you an do is to define standard interceptors the top of the event.cljs namespace:

(def standard-interceptors  [(when ^boolean goog.DEBUG debug)  another-interceptor])

And then, for any one event handler, the code would look like:

(re-frame.core/reg-event-db
   :some-id
   standard-interceptors        ;; <--- use the common definition
   some-handler-fn)

or perhaps:

(re-frame.core/reg-event-db
   :some-id
   [standard-interceptors specific-interceptor]  ;; mix with something specific
   some-handler-fn)

So that specific-interceptor could be something required for just this one event handler, and it can be combined the standard ones.

Wait on! "I see a problem", you say. standard-interceptors is a vector, and it is within another vector along side specific-interceptor - so that's nested vectors of interceptors!

No problem, re-frame uses flatten to take out all the nesting - the result is a simple chain of interceptors. And also, as we have discussed,
nils are removed.

3. Checking DB Integrity

Always have a detailed schema for the data in app-db!

Why?

First, schemas serve as invaluable documentation. When I come to a new app, the first thing I want to look at is the underlying information model - the schema of the data. I hope it is well commented and I expect it to be rigorous and complete, using Clojure spec or, perhaps, a Prismatic Schema.

Second a good spec allows you to assert the integrity and correctness of the data in app-db. Because all the data is in one place, that means you are asserting the integrity of ALL the data in your app, at one time.

When should we do this? Ideally every time a change is made!

Well, it turns out that only event handlers can change the value in app-db, so only an event handler could corrupt it. So, we'd like to recheck the integrity of app-db immediately after every event handler has run.

This allows us to catch any errors very early, easily assigning blame (to the rouge event handler).

Schemas are typically put into db.cljs (see the todomvc example in the re-frame repo). Here's an example using Prismatic Schema (although a more modern choice would be to use Clojure spec):

(ns my.namespace.db
  (:require
    [schema.core :as s]))

;; As exactly as possible, describe the correct shape of app-db 
;; Add a lot of helpful comments. This will be an important resource
;; for someone looking at you code for the first time.
(def schema           
  {:a {:b s/Str
       :c s/Int}
   :d [{:e s/Keyword
        :f [s/Num]}]})

And a function which will check a db value against that schema:

(defn valid-schema?
  "validate the given db, writing any problems to console.error"
  [db]
  (let [res (s/check schema db)]
    (if (some? res)
      (.error js/console (str "schema problem: " res)))))

Now, let's organise for valid-schema? to be run after every handler. We'll use the built-in after Interceptor factory function:

(def standard-interceptors [(when ^boolean goog.DEBUG debug)
                           (when ^boolean goog.DEBUG (after db/valid-schema?))]) ;; <-- new

Now, the instant a handler messes up the structure of app-db you'll be alerted. But this overhead won't be there in production.

Too Much Repetition - Part 2

Above we discussed a way of "factoring out" common interceptors into standard-interceptors.

But there's a 2nd way to ensure that all event handlers get certain Interceptors: you write a custom registration function -- a replacement for reg-event-db -- like this:

(defn my-reg-event-db          ;; alternative to reg-event-db
  ([id handler-fn] 
    (my-reg-event-db id nil handler-fn))
  ([id interceptors handler-fn]
    (re-frame.core/reg-event-db 
        id
        [(when ^boolean goog.DEBUG debug)
         (when ^boolean goog.DEBUG (after db/valid-schema?)) 
         interceptors]
        handler-fn)))

Notice that it inserts our two standard Interceptors.

From now on, you can register your event handlers like this and know that the two standard Interceptors have been inserted:

(my-reg-event-db      ;; <-- adds std interceptors automatically
  :some-id 
  some-handler-fn)

What about the -fx variation?

Above we created my-reg-event-db as a new registration function for -db handlers. That's handlers which take db and event arguments, and return a new db.
So, they MUST return a new db value - which should be validated.

But what if we tried to do the same for -fx handlers, which return instead an effects map which may, or may not, contain an :db? Our solution would have to allow for the absence of a new db value (by doing no validity check, because nothing was being changed).

(defn my-reg-event-fx          ;; alternative to reg-event-db
  ([id handler-fn] 
    (my-reg-event-db id nil handler-fn))
  ([id interceptors handler-fn]
    (re-frame.core/reg-event-fx 
        id
        [(when ^boolean goog.DEBUG debug)
         (when ^boolean goog.DEBUG (after #(if % (db/valid-schema? %))))
         interceptors]
        handler-fn)))

Actually, it would probably be better to write an alternative after which