martinklepsch

(ns cljdoc.pathom3
  (:require [com.wsscode.pathom.core :as p]
            [com.wsscode.pathom.connect :as pc]
            [clojure.core.async :as async]
            [clojure.set]
            [cljdoc.util.repositories :as repos]
            [cljdoc.util.pom :as pom]
            [cljdoc.config :as config]
            [cljdoc.storage.sqlite-impl :as cljdoc-sqlite]))

martinklepsch

(ns cljdoc.pathom3
  (:require [com.wsscode.pathom.core :as p]
            [com.wsscode.pathom.connect :as pc]
            [clojure.core.async :as async]
            [clojure.set]
            [cljdoc.util.repositories :as repos]
            [cljdoc.util.pom :as pom]
            [cljdoc.config :as config]
            [cljdoc.storage.sqlite-impl :as cljdoc-sqlite]))

martinklepsch

I switched the storage layer. SQLite is now used instead of lib-grimoire. More details in the respective architecture decision record: ADR-0013

This was a lot of work but it sets the project up for more interesting extensions besides API documentation and articles (think specs & examples).

Also I looked more into integrating specs but without changes to spec it is impossible to determine if a spec originates from the artefact that is being analysed or from one of it's dependencies. To fix this specs will need to support metadata (CLJ-2194) but the timeline for this is unclear.

In the light of this I'm considering focusing on examples first. More details to come.

Some more minor things that happened:

martinklepsch

2018

August

2018-08-07 15:27,

• Explored various metrics solutions to measure performance impact of new storage backend

martinklepsch

martinklepsch

(defmacro timed [expr]
  `(let [start# (. System (nanoTime))
         ret# ~expr]
     (prn (str ~(str expr) " elapsed time: " (/ (double (- (. System (nanoTime)) start#)) 1000000.0) " msecs"))
     ret#))

martinklepsch

;; Usage: execute each statement in a boot (http://boot-clj.com/) REPL, adjust as needed

(set-env! :dependencies '[[org.clojure/data.xml "0.2.0-alpha2"]
                          [clj-time "0.13.0"]])

(require '[clojure.java.io :as io]
         '[clojure.data.xml :as xml]
         '[clojure.string :as string]
         '[clj-time.format :as f])

martinklepsch

// plain function
function add_one(x) {
  return 1 + x;
}

// constructor
MyClass = function(name) {
  this.myName = name;
};

martinklepsch

(ns org.martinklepsch.rum-system
  "Rum components to start/stop com.stuartsierra.component systems and inject them into React context."
  (:refer-clojure :exclude [comp])
  (:require [rum.core :as rum]
            [goog.object :as gobj]
            [com.stuartsierra.component :as component]))

(def context-k (-> (str ::get-sys) (subs 1)))
(def context-types {context-k js/React.PropTypes.func})

martinklepsch

Originally posted on the mailinglist here. Mostly copied it here so I can add it to Pocket.

---

Here are some problems/limitations of existing OO/GF systems that I don't intend to repeat:

A) They provide only a single declaration point for all superclasses of a class B) They consider the local declaration order of superclasses to be