jdthorpe

The Problem

You're using the awesome JSDOM package to process a few hundred html files and your process crashes with an error like this:

FATAL ERROR: CALL_AND_RETRY_LAST Allocation failed - JavaScript heap out of memory

If you're sure your code it not accidentally holding on to a reference to (nodes from) prior window objects, then the most likely cause of the apparent memory leak is that

jdthorpe

The standard typings for react-dnd suck. In particular

• They ignore the fact that thre props recieved by the wrapped component are a union of the props passed in to the react-dnd HOC and props collected by react-dnd collectors
• Items are assigned type Object
• DropResults are assigned the type Object

These types are a work and progress, which means they suck too. They just suck a little less than the standard typings

jdthorpe

declare module "kld-intersections" {

    class Shape {}

    export class Point2D {
        constructor(x:number,y:number)
    }

    class Bezier2 extends Shape {

jdthorpe

{
  "requires": true,
  "lockfileVersion": 1,
  "dependencies": {
    "@types/bluebird": {
      "version": "3.5.23",
      "resolved": "https://registry.npmjs.org/@types/bluebird/-/bluebird-3.5.23.tgz",
      "integrity": "sha512-xlehmc6RT+wMEhy9ZqeqmozVmuFzTfsaV2NlfFFWhigy7n6sjMbUUB+SZBWK78lZgWHA4DBAdQvQxUvcB8N1tw=="
    },
    "@types/fs-extra": {

jdthorpe

// this requires `namedExports` to work with rollup:
import {join } from "lodash";

// this works with both webpack and rollup and yields a smaller package:
// import join from "lodash/join"; 

// this works everythere, of course: 
// import _ from "lodash"

// this works everwhere because it's an external module and isn't bundled by webpack or rollup:

jdthorpe

Extending python-docx

Background

I was recently tasked with scraping data from a few thousand of word documents each of which contained nested docx documents which were referenced by w:altChunk elements in the body of the main ./word/document.xml file, like so:

jdthorpe

<svg height="210" width="400">
  <path d="M150 0 L75 200 L225 200 Z" />
  Sorry, your browser does not support inline SVG.
</svg>

jdthorpe

/******/ (function(modules) { // webpackBootstrap
/******/ 	// The module cache
/******/ 	var installedModules = {};
/******/
/******/ 	// The require function
/******/ 	function __webpack_require__(moduleId) {
/******/
/******/ 		// Check if module is in cache
/******/ 		if(installedModules[moduleId]) {
/******/ 			return installedModules[moduleId].exports;

jdthorpe

function kernelDensityEstimator(kernel:{(x:number):number}, X:number[]) {
  return function(V:number[]) {
    return X.map(function(x) {
      return [x, d3.d3.mean(V, function(v:number) { return kernel(x - v); })];
    });
  };
}

function kernelEpanechnikov(k:number): {(x:number):number}{

jdthorpe

Having to add typings to each require is almost as useless as not having typings at all. For example,

import DB from "nedb"
const myCollection = new nedb(DB)
// insert one kind of document
myCollection.insert<{name:string}>({name:"nedb"})
// expect to find a different kind of document
myCollection.findOne<{name:number}>({},(err,doc) => console.log(typeof doc.name))