Compiling enterprise

Ivan Koshelev blog on software development

Articles for tags 'JavaScript'

Modern reusable components

Separating DOM structure from visual structure - maximum flexibility with minimum complexity

Working on a suite of applications for a big enterprise, one understands the benefits of reusable components quite quickly. Chances are, at least 50% of functionality, like a user-friendly customer lookup, will be used in the majority of your applications with just a little bit variation in in how it looks or works case-by-case. As an experienced developer, you take a base component library and build your CustomerSelectorCmp on top of it up to the current requirements. It has all the necessary logic, it is used in 20 apps and it works great. Until new requirements come:

  • A new app is being commissioned, and CustomerSelectorCmp needs to have a horizontal mode for it; This will probably be used only in this 1 app;
  • Our managers have doubled sales numbers by grabbing tablets for client visits and having our app ready to go in the heat of the moment; But CustomerSelectorCmp could be much better with touch;
  • We have 10% of managers with special visual needs – it would be very nice to have accessibility mode for CustomerSelectorCmp;

5 years ago, trying to accommodate all this functionality in a single component would inevitably have to violate Single Responsibility Principle and lead to an unmaintainable monstrosity, where key CSS properties are overridden at-least 3 times each, classes are added just to increase CSS selector specificity and !important is your best and worst friend. Well, lucky we are in 2020! Thanks to CSS-Grid and React, we can have a base component that contains all the key functionality and can change appearance drastically with just a dozen lines of CSS and JS. See the following 3 dropdowns? They are the same component, same DOM! The difference is a dozen lines of CSS!

Grab the showcase project form github github and lets dive in.

continue reading

Debugging JavaScript 02

Hints for debugging modern Web apps

This article will help you with debugging modern Web aps. ‘Modern’ here specifically means apps where source code undergoes a major transformation before execution, e.g. you are using Babel to enable features from ES2015-ES2018, TypeScript, and/or frameworks like Mobx that rely on transformations like decorators and transpile down to ES5. Because of the above, code that is actually executed by JS runtime is very different from what you see in your IDE.

You may be thinking, ‘but we’ve got source maps !’ While source maps are good, transforming ES2015-ES2018 features into ES5 changes code too dramatically, a single source line often becomes 3 or more during execution, this becomes _this, variable names change due to transpilation of let and const into var, etc. All of these changes really mess with debugging features like step over or examining variable values by their name. Debugging it can be a pain, but it does not have to.

Sample project for this article is avaliable on Github
continue reading

Debugging JavaScript 01

Advanced forensic techniques

JavaScript is among the most open-source friendly languages. Thus, when faced with a question “why doesn’t it call this function?” or “What changes this property?”, most of the time we have the code of 3rd-party libraries ready and available for us to dive into for debug. Still, sometimes you only have the minified version of a library, or a normal version so big and complicated - you give up after pressing “step into method” over 30 times. You need to track the exact place where one of your objects properties is used.

Solution? Use Chrome and the most awesome feature of ES6 – Proxies! Your users are locked into IE 8 and no ES6? What’s wrong with your organization? Ok, but you as a developer, surely can use Chrome, just for this debug session? Proxies are supported by all modern browser releases for at least a year , and current versions of Node.

continue reading