Jiří Pudil
Hello, I am
Jiří Pudil
I turn into <code>

Blog – #php

Don't bring your own HTTP client

If you are developing a library that needs to make HTTP requests, here's how not to make it difficult for your consumers.

comments
Read more

Body parser middleware that has your back

If you've ever implemented an HTTP API in PHP, you probably know that there's one thing you have to do over and over and over: parse the request body and validate the resulting structure. Now that we have PSR-15, let's use a middleware for that!

comments
Read more

Don't just float, swim with the floating point

They say you should understand floating-point precision by 35.0000000000014. This post is here to help you: it discusses how floats are (im)precise and why, demonstrates the problems you inevitably encounter when doing math with floats, and presents ways to prevent those problems by using more proper data types and structures.

comments
Read more

The frameworkless application

There's a new buzzword on the PHP block: frameworkless. The activities of PHP communities last year made it possible to build robust applications, uniquely composed of little bits, without relying on a big framework. Let's find out how they work and if it's a good approach.

comments
Read more

Keeping your dependencies up-to-date

As the business requirements grow, so does the list of dependencies. Managers ask for a new feature, you just composer-require this thing and that other thing, and wire it into the application. But how do you keep track of new features in the dependencies? Or bugfixes? Or, most importantly, security patches?!

comments
Read more

Clock needs an interface

Almost every application needs to work with time: if not for some greater purpose like expiring access credentials, entities usually store information about when they were created or generally when something happened. Telling time is indeed a crucial responsibility, and being able to rely on the mechanism across the whole codebase is really important. But it can get problematic.

comments
Read more

Announcing three new open-source projects

I've had a really productive couple of weeks and (mostly) weekends recently. I've worked extensively on one of my side projects using modern JS stack including Webpack, faced issues here and there, and found a way to solve them. In the end, I proudly announce three new open-source packages: a modern AJAX library for Nette Framework, Neon loader for Webpack, and a library integrating the Webpack build process into Nette-powered applications.

comments
Read more

CacheFactory: for better testing of cache integration

Unit testing a class that uses Nette's cache can be a pain in the you-know-what. I've found myself writing a very simple cache factory to solve this and make testing classes depending on cache a pleasing experience, so I've decided to make an extension out of it.

comments
Read more

Filtering data by user input with Kdyby/Doctrine

Listing data is essentially the most crucial part of websites. Be it products, articles, photos or whatnots, we usually need to provide the user the way to filter and/or sort the data by some preset parameters. I'll show you how to encapsulate such filtering within an object, build a user interface (in other words, a form) upon it, and use it with Kdyby/Doctrine's query objects to actually filter the data on the database level.

comments
Read more

Introduction to Kdyby/DoctrineForms

Last December, I implemented smart forms via Kdyby/DoctrineForms on this blog. Since the documentation of the package seems to be quite lacking in its scope, I decided to share with you how to overcome the pitfalls I had encountered and make it work.

comments
Read more