Michael Dennis

The End of Request::get()? Why Symfony 7.4 is a Wake-Up Call for Your Codebase. I’ve seen frameworks evolve, but few minor releases bring the level of strictness and clarity that Symfony 7.4 delivers. If you’re still relying on the "magic" of $request->get(), it’s time to modernize. In my article, I wrote about the pivotal changes in HTTP handling that every Symfony developer needs to know before Symfony 8.0 arrives. 🚀 What’s inside: ✅ The Deprecation of Request::get(): Why the days of ambiguous "bag-merging" are over and why explicit access (Query vs. Payload) is the new standard for security. ✅ Native Body Parsing: How PHP 8.4 and Symfony 7.4 finally solve the headache of parsing JSON/FormData in PUT and PATCH requests. ✅ Strict Method Overriding: Tighter control over _method parameters to prevent obscure HTTP overrides. ✅ Intelligent MIME Types: Smarter handling for complex types like application/problem+json. This isn't just about deprecations; it's about writing code that is safer, faster, and unambiguous. 👉 Read the full guide here: https://lnkd.in/d-nTF236 https://lnkd.in/d6SkWmC7 How are you preparing your legacy apps for these changes? Let's discuss in the comments! 👇 #Symfony #PHP #WebDevelopment #Symfony7 #CleanCode #SoftwareArchitecture #DevCommunity

7

Easy to look at this post from Alexander Wurts and be like "yeah duh." But the reality is that some very talented devs are falling behind and this is why. I think root cause is that many early adopters -- i.e. smart ppl in tech but not the frontier innovators -- tried coding agents early when they still sort of sucked. Now that coding agents are great, many early adopters are hesitant to offload as aggressively as they could to agents. Food for thought.

11

2 Comments

🐘 "PHP Grew Up. Did You Notice?" PHP in 2024 was a completely different language than the one you're making fun of. I get it. The old jokes write themselves. mysql_query(). Global state everywhere. $_POST['whatever']. I lived through that era too. I have the scars. But here's the thing, while a lot of us moved on and kept dunking on PHP from the sidelines, the language kept shipping. Quietly. Consistently. Every year. PHP 8.0 brought named arguments, match expressions, and union types. PHP 8.1 brought enums, fibers, readonly properties, intersection types. PHP 8.2 brought readonly classes, new random APIs, and a serious cleanup of the deprecation backlog. That's not maintenance. That's a language that's actively evolving. Over the next few weeks I'm going to walk through the features I wish I'd paid attention to sooner. Not with docs-style explanations, just real examples of how they changed the way I actually write code day to day. If you still think PHP is stuck in 2009, stick around. I think you'll be surprised. Feature 1: Readonly Classes You were writing 3 files to do what one keyword now does For years, making a truly immutable object in PHP felt like a punishment. You'd write a class, make all the properties private, write a constructor, write a getter for every single property, add a docblock, hope nobody on the team did something creative with reflection... and then do it again for the next DTO. Every. Single. Time. PHP 8.2 just added readonly class and I genuinely felt a little embarrassed about how much boilerplate I used to write. The readonly modifier on the class means every single property is automatically readonly. No setters. No accidental mutation. No defensive getters. You can still make properties public because nothing can change them after construction anyway. This is the kind of feature that makes your DTOs, value objects, and API response models genuinely pleasant to write. I've refactored probably a dozen classes since this landed. Each one got shorter. None got worse. What pattern do you use for immutable data transfer objects? Still doing the old approach? #PHP #PHP82 #WebDevelopment #CleanCode #PHPGrewUp #BackendDev #BackendDevelopment

22

He's right. Typescript is the correct thing to learn in 2025. Knowing JS isn't bad; there's a lot of code written in it, but TS is the better language to work in, especially for new projects. I ran OKC.js for about a decade, and I've given numerous technical talks on JS (possibly the most in Oklahoma) over the years and I am co-signing this, you need to know it now.

8

1 Comment

How to get scammed in software dev space? Pay $63,500 for “special treatment” like this. 👇 An Alabama founder told me his story on a call. 4 months of “development”. At the end he got: 1 Android install file. It crashed on open. No working app. No code he could see. No docs. When he asked for the source code the team said: “Send $5,500 more and we will give you the repo.” That is not a change request. That is a hostage situation. He thought he was buying an app. What he really bought was 4 months of blind trust. No clear milestones. No weekly demos. No written plan for handover. Here is what should happen instead. You see progress every week. You get builds to test. You talk through what was done and what is next. You have in writing that you own the code at the end and when you get it. You might not get everything on day 3. But you should never reach month 4 and see a broken file and nothing else. If a team wants full payment with no clear delivery points no regular demos and only “you will see everything at the end” That is maaaajor red flag. In software you are not just choosing a price. You are choosing who you trust with your product.

6

If you’re a US startup hiring mobile this quarter: screen for release ownership, not Flutter trivia. Senior candidates don’t separate “engineering” from “shipping.” They own releases, hotfixes, and the messy production stuff that actually matters. One question that works: “Tell me about a release that went sideways. What broke, how did you diagnose it, and what did you change so it wouldn’t happen again?” The answer tells you if they’ve actually owned production when it broke.

17

5 Comments

CSS Grid is underrated. Most devs stop at grid-template-columns: 1fr 1fr 1fr and miss the real power: → Named grid lines → Implicit areas with -start/-end → Full-width breakouts without negative margins → Layouts you can literally read like blueprints https://lnkd.in/eg9wtCdY

5

The simple factor that may cancel the AI software development revolution People do people things. I want to warn anyone who thinks that a technology that is "more efficient" will automatically win: this is historically not true. If you think otherwise then please explain the success of PHP to me. There was tech back then that made people more than 20x more productive than the alternatives - by some metrics. By others it was more than 100x (when taking into account the cost of security breaches, for example). Paul Graham wrote an essay about this, called "beating the averages". None of this mattered. And I don't think the wins of AI coding are that large. They only look like that because modern software development often is a serial exercise in iterative self-foot-shooting, and AI does away with that (partially, and I suspect also only temporarily). It's always been the case that many of those multi-year, large-team projects could have been done by a motivated, talented guy in a few months by just cutting the BS. But this is not how the world works. The music has a tendency to go where the people go, and "more efficient" undermines it's own popularity base by making people redundant. People will find a way, and whatever will be popular in the future will be popular because it needs people. This dynamic has played out a couple of times as far as I can tell. It looks like a solid law of nature to me. [cropped image from easy-peasy.ai. That office fragment in the middle is just hilarious]

5