<?xml version="1.0" encoding="utf-8" standalone="yes"?><rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom"><channel><title>Mattias Geniar on ma.ttias.be</title><link>https://ma.ttias.be/</link><description>Recent content in Mattias Geniar on ma.ttias.be</description><generator>Hugo -- gohugo.io</generator><language>en-us</language><managingEditor>mattias@ma.ttias.be (Mattias Geniar)</managingEditor><webMaster>mattias@ma.ttias.be (Mattias Geniar)</webMaster><lastBuildDate>Sun, 01 Nov 2026 00:00:00 +0000</lastBuildDate><atom:link href="https://ma.ttias.be/index.xml" rel="self" type="application/rss+xml"/><item><title>CVE-2026-98432: which Linux kernels are still vulnerable</title><link>https://ma.ttias.be/linux-cve-2026-98432-vulnerable-kernels/</link><pubDate>Sun, 01 Nov 2026 00:00:00 +0000</pubDate><author>mattias@ma.ttias.be (Mattias Geniar)</author><guid>https://ma.ttias.be/linux-cve-2026-98432-vulnerable-kernels/</guid><description>&lt;p&gt;&lt;a href="https://www.cve.org/CVERecord?id=CVE-2026-98432" target="_blank" rel="noopener noreferrer"&gt;CVE-2026-98432&lt;/a&gt;
is a use-after-free bug in the Linux kernel&amp;rsquo;s IPv4 TCP handshake. Anyone who can send a SYN to a listening port can reach it. Your application never sees a single byte. The bug goes back to Linux 4.4. Check your distribution and kernel flavor in the tables below.&lt;/p&gt;</description></item><item><title>Xcode 27 drew a white strip over my iPhone app's status bar</title><link>https://ma.ttias.be/xcode-27-white-strip-status-bar/</link><pubDate>Fri, 30 Oct 2026 00:00:00 +0000</pubDate><author>mattias@ma.ttias.be (Mattias Geniar)</author><guid>https://ma.ttias.be/xcode-27-white-strip-status-bar/</guid><description>&lt;p&gt;This morning I opened &lt;a href="https://ma.ttias.be/introducing-snapkin/"&gt;Snapkin&lt;/a&gt;
on my phone after a new build and the top of the screen looked wrong. Behind the clock there was a white strip, where the app&amp;rsquo;s light grey background used to be.&lt;/p&gt;</description></item><item><title>Snapkin now uploads your photos before you tap Analyse</title><link>https://ma.ttias.be/snapkin-pre-emptive-photo-uploads/</link><pubDate>Thu, 22 Oct 2026 00:00:00 +0000</pubDate><author>mattias@ma.ttias.be (Mattias Geniar)</author><guid>https://ma.ttias.be/snapkin-pre-emptive-photo-uploads/</guid><description>&lt;p&gt;Every meal you add to &lt;a href="https://ma.ttias.be/introducing-snapkin/"&gt;Snapkin&lt;/a&gt;
goes through a lot of work on the server: resizing the photos, looking for barcodes and QR codes, preparing everything for the AI call, &amp;hellip; Most of that is hard to make faster. So I moved the one part I &lt;em&gt;could&lt;/em&gt; move: when the photos leave the phone.&lt;/p&gt;</description></item><item><title>The PHP CVE scramble, in one graph</title><link>https://ma.ttias.be/php-land-last-24-hours/</link><pubDate>Thu, 22 Oct 2026 00:00:00 +0000</pubDate><author>mattias@ma.ttias.be (Mattias Geniar)</author><guid>https://ma.ttias.be/php-land-last-24-hours/</guid><description>&lt;p&gt;The last 24 hours in PHP land were a mess. Two critical RCEs in PHP core, proof-of-concepts doing the rounds before the fix was out, and a lot of sysadmins pulling the plug on their own sites rather than waiting to get popped.&lt;/p&gt;</description></item><item><title>PHP CVE-2026-109618 and CVE-2026-109619: early mitigation steps</title><link>https://ma.ttias.be/php-cve-2026-109618-early-mitigation/</link><pubDate>Wed, 21 Oct 2026 00:00:00 +0000</pubDate><author>mattias@ma.ttias.be (Mattias Geniar)</author><guid>https://ma.ttias.be/php-cve-2026-109618-early-mitigation/</guid><description>&lt;p&gt;I wrote this post on the morning of 21 October, before PHP released the fix. &lt;strong&gt;The fix is out now: PHP 8.5.12, 8.4.27, 8.3.36 and 8.2.35.&lt;/strong&gt; Installing it is the only complete fix. &lt;a href="https://ma.ttias.be/php-cve-2026-109618-patch/"&gt;How to apply the patches&lt;/a&gt;
has the steps for each distribution.&lt;/p&gt;</description></item><item><title>PHP CVE-2026-109618 and CVE-2026-109619: how to apply the patches</title><link>https://ma.ttias.be/php-cve-2026-109618-patch/</link><pubDate>Wed, 21 Oct 2026 00:00:00 +0000</pubDate><author>mattias@ma.ttias.be (Mattias Geniar)</author><guid>https://ma.ttias.be/php-cve-2026-109618-patch/</guid><description>&lt;p&gt;&lt;strong&gt;The fix is out: PHP 8.5.12, 8.4.27, 8.3.36 and 8.2.35.&lt;/strong&gt; PHP released them on 21 October at about 13:35 UTC (&lt;a href="https://www.php.net/archive/2026.php" target="_blank" rel="noopener noreferrer"&gt;PHP&amp;rsquo;s advisory&lt;/a&gt;
). Ubuntu, Debian, Red Hat, Amazon Linux, Alpine and the big third-party repositories have fixed packages. &lt;a href="#package-status"&gt;Package status&lt;/a&gt;
shows where you can get it.&lt;/p&gt;</description></item><item><title>PHP's new memory_hard_limit_percent setting</title><link>https://ma.ttias.be/php-memory-hard-limit/</link><pubDate>Wed, 21 Oct 2026 00:00:00 +0000</pubDate><author>mattias@ma.ttias.be (Mattias Geniar)</author><guid>https://ma.ttias.be/php-memory-hard-limit/</guid><description>&lt;p&gt;The PHP security releases of 21 October add a new setting: &lt;code&gt;memory_hard_limit_percent&lt;/code&gt;. It changes what &lt;code&gt;memory_limit&lt;/code&gt; means. &lt;code&gt;memory_limit&lt;/code&gt; is now a soft limit. A second, higher limit sits above it. A process that reaches that one is ended on the spot, with exit code 137. No shutdown functions, no error page, no line in your PHP error log.&lt;/p&gt;</description></item><item><title>Snapkin has an engineering blog now</title><link>https://ma.ttias.be/snapkin-engineering-blog/</link><pubDate>Sat, 17 Oct 2026 00:00:00 +0000</pubDate><author>mattias@ma.ttias.be (Mattias Geniar)</author><guid>https://ma.ttias.be/snapkin-engineering-blog/</guid><description>&lt;p&gt;Building &lt;a href="https://ma.ttias.be/introducing-snapkin/"&gt;Snapkin&lt;/a&gt;
means running a &lt;em&gt;lot&lt;/em&gt; of experiments. Which AI model reads a plate best, what a prompt change does to the accuracy, how long you wait for an answer, what each call costs, &amp;hellip;&lt;/p&gt;</description></item><item><title>From laravel-echo-server to laravel-websockets to Laravel Reverb</title><link>https://ma.ttias.be/from-laravel-echo-server-to-laravel-websockets-to-reverb/</link><pubDate>Thu, 15 Oct 2026 00:00:00 +0000</pubDate><author>mattias@ma.ttias.be (Mattias Geniar)</author><guid>https://ma.ttias.be/from-laravel-echo-server-to-laravel-websockets-to-reverb/</guid><description>&lt;p&gt;&lt;em&gt;This post replaces &amp;ldquo;Transitioning from Laravel Echo Server to Laravel WebSockets&amp;rdquo;, which I wrote for the &lt;a href="https://ohdear.app/news-and-updates?utm_source=ma.ttias.be&amp;amp;utm_medium=referral&amp;amp;utm_campaign=blogpost-from-laravel-echo-server-to-laravel-websockets-to-reverb" target="_blank" rel="noopener noreferrer"&gt;Oh Dear blog&lt;/a&gt;
on 4 December 2018. The full 2018 story is below. I added what happened to both packages since, what Oh Dear runs today, and how I&amp;rsquo;d set this up in 2026 with Laravel Reverb.&lt;/em&gt;&lt;/p&gt;</description></item><item><title>The 7 reasons Apple rejected my app, and what I changed</title><link>https://ma.ttias.be/app-store-rejection-reasons/</link><pubDate>Wed, 30 Sep 2026 00:00:00 +0000</pubDate><author>mattias@ma.ttias.be (Mattias Geniar)</author><guid>https://ma.ttias.be/app-store-rejection-reasons/</guid><description>&lt;p&gt;&lt;a href="https://ma.ttias.be/introducing-snapkin/"&gt;Snapkin&lt;/a&gt;
went live in the App Store on September 24th. It got there on the sixth submission. The first five came back rejected, with seven different reasons between them. 😅&lt;/p&gt;</description></item><item><title>Data is the application</title><link>https://ma.ttias.be/data-is-the-application/</link><pubDate>Tue, 29 Sep 2026 00:00:00 +0000</pubDate><author>mattias@ma.ttias.be (Mattias Geniar)</author><guid>https://ma.ttias.be/data-is-the-application/</guid><description>&lt;p&gt;Agentic coding lets me move &lt;em&gt;very&lt;/em&gt; fast. A feature gets built in days. A design gets implemented in a matter of hours. Changing a layout, updating a template, reworking a flow: all of it happens at a speed I couldn&amp;rsquo;t have imagined a year ago.&lt;/p&gt;</description></item><item><title>Introducing Snapkin: the calorie tracker for people who've quit calorie trackers</title><link>https://ma.ttias.be/introducing-snapkin/</link><pubDate>Sat, 26 Sep 2026 00:00:00 +0000</pubDate><author>mattias@ma.ttias.be (Mattias Geniar)</author><guid>https://ma.ttias.be/introducing-snapkin/</guid><description>&lt;p&gt;Snapkin is live on iPhone and Android. 😄&lt;/p&gt;
&lt;style&gt;
.snapkin-stores { display:flex; flex-wrap:wrap; gap:10px; margin:1.5rem 0; }
.content .snapkin-stores a.snapkin-store { display:inline-flex; align-items:center; gap:10px; min-height:52px; padding:0 22px; border-radius:999px;
background:#0B7A4E; color:#fff !important; text-decoration:none !important; border:0 !important; font-weight:700; font-size:17px; line-height:1.15;
box-shadow:0 1px 0 rgba(0,0,0,.08), 0 6px 18px rgba(11,122,78,.3); transition:transform .12s ease; }
.content .snapkin-stores a.snapkin-store:hover { transform:translateY(-1px); color:#fff !important; background:#0B7A4E; }
.snapkin-store svg { flex:none; }
.snapkin-store span { display:flex; flex-direction:column; align-items:flex-start; }
.snapkin-store small { font-size:10px; font-weight:600; letter-spacing:.04em; text-transform:uppercase; opacity:.9; }
&lt;/style&gt;
&lt;div class="snapkin-stores"&gt;
&lt;a class="snapkin-store" href="https://getsnapkin.app/go/store/apple?utm_source=ma.ttias.be&amp;utm_medium=referral&amp;utm_campaign=blogpost-introducing-snapkin"&gt;&lt;svg width="22" height="22" viewBox="0 0 24 24" aria-hidden="true"&gt;&lt;path d="M16.4 12.8c0-2.3 1.9-3.4 2-3.5-1.1-1.6-2.8-1.8-3.4-1.8-1.4-.15-2.8.85-3.5.85-.72 0-1.86-.83-3.06-.81-1.57.02-3.02.91-3.83 2.32-1.63 2.83-.42 7.02 1.17 9.32.78 1.12 1.7 2.38 2.92 2.34 1.17-.05 1.61-.76 3.03-.76 1.41 0 1.82.76 3.06.73 1.26-.02 2.06-1.14 2.83-2.27.89-1.3 1.26-2.56 1.28-2.63-.03-.01-2.46-.94-2.48-3.74zM14.1 5.9c.64-.78 1.07-1.86.95-2.94-.92.04-2.03.61-2.69 1.38-.59.69-1.11 1.79-.97 2.85 1.03.08 2.07-.52 2.71-1.29z" fill="currentColor"/&gt;&lt;/svg&gt;&lt;span&gt;&lt;small&gt;Download on the&lt;/small&gt;App Store&lt;/span&gt;&lt;/a&gt;
&lt;a class="snapkin-store" href="https://getsnapkin.app/go/store/google?utm_source=ma.ttias.be&amp;utm_medium=referral&amp;utm_campaign=blogpost-introducing-snapkin"&gt;&lt;svg width="21" height="21" viewBox="0 0 24 24" aria-hidden="true"&gt;&lt;path d="M3.6 2.6a1 1 0 0 0-.6.92v16.96a1 1 0 0 0 .6.92l9.2-9.4zM14.2 10.6l2.9-2.96-11.4-6.5zM14.2 13.4l2.9 2.96-11.4 6.5zM18.4 8.4l2.6 1.5a1.3 1.3 0 0 1 0 2.2l-2.6 1.5-2.05-2.6z" fill="currentColor"/&gt;&lt;/svg&gt;&lt;span&gt;&lt;small&gt;Get it on&lt;/small&gt;Google Play&lt;/span&gt;&lt;/a&gt;
&lt;/div&gt;
&lt;p&gt;For the past few months, I&amp;rsquo;ve been working on a calorie tracker. I originally built it just for me.&lt;/p&gt;</description></item><item><title>9000 RPM is faster than your screen refreshes</title><link>https://ma.ttias.be/9000-rpm/</link><pubDate>Wed, 02 Sep 2026 00:00:00 +0000</pubDate><author>mattias@ma.ttias.be (Mattias Geniar)</author><guid>https://ma.ttias.be/9000-rpm/</guid><description>&lt;p&gt;9000 RPM. That&amp;rsquo;s how fast a modern sports car engine can rev.&lt;/p&gt;
&lt;p&gt;I know that number, but I can&amp;rsquo;t quite &lt;em&gt;visualize&lt;/em&gt; it. The math is simple: 9000 revolutions per minute is 150 per &lt;em&gt;second&lt;/em&gt;.&lt;/p&gt;</description></item><item><title>PWAs: Personal Web Apps</title><link>https://ma.ttias.be/pwas-personal-web-apps/</link><pubDate>Wed, 19 Aug 2026 00:00:00 +0000</pubDate><author>mattias@ma.ttias.be (Mattias Geniar)</author><guid>https://ma.ttias.be/pwas-personal-web-apps/</guid><description>&lt;p&gt;I&amp;rsquo;ve found my ideal format for shipping apps for personal &amp;amp; family use: PWAs, with an offline-first focus, rendering (mostly) entirely client-side. JavaScript is powerful enough to do pretty much anything these days - especially if you don&amp;rsquo;t have to write the code yourself.&lt;/p&gt;</description></item><item><title>Deduplicating git worktrees on EXT4 with jdupes</title><link>https://ma.ttias.be/deduplicating-git-worktrees-ext4-jdupes/</link><pubDate>Fri, 24 Jul 2026 00:00:00 +0000</pubDate><author>mattias@ma.ttias.be (Mattias Geniar)</author><guid>https://ma.ttias.be/deduplicating-git-worktrees-ext4-jdupes/</guid><description>&lt;p&gt;A git worktree workflow eats a lot more disk than you&amp;rsquo;d expect. On the &lt;a href="https://ma.ttias.be/remote-coding-environment-vps/"&gt;remote box I run my coding agents on&lt;/a&gt;
, ten worktrees took a 28 GB volume down to 687 MB free in about a day. Here&amp;rsquo;s one way to limit that.&lt;/p&gt;</description></item><item><title>Setting up a remote environment for agentic coding</title><link>https://ma.ttias.be/remote-coding-environment-vps/</link><pubDate>Thu, 23 Jul 2026 00:00:00 +0000</pubDate><author>mattias@ma.ttias.be (Mattias Geniar)</author><guid>https://ma.ttias.be/remote-coding-environment-vps/</guid><description>&lt;p&gt;For the last few months my laptop has been the bottleneck in my own workflow. Not the CPU, not the RAM &amp;ndash; the fact that it&amp;rsquo;s a &lt;em&gt;laptop&lt;/em&gt;.&lt;/p&gt;</description></item><item><title>Merging multiple EPUBs into a single file for Amazon Kindle</title><link>https://ma.ttias.be/merging-epubs-single-kindle-file/</link><pubDate>Sun, 12 Jul 2026 00:00:00 +0000</pubDate><author>mattias@ma.ttias.be (Mattias Geniar)</author><guid>https://ma.ttias.be/merging-epubs-single-kindle-file/</guid><description>&lt;p&gt;Alex West wrote &lt;a href="https://alexwest.co/" target="_blank" rel="noopener noreferrer"&gt;&lt;em&gt;Not Business Advice&lt;/em&gt;&lt;/a&gt;
, a free series of 13 short books about building profitable personal projects, compiled from the daily blog posts he&amp;rsquo;s written since 2018. It&amp;rsquo;s the story of how he made over $1M from his own side projects, and I&amp;rsquo;m reading it right now. It&amp;rsquo;s &lt;em&gt;really&lt;/em&gt; good: honest about the failure, luck, and burnout, not just the wins. Go read it, it&amp;rsquo;s free.&lt;/p&gt;</description></item><item><title>What 1.8 million real website outages look like</title><link>https://ma.ttias.be/website-uptime-statistics/</link><pubDate>Mon, 29 Jun 2026 00:00:00 +0000</pubDate><author>mattias@ma.ttias.be (Mattias Geniar)</author><guid>https://ma.ttias.be/website-uptime-statistics/</guid><description>&lt;p&gt;We watch websites for a living at &lt;a href="https://ohdear.app/?utm_source=ma.ttias.be&amp;amp;utm_medium=referral&amp;amp;utm_campaign=blogpost-website-uptime-statistics" target="_blank" rel="noopener noreferrer"&gt;Oh Dear&lt;/a&gt;
, which puts us in a unique position: we have data on a lot of different outages.&lt;/p&gt;</description></item><item><title>Dealing with concurrent bridge-network creates &amp; host-port races in Docker</title><link>https://ma.ttias.be/docker-bridge-network-host-port-races/</link><pubDate>Mon, 22 Jun 2026 00:00:00 +0000</pubDate><author>mattias@ma.ttias.be (Mattias Geniar)</author><guid>https://ma.ttias.be/docker-bridge-network-host-port-races/</guid><description>&lt;p&gt;I&amp;rsquo;ve been moving our CI off GitHub-hosted runners onto our own arm64 hardware. The plan was straightforward: a pool of ephemeral runners on a dedicated CI box, and each test shard spins up its own MySQL, Redis, and ClickHouse as service containers, all under rootless Docker.&lt;/p&gt;</description></item><item><title>Laravel Octane (FrankenPHP) vs. PHP-FPM: what I measured, and why we went back</title><link>https://ma.ttias.be/laravel-octane-vs-php-fpm-lessons-learned/</link><pubDate>Mon, 22 Jun 2026 00:00:00 +0000</pubDate><author>mattias@ma.ttias.be (Mattias Geniar)</author><guid>https://ma.ttias.be/laravel-octane-vs-php-fpm-lessons-learned/</guid><description>&lt;p&gt;Laravel Octane promises a faster Laravel: keep your app booted in memory, skip the framework bootstrap on every request, do less work per request. We run a fleet of uptime-checker boxes at Oh Dear that do nothing but fire HTTP, TCP, and ping checks at the sites we monitor, so I figured they&amp;rsquo;d be a good, low-risk place to try it. We moved a handful of them from PHP-FPM to Octane on &lt;a href="https://frankenphp.dev/" target="_blank" rel="noopener noreferrer"&gt;FrankenPHP&lt;/a&gt;
, measured it properly, and then moved every one of them back.&lt;/p&gt;</description></item><item><title>Turn your Filament panel into an AI-native dashboard</title><link>https://ma.ttias.be/filament-mcp/</link><pubDate>Mon, 22 Jun 2026 00:00:00 +0000</pubDate><author>mattias@ma.ttias.be (Mattias Geniar)</author><guid>https://ma.ttias.be/filament-mcp/</guid><description>&lt;p&gt;You can already point an AI agent at your Filament admin panel and have it click around like a human. Take a screenshot, find the button, fill the form, submit, screenshot again to check it worked. With computer-use or a browser MCP, that works today. It&amp;rsquo;s also slow, brittle, and spends a small fortune in tokens describing pixels to figure out where the &amp;ldquo;Save&amp;rdquo; button is.&lt;/p&gt;</description></item><item><title>Letting Cloudflare serve Brotli-compressed data with nginx</title><link>https://ma.ttias.be/letting-cloudflare-serve-brotli-with-nginx/</link><pubDate>Mon, 15 Jun 2026 00:00:00 +0000</pubDate><author>mattias@ma.ttias.be (Mattias Geniar)</author><guid>https://ma.ttias.be/letting-cloudflare-serve-brotli-with-nginx/</guid><description>&lt;p&gt;Google PageSpeed flagged a stylesheet on ohdear.app the other day: &lt;code&gt;front.css&lt;/code&gt;, the bundle for its public-facing pages. That surprised me, because Cloudflare sits in front of ohdear.app and Cloudflare does Brotli, and a Brotli&amp;rsquo;d copy of that file is around 40 KB. Browsers were downloading just over 50 KB of gzip instead. So why wasn&amp;rsquo;t Brotli kicking in?&lt;/p&gt;</description></item><item><title>Linux debugging tools I use daily</title><link>https://ma.ttias.be/linux-debugging-tools-i-use-daily/</link><pubDate>Fri, 12 Jun 2026 00:00:00 +0000</pubDate><author>mattias@ma.ttias.be (Mattias Geniar)</author><guid>https://ma.ttias.be/linux-debugging-tools-i-use-daily/</guid><description>&lt;p&gt;Every server I run, including the fleet behind &lt;a href="https://ohdear.app?utm_source=ma.ttias.be&amp;amp;utm_medium=referral&amp;amp;utm_campaign=blogpost-linux-debugging-tools-i-use-daily" target="_blank" rel="noopener noreferrer"&gt;Oh Dear&lt;/a&gt;
&amp;rsquo;s uptime checks, has the same set of debugging tools installed before anything goes wrong. Not because the application needs them, but because the one time you reach for &lt;code&gt;strace&lt;/code&gt; is at 2am, the box is misbehaving, and &lt;code&gt;apt install&lt;/code&gt; is the last thing you want to be doing while you work out what broke.&lt;/p&gt;</description></item><item><title>The backup SSH daemon I run before every do-release-upgrade</title><link>https://ma.ttias.be/backup-sshd-do-release-upgrade/</link><pubDate>Wed, 10 Jun 2026 00:00:00 +0000</pubDate><author>mattias@ma.ttias.be (Mattias Geniar)</author><guid>https://ma.ttias.be/backup-sshd-do-release-upgrade/</guid><description>&lt;p&gt;I spent a chunk of the last few weeks upgrading a fleet of Ubuntu servers in place, one LTS to the next, with &lt;code&gt;do-release-upgrade&lt;/code&gt;. Dozens of boxes, mostly stateless, mostly boring once you&amp;rsquo;ve done the first few.&lt;/p&gt;</description></item><item><title>AI is a confidence booster</title><link>https://ma.ttias.be/ai-is-a-confidence-booster/</link><pubDate>Tue, 09 Jun 2026 00:00:00 +0000</pubDate><author>mattias@ma.ttias.be (Mattias Geniar)</author><guid>https://ma.ttias.be/ai-is-a-confidence-booster/</guid><description>&lt;p&gt;&lt;a href="https://www.acquisition.com/bio-alex" target="_blank" rel="noopener noreferrer"&gt;Alex Hormozi&lt;/a&gt;
has a solid definition of &lt;em&gt;confidence&lt;/em&gt;:&lt;/p&gt;
&lt;blockquote&gt;
&lt;p&gt;You don’t become confident by shouting affirmations in the mirror, but by having a stack of undeniable proof that you are who you say you are.&lt;/p&gt;</description></item></channel></rss>