Hacker News: merb

New comment by merb in "Typst makes big strides"

merb — Fri, 25 Sep 2026 21:56:57 +0000

I just compared typst with peachpdf (c#) and others to generate big reports, invoices, etc. coming from itext 4 the open source version which was lgpl. And was blown away. Itext was quite efficient because layouting did not do a full css thingy. Which peachpdf does, which of course produces a better looking pdf. But full css eval, could byte your ass with flow elements and div wrappers. You can be extremely non performant and ram hogging pretty fast, especially since peachpdf is not optimized.

However typst? God , it is so fast that downloading and rendering it inside a pdf viewer is the new bottleneck… it generates 2000 page pdfs in second with maximum of 2 gb , however most of the time it used 250mb memory. Peachpdf used 15-30gb for that, heck even itext used 4 gb. Cpu everything was so fucking fast, even while going through p/invoke.

I just fucking love it. I mean sometimes the syntax is weird, but it’s the best layout+data to pdf renderer.

New comment by merb in "Microsoft killed FoxPro in 2007. Anyway, here's FoxPro revived"

merb — Wed, 23 Sep 2026 04:25:02 +0000

Actually you can use foxpro with mssql and ole, you could than even add row level security for even more security

New comment by merb in "Performance Improvements in .NET 11"

merb — Thu, 17 Sep 2026 00:33:11 +0000

With Claude I never had that problem. However Claude does not use the pattern by default

New comment by merb in "Alternatives to MinIO for single-node local S3"

merb — Tue, 15 Sep 2026 17:00:47 +0000

I use that aswell since it can be used and expose a posix filesystem. This is helpful when we want to add s3 to existing storage like a nas. (Well we had some problems with it in the past tough)

Sadly most new projects do not offer posix anymore.

New comment by merb in "We found a division by zero bug in FFmpeg with a vibecoded fuzzer"

merb — Thu, 27 Aug 2026 21:26:00 +0000

My plan for bigger things is mostly:

Generate multiple solutions- they do not to work 100% correctly. And than I check which I would prefer. Which is more to our applications taste.

And than I would take the vibe output as a kind of a ‚plan‘ which I use to implement but not follow 100% and at the end I take my solution and review it. I gain speed with that because I often can quickly see the pros and cons of a solution way better than when I would manually do it and hang on a major roadblock and also I even see such roadblocks in the vibe output - it’s mostly the part with an unnecessary amount of new code that looks nonsensical.

New comment by merb in "Walgit – a Git server that is one binary in front of an object store"

merb — Tue, 25 Aug 2026 04:34:40 +0000

Fun thing Google cloud storage supported that for a long time now. Just AWS took its time.

New comment by merb in "Splitting a Git Commit"

merb — Tue, 18 Aug 2026 16:50:29 +0000

Which gui? I’m on Mac and somehow I could never really be happy with gui‘s for git. Gitkraken is the last I tried and I found it too much cluttering. Paid choices weren’t that great aswell either (yet)

New comment by merb in "Linux 7.3 improves performance when running out of vRAM"

merb — Tue, 18 Aug 2026 16:45:32 +0000

On servers with k8s the default is preempt based on prio, followed by oom kill and ‚newly’ memory qos:

https://kubernetes.io/docs/concepts/scheduling-eviction/pod-...

https://kubernetes.io/docs/concepts/configuration/manage-res...

https://kubernetes.io/docs/concepts/workloads/pods/pod-qos/#... (New feature)

So normally if you overcommit it might be undefined behavior if you not have enough memory. However if everything is fine pods can preempt based on prio to new nodes.

With the right settings only the faulty app might fail. Without it it can be a real pain to find the rouges

New comment by merb in "Quake Shareware, a CD-ROM just a little too full"

merb — Tue, 18 Aug 2026 10:32:32 +0000

Some collectors editions do come without a game cd nowadays.

New comment by merb in "Quake Shareware, a CD-ROM just a little too full"

merb — Tue, 18 Aug 2026 04:12:14 +0000

there isn't a market for gamers buying games on DVDs these days. BlueRay the same.

Of course there are voice that do want real disks but sadly that time is over. Digital sales won.

New comment by merb in "Mistral OCR 4.1"

merb — Thu, 13 Aug 2026 21:30:56 +0000

It‘s also cheaper than the murican ones from Google, Amazon, …. And tesseract was an example. Heck you can go xberg and use paddleocr. Most often layout is less of a problem for ocr. Most often you need high accuracy, which tools like these are often worse in the 95 percentile.

New comment by merb in "Mistral OCR 4.1"

merb — Thu, 13 Aug 2026 19:03:25 +0000

Tbf even with tesseract you already get shit ton of accuracy and you can probably do these 1000 pages for way less than 3.5€. For 3.5€ you can spin up a cloud instance with 8vCPU+32gb on gcloud for 11 hours (or 11 instances for an hour) which can do way more than 1000 pages per hour on tesseract. It takes you around 6 second per page +-4 seconds start/stop depending on what you are doing on that instance size without too much optimization (you can probably even run multiple processes on a single node)

Google documentai costs 1.5$ per 1000 which is probably better in quality and speed.

New comment by merb in "Mistral OCR 4.1"

merb — Thu, 13 Aug 2026 17:52:13 +0000

1000 Pages / 3.5€ this is expensive as hell. If this is not fastly superior than something like tesseract it is not worth it.

New comment by merb in "Vlt 1.0 and Hosted Package Registries"

merb — Tue, 11 Aug 2026 19:52:29 +0000

A little late reply:

Preferably I do not use a client on my machine to publish a package. That’s what I meant, modern registries should always support working with oidc providers to support publishing credentials from ci/cd systems. Sometimes called trusted publishing: https://repos.openssf.org/trusted-publishers-for-all-package...

New comment by merb in "What I learned by putting GitHub Copilot behind a MitM proxy"

merb — Tue, 11 Aug 2026 17:22:08 +0000

It’s done via uprobes which allows ebpf to attach to users processes or libraries, basically like ld_preload but built into the kernel. First link that explains it https://blog.quarkslab.com/defeating-ebpf-uprobe-monitoring.... but there are many more links about it and it is a pretty useful tool for debugging in prod environments. (Also first link: https://blog.px.dev/ebpf-function-tracing/) this also the intended use case. px.dev is btw a cncf project which also helps for ebpf debugging in k8s (comes from new relic)

New comment by merb in "Vlt 1.0 and Hosted Package Registries"

merb — Tue, 04 Aug 2026 18:50:46 +0000

Without identity federation via oidc it’s basically not really a replacement for verdaccio

New comment by merb in "Claude Opus 5"

merb — Fri, 24 Jul 2026 17:04:47 +0000

Same as 4.8

New comment by merb in "Passkeys were invented by engineers with zero understanding of consumer brain"

merb — Wed, 22 Jul 2026 21:29:08 +0000

Wouldn’t it be better to use envelope encryption and sync the passkeys? I mean with envelope encryption somebody would probably be able to create a rfc that would support cross browser/password manager sync a passkey by using a encrypted envelope that would be able to be deceptively wither via tpm or by using credentials so when syncing it would ask for upn/password of the original device or something so it can even be e2e encrypted, so some kind of federation protocol for password managers that you own

New comment by merb in "'VPNs are lawful technical tools,' says EU Court in landmark copyright ruling"

merb — Tue, 21 Jul 2026 20:44:01 +0000

Btw this also could come with negative consequences, like that they want to force id checks to access copyrighted material

New comment by merb in "Rewriting Bun in Rust"

merb — Wed, 08 Jul 2026 23:36:15 +0000

1 Developer with a 200k budget for tools.