Jorropo
I break code, both mine and not mine. Mainly in Go. 🇫🇷
IPv6 maximalist
- Today the toolchain broke, so I've hexedit-ed the binary to fix a bug. « Everything is open source if you know assembly. » might not be as wrong as it looks.
- What they don't tell you about software engineering is that on 11 April 2025 at 04H07 UTC+2 you will spend 40 minutes stare at the screen thinking, read thousands of words across multiple issues to catch back up, do a couple greps to check how things were done just to review a 1 LINE CODE CHANGE.
- - Nobody: Are you felling good today ? - Me: 😒 - Nobody: Why ? - Me:
- Reposted by Jorropo[Not loaded yet]
- My life: - Lurk in an open source project. - Spend a couple hours coding a feature wanted for months. - 😎 - Turn out you need to spend 10 more minutes fixing an edge case. - It now only works for you because you're the only Arch user and Debian ships a 20-month-old version of libuv.
- Reposted by Jorropo@thepudds.bsky.social is my hero So far, up to 15x faster escape analysis for the Go compiler when compiling the TS checker package. End to end that's an 8x faster compiler of the checker package for us, bringing it down from 50 seconds to 6 seconds. github.com/golang/go/is...
- Reposted by JorropoTypeScript team: rewrites compiler in Go. Go community: what do you mean your new compiler takes more than a minute to compile? Unacceptable. Dishonorable even. We are so sorry for this sub par experience. Not how we do things around here. Two days later: WIP 5x speedup. HN: why pick Go anyway?
- Awesome work by @thepudds.bsky.social github.com/golang/go/is... Compiling the typescript compiler went from 70s to 30s !
- Reposted by JorropoToday we're thrilled to announce our effort to port the TypeScript compiler and language service to native code, gaining a 10x speed boost in build times and editor responsiveness! devblogs.microsoft.com/typescript/t...
- The most surprising thing in the typescript-go story is the number of C# devs feeling betrayed by Microsoft.
- Useless go fact: Untyped expressions are not limited to the const realm. Using << and >> operators on an untyped LHS with a variable RHS produce an untyped variable expression.
- The ESP32 backdoor conclusion from an offensive PoV: If you are running code in ring0* you can read, write and execute arbitrary code. *note that the Core used in the first ESP32 lacks an MMU like the vast majority of MCUs.
- Tomorrow's side quest, figure out why the first function compiles better than the second one: func good(x, y uint) uint { return min(x, y) } func good(x, y uint) uint { return min(y, x) }
- Today's mood: - 12h00 I guess I am writing Java - 15h00 Fuck Java why am I doing this ?
- So much fun waiting to be had, I just noticed the go compiler lacks RISCV64 C & B support even tho GORISCV64=rva2020u64 & rva2020u64 respectively allows us to use them.
- Wake up babe new Agile manifesto just dropped ! agvxov.github.io/fdd/fdd_mani...
- Ok hear me out, Factorio asynchronously autosave while playing by forking to create a COW of the memory space. What if we did the same thing for Go tests ? The scheduler would fork on each scheduling decision and run all paths, this allows to prove some concurrent is always correct via bruteforce.
- Hype new interprocedural analyzer for go1.25's optimizer: - Interprocedural backward register allocation go-review.googlesource.com/c/go/+/636839 - Side effects marking for functions go-review.googlesource.com/c/go/+/637936 - backward propagation of proofs (I'll code that later) - ...
- Tips: - Types, use them !
- I just tried github.com/Zygo/bees Really awesome piece of software that can only work thx to btrfs, more reasons why we need to stop advising EXT4 as the default config for systems.
- @why.bsky.team where is IPv6 support for bsky ? I feel the need to create a lack of IPv6 shaming website right now.
- In 6.1 ffmpeg will now ship with AM and FM transcoding using SDR hardware. I expect ffmpeg 7.4 to have dream reading capabilities by transcoding my brain waves to av1 if we continue at this rate.
- I was nerd snipped into writing a hash function exclusively using AES. I present to you a modern parallelisable, incrementally veriable merkle tree hash function: github.com/Jorropo/aesmthash/ Is it safe ? Hopefully. Is it fast ? Not really. Should you use it ? Definitely not.
