emanuele-em

11 ms·

1.
by emanuele-em 5mo ago
for parsers, malformed input probably has to be part of the actual spec, not just an edge case. A small set of bad-input tests plus fuzzing seems more useful than chasing the last few percent of line coverage
2.
by emanuele-em 5mo ago
Tables and images are the part where every "just use a rope" answer falls apart, so going B-tree feels right. I tried building a minimal rich text editor last year and got stuck exactly at the point where tables stopped being atta
3.
by emanuele-em 5mo ago
cmd+backtick breaks the second you minimize a window, been annoyed by this forever. nice to see a fix.
4.
by emanuele-em 5mo ago
yes I totally agree, my message was a compliment :)
5.
by emanuele-em 5mo ago
The 16MB IndexedDB cap per file is going to be tight if anyone tries loading the Atomic Edition with all the user maps, but honestly the fact that this boots at all in a browser with working MIDI is wild.
6.
by emanuele-em 5mo ago
Why Zig 0.15.2 specifically for the vendored libghostty-vt, is that tracking Ghostty upstream or did you pin it for a reason?
7.
by emanuele-em 5mo ago
Per-bucket DEKs with HKDF, hashed policy keys to kill enumeration, HMAC audit chain. This is the kind of boring-correct crypto design I rarely see in Go libraries. memguard for the master key is a nice touch too.
8.
by emanuele-em 5mo ago
Interesting that you went the full custom-language route instead of just better gopls completion inside string literals. The tree-sitter grammar plus language server proxying to gopls is a lot of moving parts to maintain. Curious how you ha
9.
by emanuele-em 5mo ago
How does the manual approval flow work when running this in CI? If a scheduled experiment fires and nobody approves within the TTL window, does the run just get marked expired or does it also block future scheduled runs from firing?
10.
by emanuele-em 5mo ago
Replacing ptrace with LD_PRELOAD is the obvious move but the devil is in the details. How do you handle things like /proc/self/exe emulation and statfs without a kernel-level hook? Proot's been painfully slow on phones f
11.
by emanuele-em 5mo ago
Six protocol boundaries between psql and S3 and it just works? What's the write latency like in practice?
12.
by emanuele-em 5mo ago
I've wasted so many hours doing the /etc/hosts + mkcert dance by hand and forgetting to clean up after. Having Pingora handle the proxy layer with longest-prefix routing and automatic cleanup on shutdown is exactly right. The
13.
by emanuele-em 5mo ago
The bash permission parsing is what sold me. Most agents treat `git diff && rm -rf /` as one blob, using tree-sitter to split that into separate permission requests is smart.
14.
by emanuele-em 5mo ago
The 10-second correlation window for file access + network connection is a smart heuristic for catching exfiltration. Scanning /proc/pid/fd every few seconds is lightweight enough to not annoy people. How granular is the netw
15.
by emanuele-em 5mo ago
Finally a single tool that handles Snowflake presets for Twitter/Discord/Instagram. The pipe support is great too, `idt gen ulid -n 5 | idt sort --show-time` is exactly the kind of composability you want in a CLI. Written in Rust?
16.
by emanuele-em 5mo ago
The clock injection for testing temporal nodes is a really nice touch. Most BT libraries force you to actually wait or mock everything yourself. This would work well for orchestrating retries and fallback logic in microservices, not just ga
17.
by emanuele-em 5mo ago
Really cool to see someone actually prove that the NVIDIA vs Apple efficiency gap is mostly a software problem. A 2020 GPU matching M5 Max tok/J at 1.8x the throughput just by fusing all 24 layers into one persistent kernel is a strong
18.
by emanuele-em 5mo ago
That feature comparison table is brutal for httpx and aiohttp. HTTP/3 over QUIC, DNSSEC, post-quantum, and still 4x faster than httpx? How stable has QUIC negotiation been with the big CDNs in practice?
19.
by emanuele-em 5mo ago
Smart call on the tiered lookup, hitting SQLite first and falling back to FLOPs/TFLOPS estimation. One thing I'm wondering about the 20% overhead in Tier 2, does that factor in framework overhead or just raw model weights? That ma
20.
by emanuele-em 5mo ago
What made you go with a full GUI over a TUI for this? I usually just chain ss and lsof in a terminal and call it a day, so I'm curious if having the port table always visible changed how you debug network stuff.
21.
by emanuele-em 5mo ago
This is a great Easy68K replacement. MOVEM and displacement addressing working in the browser with full undo/redo is already more than what most students get on native tools. If you haven't already, consider adding breakpoints, th
22.
by emanuele-em 5mo ago
Wait, so you're dlopen-ing into MultitouchSupport.framework because arm64e PAC breaks direct extern calls? That's a great find. The braille heatmap in the terminal is super cool too. How often does a macOS update break things give
23.
by emanuele-em 5mo ago
301 bytes! The base64 one-liner install is a nice flex. Accepting an infinite loop when energy_full doesn't exist is peak code golf, perfectly reasonable when every byte counts. Is there a writeup on the assembly somewhere?
24.
by emanuele-em 5mo ago
Honestly this is way more appealing than fighting mergerfs when you just want explicit disk placement. Doctor + prune for orphaned symlinks is exactly what you'd need to keep things sane over time. Saw it's written in Zig, how&#x2
25.
by emanuele-em 5mo ago
Ok the MUL workaround got me. MUL does (a b)/64 so you have to DIV first to get a 64, then MUL finally gives you a*b. And recursive FDEFs because there's no WHILE? All in 6.5KB? What kind of frame rate do you actually get out of t
26.
by emanuele-em 5mo ago
Nice and minimal. How does scanning speed compare to dust or ncdu on large trees? The "hide largest item to compare smaller ones" interaction is clever, haven't seen that in other du tools.
27.
by emanuele-em 5mo ago
This is seriously impressive for a single project. Rust + CUDA kernels for MD, fallback to SIMD thread pools, pharmacokinetics inference, docking, and it compiles to a standalone binary? How does the MD throughput compare to GROMACS on equi
28.
by emanuele-em 5mo ago
Using pubgrub for resolution is the right move. 18s down to 8.5s is nice for something this early. Did you port the algorithm from pubgrub-rs or reimplement from the paper? Also curious how you plan to handle composer plugin hooks, that&#x2
29.
by emanuele-em 5mo ago
sorry, did't see it. Great
30.
by emanuele-em 5mo ago
The finding that naive single-op benchmarks overestimate dispatch cost by ~20x is wild. Curious how much the torch-webgpu backend could close the gap with CUDA if you went aggressive on kernel fusion, 53% improvement on Vulkan already is si

More ›