wolf550e

https://github.com/rustls/rustls-ffi I guess

Frank Denis, maintainer of libsodium, is also a reputable engineer.

Didn't they invent MTA-STS so they could not use DANE? They use MTA-STS.

They prefer tcl as the scripting language and they like the maintainer.

I don't think they are actually selling those machines so well. They have a captive legacy customer base, who else is buying those?

The amount of throughput you can get out of AMD EPYC zen5 servers for the price of a basic mainframe is insane. Even if IBM wins in single core perf using absurd amount of cache and absurd cooling solution, the total rack throughput is definitely won by "commodity" hardware.

They also removed the gridfins and their engines.

https://blogs.nasa.gov/europaclipper/wp-content/uploads/site...

The Case for Mars is the book about Mars Direct

https://en.wikipedia.org/wiki/The_Case_for_Mars

The difference is that Mars Direct was a conservative plan, wrt rocket technology. It just used existing Shuttle parts and was meant to fly from Shuttle pads in between Shuttle flights. It didn't require developing anything new, that's why they were sure it would be cheap and fast, and flight rate would make it possible for it to be reliable. SLS is that idea turned into a jobs program.

Robert Zubrin's Mars Direct would have got "flags and footprints" on Mars without needing something like Starship. He worked on a version of SLS in 1989 in Martin Marietta that would have cost $1B and been ready in the second half of the 90s and would be retiring soon after many successful flight to Mars (and to the Moon), but instead we got SLS and Orion so we need Starship to get anything done.

They dropped transpiration cooling, at least for now.

It updated my corp laptop a few days ago

Sorry, bad reading comprehension on my part! I see honorary PhDs often used to express political support, instead of honest admiration of achievements.

It's not original research. A Master's Thesis at most, or only a Bachelor's Degree Final Project. Regardless, doing such amazing work for a blog (and geek fame) is amazing.

What are "real types and the associated constraints"? It has strict tables:

https://www.sqlite.org/stricttables.html

I don't know Rust. Can someone explain how this data structure stores whether a node is itself a valid word or whether it only leads to more nodes? For example the node "do" in their (“and”, “ant”, “dad”, “do”, & “dot”) example. I guess it's a discriminated union provided by Rust algebraic data types or something like that, but why not show the full bit pattern in memory?

I get the feeling Intel needed to "bribe" the Navajo by ordering this work. Why did they need to do that? Was it part of an agreement with tribe leadership?

The macOS IO benchmark and pgbench (also IO) makes me think macOS cheats somehow, for example it doesn't actually cause the SSD to write to flash when the app performs fsync, and some other "optimizations".

The adapters are installed on the station, the vehicles don't carry them.

https://en.wikipedia.org/wiki/International_Docking_Adapter

VSS6 had file checkout, which IIRC locked the file preventing other devs from editing it. It was popular with Visual C++ 6.0, circa the year 2000.

https://en.wikipedia.org/wiki/Microsoft_Visual_SourceSafe

That's one CVE, right? How many other vulnerabilities were caused by compiler optimizations, whether they were bugs in the compiler or allowed by the spec?

> It would be interesting to study what percentage of security failures can be partly or entirely attributed to compiler "optimizations".

I bet it's roughly none.

But you can use external storage for code and data, and that can be bigger than what can be addressed by the pointer size. The universe is finite, but most computations you would want to perform treat AWS S3 as infinite. Unless you want to brute force AES-128 or something.

regehr et al use alive2 which uses z3

You can follow Hector Martin @marcan at https://social.treehouse.systems/@marcan/

He works on Asahi Linux, a Linux port to arm64 Apple hardware.

I think the trick with dereferencing unmapped memory is cool, but I only really care about techniques that work reliably and I can use in production.

It is possible to construct a worst case scenario for bytecode execution, for example very complex expressions in the WHERE and/or SELECT clauses that compute values, and a query plan that performs a full table scan over a table with say 100 million rows that is cached in RAM (or maybe use generate_series, whatever works best).

Computing the expressions should dominate execution time, right?

Then, to compare against the best possible case, we can write a custom C program that uses sqlite internals to perform the same task (full scan of the table, extract values from row) and does not use the bytecode VM and computes the complex expressions in regular C code (e.g. a function that accepts floats and returns a float or whatever).

Then comparing the two implementations will tell us how much faster sqlite can be if it had a "perfect" JIT.

The page is the result of this exchange on Twitter:

https://twitter.com/gorilla0513/status/1784756577465200740

I was surprised to receive a reply from you, the author. Thank you :) Since I'm a novice with both compilers and databases, could you tell me what the advantages and disadvantages are of using a VM with SQLite?

https://twitter.com/DRichardHipp/status/1784783482788413491

It is difficult to summarize the advantages and disadvantages of byte-code versus AST for SQL in a tweet. I need to write a new page on this topic for the SQLite documentation. Please remind me if something does not appear in about a week.

Eh.

Russian folk wisdom says "Жид крещёный - что вор прощеный. Веры нет" - "A converted Jew (slur) is like a forgiven thief - no faith/trust".

And in later times, "бить будут не по паспорту, а по морде" - you'll get beaten up on your face, not on (according to) your identification papers, a play on words meaning if you look Jewish, it doesn't matter that your papers say otherwise.