huge wrench in my plans. python only supports building against openssl which is a much larger surface area than curl with gnutls
Discussion
Loading...
however i'm realizing this is may also possibly be exactly what i want. because while python's pathlib is easy the best vfs interface, its networking protocols are outrageously unfit
this spans the range from parsing json, to url quoting, to its intentionally broken "email" parsing, to its disastrous reliance on implicit context like the User-Agent string sent to pypi upon every request, which after 5 years of working on pip came as a revelation to me
a revelation in the most literal sense that the word admits. i have spent many hours and finally succeeded at deriving the simplest version of a type system that mypy and corporate greed stole away from me all those years ago. it ends up looking like pydantic on the surface but with significantly more care for crafting the transitions between states such that an unambiguous unidirectional data flow is achieved and described
this is significant because it is precisely such transitions which the pants v2 engine fails to provide language to express. instead, it has synchronous and asynchronous (coroutine) paths in the graph. it has no conception of resource allocation or RAII. it has no conception of failure, and therefore not of trying again
this is not because we had failed to consider whether such affordances would enable safer faster more robust build tasks. Retry (a wrapper over a path in the task graph which could be separately tracked) was one of the several incredibly compact prototypes i developed (along with the composable --query syntax) which were poorly received
it remains shockingly clear that the model of pure functions alone cannot hold. the embarrassing parallelism for scala compiles that win wang and i developed the first successful model for was very specifically enabled by codifying two separate forms of persistent state, locally and remotely, allocating a local jvm jit to run eugene burmako's rsc and allocating remote scoot workers from drew gassaway's squadron within a compile context that enabled them to temporally and spatially share resources for the so-called "ephemeral" scalac workers
~128 more replies (not shown)
it's very important that the task graph is able to represent more than pure functions because in fact literally every single i/o operation involves negotiating shared access to a limited resource. this is what makes it an i/o operation.
when i read that the zstd project categorically refuses to consider any benchmark outside of purely in-memory operations (they say this on the github in PRs, they're very open about it and it is used to justify extremely precise modifications to the code) was when i realized that project is not worth forking
resource pooling is not really an optimization technique so much as it is the bare minimum prerequisite to having any idea of what the system you've built is running up against. "explicit vs implicit" is far too shallow of a distinction to do justice to the difference in behavior of e.g. a rust program doing i/o with tokio's threaded async executor, vs a rust program that directs each thread to perform its own hot loop and simply sleeps until there's more data https://github.com/zip-rs/zip2/pull/236
coroutines themselves are not the problem. in fact it's well understood that ring buffers from the os are not just mario pipes but indeed serve the purpose of flow control with "backpressure". through coffeescript and atom-shell (electron before it became evil) i found node.js to have developed a wonderfully rich appreciation across the community for how to write in essentially continuation-passing style. it's a big relay race! we're doing it together! no wonder rust did the big rewrite from libuv barely 10 years ago!
but there is another presence in this world. not another tarnished that the light of grace would touch as prophesied by the elden ring cutscene, whose adventure begins at level 5, who uses a mirror to change their gender on demand
james baldwin wrote this 64 years ago https://www.newyorker.com/magazine/1962/11/17/letter-from-a-region-in-my-mind and it will haunt me long after the unix epoch transgresses 64 bits in seconds dulling our collective memory of the nixon presidency
Every effort made by the child’s elders to prepare him for a fate from which they cannot protect him causes him secretly, in terror, to begin to await, without knowing that he is doing so, his mysterious and inexorable punishment. He must be “good” not only in order to please his parents and not only to avoid being punished by them; behind their authority stands another, nameless and impersonal, infinitely harder to please, and bottomlessly cruel.
this sentential clause is written into the folds of my cranium— not burned or etched, but cryptographically signed into the ROM sector, the inner workings of it exposed by a clever reverse engineer
He does not know what the boundary is, and he can get no explanation of it, which is frightening enough, but the fear he hears in the voices of his elders is more frightening still. The fear that I heard in my father’s voice, for example, when he realized that I really believed I could do anything a white boy could do, and had every intention of proving it, was not at all like the fear I heard when one of us was ill or had fallen down the stairs or strayed too far from the house. It was another fear, a fear that the child, in challenging the white world’s assumptions, was putting himself in the path of destruction.
A child cannot, thank Heaven, know how vast and how merciless is the nature of power, with what unbelievable cruelty people treat each other. He reacts to the fear in his parents’ voices because his parents hold up the world for him and he has no protection without them.
the following takes the form of a thought experiment: perhaps of the form a nazi physicist might propose, in which he claims that the stochastic behavior einstein challenged until his death is responsible for whether a creature placed into his care lives or dies. schrödinger suddenly shows up with a flask of poison and starts demanding einstein take him seriously
have you ever looked up which poison he proposes for the act? it's a very specific poison. there are many poisons you could imagine might be able to kill a cat. but to erwin schrödinger the cat is largely immaterial to the entire enterprise
recall again that this is in response to einstein's scientific paper with two other jewish scientists https://en.wikipedia.org/wiki/Schr%C3%B6dinger's_cat
One can contrive even completely burlesque [farcical] cases. A cat is put in a steel chamber along with the following infernal device (which must be secured against direct interference by the cat): in a Geiger counter, there is a tiny amount of radioactive substance, so tiny that in the course of an hour one of the atoms will perhaps decay, but also, with equal probability, that none of them will; if it does happen, the counter tube will discharge and through a relay release a hammer that will shatter a small flask of hydrocyanic acid.
wikipedia has his original text in german inline as a citation, which unambiguously describes hydrocyanic acid, using the standard scientific terminology https://en.wikipedia.org/wiki/Hydrogen_cyanide#As_a_poison_and_chemical_weapon
Perhaps its most infamous use is Zyklon B (German: Cyclone B, with the B standing for Blausäure – prussic acid; also, to distinguish it from an earlier product later known as Zyklon A),[72] used in the Nazi German extermination camps of Majdanek and Auschwitz-Birkenau during World War II to kill Jews and other persecuted minorities en masse as part of their Final Solution genocide program.
none of the "interpretations" that are given scientific credence represent einstein's interpretation of the EPR thought experiment, which they devised and published together not as a philosophical quandary but a measurement of physical reality
when i attempted to request on the talk page where the work first received the derogatory term "paradox", i had my very minimal edits reverted with extreme prejudice https://en.wikipedia.org/wiki/Einstein%E2%80%93Podolsky%E2%80%93Rosen_paradox a day later the most obnoxiously authoritative replier followed up to say he believed it was schrödinger who said it. my addition of a template to request a source for this non-neutral point of view remains effectively censored. this is what einstein worked on for decades. it's not nothing
my thought experiment will make no claim to rival james baldwin nor albert einstein, nor their searing observation of what may be fancifully conflated to constitute the celestial mechanics of white supremacy. this name invokes the theory of epicycles, a historically controversial approach to orbital mechanics which baby physicists are taught as one of their first worked examples motivating their modern smug superiority over the past https://en.wikipedia.org/wiki/Deferent_and_epicycle#As_an_example_of_bad_science
galileo is considered a sympathetic figure in the history of epicycles. i was specifically told in school that he was "excommunicated" from the church, which i was informed constituted the harshest possible punishment. that appears to have been false, but what's more interesting is what incites that response. https://en.wikipedia.org/wiki/Galileo_affair
so the pope was quite literally a galileo patron and gave him explicit consent to publish on geocentrism "as long as he treated it as a hypothesis". given that telescopes in 1632 were limited to areas of clear sky with visible light and the robustness of ground glass (glass was still the weakest link for the hubble space telescope 400 years later amidst the cold war), the pope's admonishment is less harsh than modern funders of research into the dynamics of slop machines
but galileo had decided any resistance be it philosophical or even scientific critique was indicative of what he deemed an inferior intellect. he would go a lot further than needed with this. it was clearly really important to him:
Galileo complained that some of the philosophers who opposed his discoveries had refused even to look through a telescope:[11]
slopscope. teleslop
he would write a treatise regarding "the two chief world systems" https://en.wikipedia.org/wiki/Dialogue_Concerning_the_Two_Chief_World_Systems i mentioned the cold war earlier. bear this in mind. galileo would be wrong in many ways here, such as his impressive disregard for non-circular orbits. elliptic curves are more complex—and correct. what gets him into hot water was when he named a character after an ableist insult and used them to represent geocentrism
the term a priori is used here to trash aristotle. the only time i've ever seen that phrase italicized before in my entire life is the 1977 ziv and lempel compression paper which aggressively fails to cite shannon, claims to be independent of probabilistic assumptions, and is responsible for compression slop machines that are now the only type of compression anyone fucking uses
He adds "we possess a much better basis for reasoning about celestial things than Aristotle did...Now we, thanks to the telescope, have brought the heavens thirty or forty times closer to us than they were to Aristotle, so that we can discern many things in them that he could not see;"
thirty or forty times. this guy can't be bothered to read kepler's book kepler had personally sent to him which calculates the elliptical orbit of mars because he thinks seeing "30 or 40 times closer" is enough to start writing about the two competing world systems. elon musk behavior
in the immortal words of robert pike https://go.dev/talks/2012/splash.article "it's still bigger than we expected"
For the record, we measured the compilation of a large Google program written in Go to see how the source code fanout compared to the C++ analysis done earlier. We found it was about 40X, which is fifty times better than C++ (as well as being simpler and hence faster to process), but it's still bigger than we expected.
40X, which is (as we all know) fifty times better than C++
