Notices by Tony Finch (fanf@mendeddrum.org)

from my link log —

High performance correctly rounded math libraries for 32-bit floating point.

https://blog.sigplan.org/2021/08/26/high-performance-correctly-rounded-math-libraries-for-32-bit-floating-point-representations/

saved 2021-08-27 https://dotat.at/:/IV9QT.html

@anildash was reddit the first popular website with markdown support? i had the vague idea aaron swartz added markdown to reddit when he was working on it, and it spread from there … or have i got the wrong impression?

the rise and fall of stackoverflow

the graph of posts per month is quite something

the decline over the last four years is almost a mirror of the rise 2008-2012

and they are now down to near zero

https://data.stackexchange.com/stackoverflow/query/1926661#graph

from my link log —

"DevOps didn't exist when I started as a developer."

https://circleci.com/blog/devops-did-not-exist/

saved 2019-09-13 https://dotat.at/:/THY3G.html

@whitequark oh dear lolsob

https://go.googlesource.com/go/+/refs/heads/master/src/net/dnsclient.go#75

@whitequark good question

i dunno the go libraries well enough, but i’d probably look around to see if there’s existing an existing hostname checker i could clone and hack

but a regex would do if i didn’t find code i like (i have used them in the past for this, but without _, heh)

one tricky thing is hostnames have per-label length limits which a regex can handle ok, but also an overall length limit which is probably best done as a separate check

oh wait, the tr46 algorithm includes length checks already!

so it might be reasonable to just check [a-z0-9._-]+

(sorry that turned into a ramble)

@whitequark yeah idna is difficult

i think the key sentence is “This option corresponds to the UseSTD3ASCIIRules flag in UTS #46.” and in tr46 that flag basically enables or disables all hostname syntax checks, hence allowing * and everything else i bet

the approach i would take is to do idna conversion with UseSTD3ASCIIRules=false then check that the resulting punycode obeys the relaxed hostname-with-underscore syntax

on the basis that idna is designed so that it works when trad syntax is enforced by various lower layers

from my link log —

Why ed(1)?

https://blog.thechases.com/posts/cli/why-ed1/

saved 2025-12-04 https://dotat.at/:/X2EXL.html

@joe @dotstdy @zwarich yeah i sketched a simplified UCS along those lines a few months ago which i called (tongue in cheek) the penultimate conditional syntax https://dotat.at/@/2025-05-13-if-is.html

definitely not the last word on the matter, in particular it needs less noisy punctuation

@whitequark yeah pem is surprisingly old and influential

hmm that wikipedia article undersells it, the first pem rfc was published in 1987 which i find kind of incredible https://www.rfc-editor.org/rfc/rfc989 tho it is mostly the first definition of base64 and ascii armoring, the DER business came later

john venn’s memorial window in gonville & caius college dining hall

@leah @dalias @ayke reminds me of one of the first y2038 bugs that occurred when a sleep 1000000000 stopped sleeping

from my link log —

EBCDIC is incompatible with GDPR.

https://shkspr.mobi/blog/2021/10/ebcdic-is-incompatible-with-gdpr/

saved 2024-10-28 https://dotat.at/:/J1ZQP.html

on my blog!

performance of random floats

in which i examine the relative cost of bithacking floating point numbers vs more conventional conversions

https://dotat.at/@/2025-06-08-floats.html

oh good grief i have discovered that a better compiler significantly changes the results on arm64

also bloody amd64 is running my baseline slower than my functions that do work

how am i supposed to measure these things i ask you

hmm no, `cpuid` is not a good replacement for `isb sy`

aha!

i think i have succeeded by passing arguments and return values via pointers and putting an MFENCE just before and just after the code under test

RIGHT

i added a baseline to my benchmark which revealed my amd64 numbers were nonsense, so i revamped it and now the benchmark results are more plausible

but the conclusion of the blog post is rather different!

newly rewritten article now up at the same url https://dotat.at/@/2025-06-08-floats.html

hmm, Safari has enormous problems with the new Lobsters CSS grid layout

the slowness is visible in how long it takes to reflow when the window is resized (on desktop)

also typing speed in the comment box is like 1 character per second

this page is particularly bad https://lobste.rs/s/ewinm7/on_bloat

it takes about 17 seconds to render on my iPad or 14s on my iPhone

@pushcx