Nostr notes by sarah tonin :wlfBlep:

honestly why dont you call him out on it

2026-10-28T20:21:13Z

In reply to nevent1q…sf0x
_________________________

honestly why dont you call him out on it

hey that's how i worked on first semester of uni

2026-10-21T21:04:05Z

In reply to nevent1q…kdf5
_________________________

hey that's how i worked on first semester of uni

an array is an object and all objects are copied as references

2026-10-19T20:58:24Z

In reply to nevent1q…8d3d
_________________________

an array is an object and all objects are copied as references

by which banger this time

2026-10-18T19:47:24Z

In reply to nevent1q…uvn6
_________________________

by which banger this time

they are fucked enough but not loved enough. we need to cuddle ...

2026-10-16T16:59:06Z

In reply to nevent1q…sf0s
_________________________

they are fucked enough but not loved enough. we need to cuddle the computer.

probably not unit tests ? though there will be some. more end to ...

2026-10-15T18:31:55Z

In reply to nevent1q…qstk
_________________________

probably not unit tests ? though there will be some. more end to end tests, though there WILL be unit tests for a few parts (diffing / diff rendering / merge)

Nostr event nevent1qqs80h4yelnjmecasn5l39s4p644qq3vn23nxmf0rcnnxzsk6kj6efszyrv8snsjxhjemzs04upqcdw6c0z7sr8dap9kc5qfjthzkurz42lwkj2z4mn

2026-10-15T18:29:39Z

yea i need to write some tests for my toy vcs

just learnt the french created the word "vroomer" for ...

2026-10-15T15:26:33Z

just learnt the french created the word "vroomer" for people who like car centered infrastructure a little too much and i think the english language could benefit from this

minecraft really doesn't have a focus on making the anticheat ...

2026-10-15T10:49:23Z

In reply to nevent1q…2993
_________________________

minecraft really doesn't have a focus on making the anticheat good though, so it is only the basics

i wonder what isn't clear in "i don't play on those ...

2026-10-15T06:17:28Z

In reply to nevent1q…c0u9
_________________________

i wonder what isn't clear in "i don't play on those servers". i don't. i might dhow up once or twice but i do not spend significant time there

oh well i don't play survival on those

2026-10-15T06:14:08Z

In reply to nevent1q…t968
_________________________

oh well i don't play survival on those

it's a bot, no one is controlling it, and if there is, the ...

2026-10-15T06:11:18Z

In reply to nevent1q…8pv6
_________________________

it's a bot, no one is controlling it, and if there is, the mod already knows anyway

i can have a server mod create a fake player and have it attach ...

2026-10-15T06:08:24Z

In reply to nevent1q…jffd
_________________________

i can have a server mod create a fake player and have it attach my game to it, so i can control it

better than carpet bots that i teleport wherever they're useful

if i ever need more than one, i'm a developer, i can fix it ...

2026-10-15T06:02:06Z

In reply to nevent1q…eeec
_________________________

if i ever need more than one, i'm a developer, i can fix it myself

i never needed more than one, i got it during covid and one is ...

2026-10-15T06:01:15Z

In reply to nevent1q…kkgw
_________________________

i never needed more than one, i got it during covid and one is enough

i only have one account and honestly i hate how mojang does fomo ...

2026-10-15T05:55:38Z

In reply to nevent1q…ydu2
_________________________

i only have one account

and honestly i hate how mojang does fomo with capes. really unhealthy imo

dangerous to give me that much freedom

2026-10-08T19:08:17Z

In reply to nevent1q…g70e
_________________________

dangerous to give me that much freedom

okay then, next question, what qualifies as a game is a game ...

2026-10-08T19:04:41Z

In reply to nevent1q…huk4
_________________________

okay then, next question, what qualifies as a game

is a game within another game a different game or the same game ? where is the line of what a game is

what kind of epic are we talking

2026-10-08T18:54:29Z

In reply to nevent1q…g576
_________________________

what kind of epic are we talking

which one this time

2026-10-04T17:23:38Z

In reply to nevent1q…79c8
_________________________

which one this time

there's a bunch of blog posts about what defaults everyone ...

2026-10-01T19:28:47Z

In reply to nevent1q…xxy5
_________________________

there's a bunch of blog posts about what defaults everyone probably should override, i disagree with a few of them, but generally, diff.algorithm = histogram, commit.verbose = true, branch.sort = -committerdate are the ones i would recommend (diff.algorithm for better git with refactors, commit.verbose to show the full diff in the editor when you write the message, allowing you to doublecheck what is being committed with better navigation than less (or at least i hope your editor has better navigation than less), and branch.sort to show first those that have recently changed rather than alphanumeric sort

a lot will also recommend diff.colorMoved = true but i think it's confusing bc the colours are badly chosen (imo) and it doesn't highlight where something moved, only that it did move from somewhere or to somewhere

hey, you. yea, you. do yourself a favour, run git config --global ...

2026-10-01T18:53:17Z

hey, you. yea, you. do yourself a favour, run git config --global diff. algorithm histogram, and look at how much better life could be if git has reasonable defaults

a full year into uni a dude (who was into nfts because that was ...

2026-09-28T21:11:45Z

In reply to nevent1q…n0k3
_________________________

a full year into uni a dude (who was into nfts because that was the hype back then. remember them ?) asked me what a for loop is

the real tony was the friends we made along the way

2026-09-27T09:24:05Z

In reply to nevent1q…hwaf
_________________________

the real tony was the friends we made along the way

well technically savings accounts are made so they give less ...

2026-09-25T19:19:12Z

In reply to nevent1q…hw66
_________________________

well technically savings accounts are made so they give less return than investing / dividends, so it's not entirely wrong, very misleading though

there's a reason why i always recommend it

2026-09-24T18:13:19Z

In reply to nevent1q…3fgx
_________________________

there's a reason why i always recommend it

ant using xml is fine, if i were reinventing it i would probably ...

2026-09-24T18:04:26Z

In reply to nevent1q…557u
_________________________

ant using xml is fine, if i were reinventing it i would probably use scfg instead, but xml is fine enough. it makes more sense for ant than for maven imo

too late, and i like it

2026-09-24T17:58:34Z

In reply to nevent1q…er5k
_________________________

too late, and i like it

yea git only sees changes over whole lines of text, that's ...

2026-09-23T15:47:44Z

In reply to nevent1q…7kf7
_________________________

yea git only sees changes over whole lines of text, that's why git is terrible for everything that isn't code (there, the assumption that if the line has changed then the meaning is potentially completely different is actually reasonable, but it isn't for prose)

personally i am unfortunately not aware of a version control system that is actually competent for prose, it could exist though, and it's definitely possible if not

if im doing it how i think i may do it i'll need to write an ...

2026-09-23T08:21:17Z

In reply to nevent1q…38pe
_________________________

if im doing it how i think i may do it i'll need to write an ant task, fun

Nostr event nevent1qqsz0cxvuwrp22rxfh5gcegzje58q5y865yldkf7g883jmkpd87dv7czyrv8snsjxhjemzs04upqcdw6c0z7sr8dap9kc5qfjthzkurz42lwkp8hn7v

2026-09-23T08:15:52Z

i should write tests for git teaching thing

40th time's the charm

2026-09-22T20:02:11Z

In reply to nevent1q…5eve
_________________________

40th time's the charm

in tandem, sure, or in parallel, but not stuck together in a ...

2026-09-22T07:30:53Z

In reply to nevent1q…lq5x
_________________________

in tandem, sure, or in parallel, but not stuck together in a single history, esp since that would probably be the code vcs, which sucks for anything other than code. git sucks for anything that isn't code, because it wasn't made to handle anything that isn't code. translations, famously, aren't (usually) code, and therefore we shouldn't force people to use a tool that sucks for what they do

hence, a submodule like thing but across VCSs

code history and translation history and assets history aren't wholly separate, true, but they shouldn't be stuck together because, no, one tool will never work for every job, and we need to accept that and work into making things that do work

and why ? you're looking at translations, why would you use a ...

2026-09-21T20:52:33Z

In reply to nevent1q…3u2u
_________________________

and why ? you're looking at translations, why would you use a tool meant for handling code and not translations ? what about textures ? will git be useful when looking at PNGs ? no! you need something custom for whatever you're handling

yes! what i want is what submodules are like but for arbitrary ...

2026-09-21T20:50:50Z

In reply to nevent1q…84qd
_________________________

yes!

what i want is what submodules are like but for arbitrary VCSs (with permissions, of course) and actually working and not the half dead horse that is git submodules

yes, it's in the translations log. you can do that with ...

2026-09-21T20:49:17Z

In reply to nevent1q…x0nk
_________________________

yes, it's in the translations log.

you can do that with multiple repos. in fact, it's even easier, as translation history is only in translations repo!

just a store isn't enough, those get worked on, so there ...

2026-09-21T20:47:16Z

In reply to nevent1q…e8ac
_________________________

just a store isn't enough, those get worked on, so there shoukd be a full featured VCS

nope, the translations never hit the code repo, the two are ...

2026-09-21T20:45:49Z

In reply to nevent1q…3ll0
_________________________

nope, the translations never hit the code repo, the two are completely separate. a typo being fixed has no consequence on the code, and a fix in the code has no consequence on the translation data, they have no reason to be kept together in a weird mangled history that will just be annoying gor both parties

git's model is completely tied down to files of text with ...

2026-09-21T20:43:34Z

In reply to nevent1q…wlge
_________________________

git's model is completely tied down to files of text with short lines. a very reasonable assumption for code, but terrible for everything else. to me, a meta VCS that allows common parts and maybe database formats to be shared would be reasonable, but a lot of things would need to be reworked, and each component (not specifically code, more assets and data and such) should be in individual repositories

for example, translations make no sense to have alongside code, ...

2026-09-21T20:40:45Z

In reply to nevent1q…uxya
_________________________

for example, translations make no sense to have alongside code, like, what, when you update translations you're gonna put a commit like "bump translations" ? how unserious

stack arrays ? vlas ? malloced arrays ?

2026-09-21T20:39:12Z

In reply to nevent1q…xx4g
_________________________

stack arrays ? vlas ? malloced arrays ?

to me, having a general purpose VCS that can handle everything ...

2026-09-21T20:38:27Z

In reply to nevent1q…ds6c
_________________________

to me, having a general purpose VCS that can handle everything with general purpose diff/merge/etc algorithms that work with binary data is completely off the mark, the goal should be having things specific to what you're working with so that you can work with them well

VCSs shouldmake working with other repos (and also other VCSs) as ...

2026-09-21T20:35:46Z

In reply to nevent1q…m77l
_________________________

VCSs shouldmake working with other repos (and also other VCSs) as easy as possible so that people aren't stuck with using git for everything, and git submodules is, as tooling, wholly subpar and incapable of any serious work

like how imo having models, assets, translations, etc in the same ...

2026-09-21T20:32:17Z

In reply to nevent1q…2azu
_________________________

like how imo having models, assets, translations, etc in the same repo as code is messed up, and those aren't code so we should be able to have better VCSs specifically for them, and our VCSs are completely unable to work with eachother as it is, which imo is a failure

that kind of thing

many things about VCSs

2026-09-21T20:27:16Z

In reply to nevent1q…yqgc
_________________________

many things about VCSs

i have so many things to write about but i can never seem to get ...

2026-09-21T19:36:10Z

i have so many things to write about but i can never seem to get in the right mind space to actually do it

i fucking hate having adhd

i change pfp too often for that

2026-09-16T05:50:25Z

In reply to nevent1q…t95p
_________________________

i change pfp too often for that

i don't have one

2026-09-16T05:46:33Z

In reply to nevent1q…7v93
_________________________

i don't have one

this is so gamecore pack

2026-09-15T06:46:44Z

In reply to nevent1q…kymu
_________________________

this is so gamecore pack

Nostr event nevent1qqsxahccstuk5zlvr44ruzhpdrmkdcu3zefcpnwuk9te5ng5cg25c9qzyrv8snsjxhjemzs04upqcdw6c0z7sr8dap9kc5qfjthzkurz42lwk0kc3ud

2026-09-14T08:41:18Z

i must wag but i have no tail

"i prefer not to pollute root project directory with one test ...

2026-09-13T21:29:29Z

In reply to nevent1q…vsfh
_________________________

"i prefer not to pollute root project directory with one test config directory" people when CONTRIBUTING, LICENSE, MAINTAINERS, README, BUILDING, [insert package manager].toml, [insert package manager].lock ...

kid called echo "run/" > .git/info/exclude

2026-09-13T21:24:00Z

In reply to nevent1q…0794
_________________________

kid called echo "run/" > .git/info/exclude

and no error either ? odd

2026-09-13T18:18:22Z

In reply to nevent1q…phed
_________________________

and no error either ? odd

maybe me saying java and js are fine, actually, and that how ...

2026-09-13T18:14:47Z

In reply to nevent1q…uz4w
_________________________

maybe me saying java and js are fine, actually, and that how badly they get abused is a problem, is too radical

lol they blocked me

2026-09-13T18:09:50Z

In reply to nevent1q…us5n
_________________________

lol they blocked me

@npub19v8…gsx7 i think @npub1e9p…zeum has no fucking clue ...

2026-09-13T06:39:21Z

In reply to nevent1q…y676
_________________________

FoolishOwl (npub19v8…gsx7) i think Faraiwe (npub1e9p…zeum) has no fucking clue what they're talking about and are just doing virtue signalling based on what is unpopular among mainstream developers

guess what ? js is fine enough, browsers and the web are the problem, and java is fine, enterprise java code is the problem

also the jvm is good

2026-09-13T06:34:30Z

In reply to nevent1q…l92r
_________________________

also the jvm is good

you're in school, unless you're specifically in a ...

2026-09-11T14:30:54Z

In reply to nevent1q…wxcz
_________________________

you're in school, unless you're specifically in a performance oriented class, they won't get into specifics

if you access it randomly that's a fair assessment, and in ...

2026-09-11T14:28:22Z

In reply to nevent1q…wvmn
_________________________

if you access it randomly that's a fair assessment, and in certain data structures it also is a fair assessment

6 morbillion lines long --help

2026-09-08T20:03:19Z

In reply to nevent1q…6a6p
_________________________

6 morbillion lines long --help

"daddy (lesbian)"

2026-09-06T18:39:22Z

In reply to nevent1q…3jzv
_________________________

"daddy (lesbian)"

even if it were better tgan me idc i like doing it

2026-08-31T20:07:22Z

In reply to nevent1q…9h8l
_________________________

even if it were better tgan me idc i like doing it

"haha if you use discord in light mode you're a ...

2026-08-31T12:12:45Z

"haha if you use discord in light mode you're a heretic" how about go fuck yourself and let people use their computer however they like / works best for them ?

believe me, if you've seen enterprise informix4gl, you'll ...

2026-08-23T07:25:03Z

In reply to nevent1q…qvut
_________________________

believe me, if you've seen enterprise informix4gl, you'll find enterprise java sexy

"dominant puppygirl" close enough

2026-08-21T18:38:22Z

In reply to nevent1q…dxtq
_________________________

"dominant puppygirl" close enough

the ux on forgejo is already miles ahead on many of the fringe ...

2026-08-21T06:43:04Z

In reply to nevent1q…lv06
_________________________

the ux on forgejo is already miles ahead on many of the fringe parts, which are great for new maintainers

like, just compare the force push on prs on the two: github will show you the entire diff of the pr, forgejo gives you what commits have been pushed and allows you to see each individual diffs ... i was going to say look on codeberg but now clicking on the "compare" button seems to return a 404

from my point of view, forgejo is basically github, but foss, and ...

2026-08-21T06:27:43Z

In reply to nevent1q…7jcv
_________________________

from my point of view, forgejo is basically github, but foss, and somewhat improved

i hope for their versions of "stacked prs" they make an actual tree / dag dependancy system rather than the extremely narrow (heh) github style linear only

bonjour

2026-08-21T06:21:33Z

In reply to nevent1q…vmtq
_________________________

bonjour

my guess would be yes, but i am not sure

2026-08-21T06:21:16Z

In reply to nevent1q…864u
_________________________

my guess would be yes, but i am not sure

github is like freud, it started a bunch of interesting ...

2026-08-20T19:39:19Z

In reply to nevent1q…ahfp
_________________________

github is like freud, it started a bunch of interesting discussions and research, but we shouldn't rely on it because everything is wrong, misleading, outright false, and powered by child labour

yeaaaaaa but 1 that's a fucking hack 2 it should be built in ...

2026-08-20T19:36:33Z

In reply to nevent1q…cmhu
_________________________

yeaaaaaa but 1 that's a fucking hack 2 it should be built in directly

i can guarantee you many new maintainers have seen code that's fine, saw the pretty green button, "no problem" the megacorp assures them, they click, and there is in fact a probkem bc now the whole project doesn't work, and the new maintainer has on their repo the Stain of a non working commit (shown by ci)

but, well, ig it would only be one if many among the "fix ci (for real this time please please please)" commits bc github has decided that local first ci is for losers

yea i want on the repo after a merge but before that merge is ...

2026-08-20T19:31:22Z

In reply to nevent1q…n052
_________________________

yea i want on the repo after a merge but before that merge is actually added to the target branch

and squashing only collapses the history of the pr into a single ...

2026-08-20T19:23:06Z

In reply to nevent1q…szlr
_________________________

and squashing only collapses the history of the pr into a single one, the trees being merged are the same as normal merge, so if normal merge succesfully merges but results in a broken project, squash would be the same

and no it defaults to normal merge anyway

it runs on the merge commit after merge bc then it's head of ...

2026-08-20T19:20:31Z

In reply to nevent1q…x7jx
_________________________

it runs on the merge commit after merge bc then it's head of repo

i want before, when still in pr screen

also i hate hate hate how github will show you a big bright green ...

2026-08-20T19:11:05Z

In reply to nevent1q…pr5w
_________________________

also i hate hate hate how github will show you a big bright green button to merge if it succesfully merges and passes ci, but doesn't run ci on the merge itself, which is bad bc a succesful merge doesn't guarantee a working commit

yea like, for example, in git, rebase can be seen as being built ...

2026-08-20T19:08:49Z

In reply to nevent1q…nyrc
_________________________

yea like, for example, in git, rebase can be seen as being built on top of cherry pick, which itself can be seen as built on top of merges

i probably commit code that doesn't work a handful of times a ...

2026-08-20T19:07:06Z

In reply to nevent1q…h356
_________________________

i probably commit code that doesn't work a handful of times a year, i am extremely anal about keeping the history clean, code can be fixed later but history can't (well, it can, but if someone else uses the repo it's as if you can't)

the way i work is i go through the implementation of a feature, committing as i go, when the code is in a state i am content with, i erase that temporary history, redo it only based on what is in that state (that removes code being added and deleted which is mostly useless in the context of the project's history as a whole) and then push or pr that

and i mean cases where a feature is only partially added in a way such that it doesn't work, or has 5 prs adding random things to a just added feature even though all necessary parts were known from the beginning

like if it's "oh i need a refactor of this part" then ...

2026-08-20T18:47:48Z

In reply to nevent1q…u3wd
_________________________

like if it's "oh i need a refactor of this part" then sure, but pushing in half a feature is a No

idk maybe it is pr dependencies but it seems weirdly obtuse and ...

2026-08-20T18:43:53Z

In reply to nevent1q…9m6h
_________________________

idk maybe it is pr dependencies but it seems weirdly obtuse and arbitrarily restricted, pr dependencies aren't linear, but everything i read tells my brain that it's purely linear, which is just not correct

also i think splitting in pr 1 pr 2 and or 3 leaving partial work in the repo is a profound mistake, main branch should only ever have finished work, what's the point of wip branches if main is itself a main branch

it may just be me being tired but 1 i don't see how this is ...

2026-08-20T18:37:09Z

In reply to nevent1q…qft8
_________________________

it may just be me being tired but 1 i don't see how this is better than pull request dependencies (as in where some pr depends on another) and 2 is it me or does their docs not have a way to show how to make them without the cli

yes, even if it involves a secret, there should be a way to shim ...

2026-08-18T17:21:06Z

In reply to nevent1q…8lh0
_________________________

yes, even if it involves a secret, there should be a way to shim it locally and test it, and if not, the ci system is a failure

once again the "CI must be local first and easy to run" ...

2026-08-18T17:17:19Z

In reply to nevent1q…9trg
_________________________

once again the "CI must be local first and easy to run" argument that forge developers don't want to hear

that thing on github is honestly completely worthless, and this ...

2026-08-18T14:48:47Z

In reply to nevent1q…wdc9
_________________________

that thing on github is honestly completely worthless, and this is one of the reasons why

what's worse is git allowing you to create commits at arbitrary times have a real use: it's how repositories imported from other VCSs keep the times right

i mean i still play 1.7.2

2026-08-12T17:10:44Z

In reply to nevent1q…0ltm
_________________________

i mean i still play 1.7.2

as a tech player im used to it tbh

2026-08-12T17:01:50Z

In reply to nevent1q…ewy9
_________________________

as a tech player im used to it tbh

the website lags on firefox on my phone, which generally ...

2026-08-11T09:26:35Z

In reply to nevent1q…d9an
_________________________

the website lags on firefox on my phone, which generally doesn't inspire me much confidence in the efficiency of the software

"security is more important than accessibility" look ...

2026-08-02T09:49:22Z

"security is more important than accessibility" look buddy, if i want to stay safe on my computer while not being able to use it, i may as well not power it on

eh ?

2026-07-26T06:10:34Z

In reply to nevent1q…chjv
_________________________

eh ?

yesterday i was amazed by past me's choices for two apis, and ...

2026-07-24T05:49:46Z

In reply to nevent1q…0kks
_________________________

yesterday i was amazed by past me's choices for two apis, and how easy deciding well beforehand what things will be like makes everything afterwards easier

name a single daemon that doesn't (cron and co work, their ...

2026-07-23T08:08:20Z

In reply to nevent1q…t4hv
_________________________

name a single daemon that doesn't (cron and co work, their input is a config and the output is starting stuff

yea right putting politics aside will totally lead to a better ...

2026-07-13T11:12:46Z

In reply to nevent1q…54eq
_________________________

yea right putting politics aside will totally lead to a better ecosystem and not just one slapped together faster

welp have fun

2026-07-09T09:31:25Z

In reply to nevent1q…rzum
_________________________

welp have fun

is there really no other way this could have been done ? sounds ...

2026-07-09T09:08:40Z

In reply to nevent1q…9hze
_________________________

is there really no other way this could have been done ? sounds suspicious tbh

like, maybe a list of async / sync events and a kind that's the event's name ? wouldn't that have been fine ? though it does depend on how the library is structured

well sure but i'm sure you could simplify a lot of that

2026-07-09T09:05:02Z

In reply to nevent1q…hxxq
_________________________

well sure but i'm sure you could simplify a lot of that

damn, not even aliasing that ?

2026-07-09T05:13:05Z

In reply to nevent1q…sqz7
_________________________

damn, not even aliasing that ?

aight, gonna add you when i have time, won't have much today ...

2026-06-29T06:39:50Z

In reply to nevent1q…jjex
_________________________

aight, gonna add you when i have time, won't have much today though

i have discord and signal

2026-06-29T06:12:25Z

In reply to nevent1q…lesm
_________________________

i have discord and signal

it says right on the doc it supports ninja and makefile ...

2026-06-28T08:45:51Z

In reply to nevent1q…66v6
_________________________

it says right on the doc it supports ninja and makefile generators

also maybe ...

2026-06-28T08:38:18Z

In reply to nevent1q…jcw3
_________________________

also maybe https://cmake.org/cmake/help/latest/variable/CMAKE_EXPORT_COMPILE_COMMANDS.html ?

worse than cmake ? impossible

2026-06-28T08:36:57Z

In reply to nevent1q…4vyw
_________________________

worse than cmake ? impossible