@lanodan@queer.hacktivis.me e.g. the verbatim search "OnPnpNewDevnode" gives me almost nothing to go on in kagi, but gives me juuuust slightly more in google, and that tiny bit matters
@lanodan@queer.hacktivis.me kagi is usually pretty good about finding things in archive.org if that's the best place to look for them, especially if i've set it to look for pdfs (and because i've raised its priority, which is a thing kagi lets you do)
but it still misses out a lot
and there's stuff archive.org doesn't have (or doesn't have indexed) which google finds in google books / scholar, and kagi does not
trying to [write / sorta-reverse-engineer / understand] windows 95 drivers in 2023 is actually really difficult because there is almost no information remaining on the internet
on any given topic i can only find just enough information from old forum posts or mailing lists to assure me that, once, long ago, there were answers, and tools, and ways to learn.
my saving grace is that i have managed to get a pdf copy of what i believe is the exact "how to write a driver" book that the authors of the driver i am trying to re-write used
and i believe their driver began as a carbon copy of all of the example code from that book, shoved together into a single project
most of the rest of the functionality looks like stuff that started out in the userland side of the stack (which i have source for), and was eventually #ifdef'd out and moved into the kernel driver (conveniently with a comment to that effect in some instances)
now, the trick is, i'm not actually rewriting the windows 95 driver at the moment. i was thinking about doing that, but even with the stuff being released under MIT license, i decided if i'm going to be writing a kernel driver based on something partly reverse-engineered from a binary, i'm gonna have a different target for the driver to quash potential accusations of impropriety.
so i'm targeting NT4 instead, since it still uses tools from the same era, will run the exact same unmodified .dlls and binaries, but has a dramatically different driver model
... i was also getting fed up with win9x crashing on me constantly, and NT4 is proving very stable after updates, with only one BSOD that was due to the hard drive dying underneath it
it's really funny that i've been able to run a usable matrix server (conduit) on a literal 486, but i can't run a usable matrix client on like. basically anything but my phone and the two non-POWER desktop computers i have with a manufacturing date greater than 2010 (and my 2008 laptop if i am desperate)
i want a matrix "client" that just lets me connect, like, AIM, to it
i dont want a matrix "bridge" that tries to synchronize the AIM world and the matrix world, in order to let me use AIM to speak to both AIM or matrix users (or the other way around).
i want to connect my 20 year old AIM client to my matrix server and use it as a matrix client
my use case is not "connect matrix, irc, and AIM users all in the same room"
my use case is i want to send my wife a screenshot of a weird thing i found on my windows nt 4.0 computer, and we both use matrix, when it's not possible to use a matrix client on a computer that is more than 15 years old (or which is newer than that but slower than a 15 year old pc)
hey! hey you! wanna run a native matrix client on a brand new POWER9 workstation? (or at least a client that pretends to be native even though it's a web browser?)
well, tough luck. all you get is iamb. nheko might work if you let it sit for about a week after you log in, but good luck with that
and if you think you can get away with using a browser client, i hope you like google chrome!
do take note that this is not a call for harassment. it is wrong to read it as such.
the right to spread the knowledge of the harm caused by someone (especially a public figure), so that people can take action in order to protect themselves and others around them, is not harassment. many people who would perpetrate harm would prefer to frame it as such, but it is not.
i do not think people should respond to RMS's harms by harming him in return. he should not be ostracized. he should not receive hate mail. people who are comfortable doing so should continue to have respectful discussions with him.
but i do think that it is appropriate to take steps to limit the harm he can cause - such as ensuring he does not remain as figurehead in the free software movement, wielding an amount of power in his words that i do not think he comprehends that he has.
i received a DM earlier. i am not going to call out the person who DM'd me, but i am going to publicly share my response, because it is important. the context is @drewdevault@fosstodon.org's blog post earlier today about RMS. the person DMing me suggests for the "proper" place for this discussion was in the FSF channels and with RMS himself.
my response is this, verbatim:
"no. if we cannot discuss issues like this in places not directly controlled by the people at the root of the issue - e.g. in this instance, RMS and people sympathetic to him - then the reality is that we cannot discuss them at all.
we cannot restrict the discussion of difficult topics to forums controlled by people who might rather not have those topics talked about. there is a power imbalance, whether or not that power is used.
is free speech still free speech if, whether by rule or by cultural norm, you are only to criticize a king from within his castle, surrounded by his guards, with only the audience of himself and those who have chosen to be in that environment? even if the guards take no action against you, and allow you to freely speak your mind, there is no question in that situation that you do not hold the power, and you are at the mercy of those who do.
and, ever so frequently, they do exercise the power to silence or eject you, and if we do as you suggest, you are left with no acceptable avenue in which to exercise your right to speak.
you must be able to criticize the king in your home, or in the tavern, or on the streets, or wherever you choose, in places where his guards have no power. you must be able to do this with other people who would not have chosen to subject themselves to the king's court. you must be able to loudly proclaim his faults, so that those who would never have had reason to find themselves in his castle may hear, and understand.
the king will not dethrone himself because someone entered his court and made a good point. he will dethrone himself because crowds of people outside his court are making their opinions known, and he knows that if he does not, the castle he worked so hard to build will fall with him inside."
people like stallman being figureheads, and people like you deflecting discussions like this away from people like him, are a substantial part of why i am uncomfortable participating in communities around free software
believe me when i say, drew's article is serving to amplify marginalized voices.
you, a privileged white dude, are serving to suppress that discussion, and by extension, those voices. that is not the way to move the issue forward.
i am a bit confused, because an optimization i am pursuing in 486quake is making things slower in dosbox
there is a special case under which a different version of a drawing function can be used. i have a toggle that lets me switch, at runtime, whether it calls the normal version or the special-case version of the function when the special case is detected
the only difference between the regular version and the special-case version is that the special-case version *doesn't* have to do some math that the regular version does. and so that math is removed
and yet for some reason the special-case version takes longer
some girl who writes code sometimes-i live in the midwest and do tech things-i like old computers and weird operating systems and writing emulators-i have many girlfriends-sometimes i make music as 'ersatz waterfall'-consumerism is destructive-if you send me a follow request and you already follow me on twitter, or recently migrated instances, dm me so i know who you are