Replying to
@atax1a@infosec.exchange Writing software in this language is referred to as "taking it out for a spin."
Remote
Simon Richter
@GyrosGeier@hachyderm.io
mastodon 4.7.3Geek.
9C43253495E4DCA837945F5BEBF67A846AABE354
408 Followers
177 Following
50 Posts
Joined December 13, 2022
pronouns:
he/him
github:
Replying to
@mwl@io.mwl.io the current trend to disregard the "free software" aspects and focus on the "open source" ones is nonetheless worrying, especially the bits where people effectively emulate commercial software publishing, with its clear distinction between suppliers and consumers (and "power users" who are locked into an ecosystem).
Replying to
@threedaymonk@sonomu.club three toots further on my timeline: https://mastodon.social/@Thayer/116612369342129117
Not an exact fit, though.
Replying to
If people are unwilling to parent their children, that is not something we can solve with technology.
We can, however, by acceding to these demands, maneuver ourselves into a situation where we have invested a lot of time into this, indicating that it is a viable idea, and then have to admit failure because it is simply impossible to solve with technology that leaves even a modicum of control in the users' hands.
(2/2)
Replying to
@cstross Warhammer 40k already exists though, and there's a lot of chuds who think they'd be the elite soldiers, same as they think of themselves as temporarily embarrassed billionaires right now.
What we need more of is a hopeful vision of the future, one where wasting dwindling resources isn't a status symbol.
Large parts of my timeline are now debating whether resisting the introduction of the "birthdate" column in the user database is part of a fascist takeover and should be resisted.
It is a distraction from the fascist takeover. The walls will come up somewhere else.
If it's not maintained by someone trusted by the government (i.e. an entrenched corporation), it *is* just a database column.
Debating it wastes a lot of energy we would need elsewhere: there is a lot of pending legislation that everyone now treats as if it had already passed, and so the only thing left to do is take a heroic last stand, to at least not be complicit.
We're not at this point yet. There are still things that can be done, but that means organizing against big tech, which means both political opposition, and building out alternate infrastructure if you're too exposed for direct political action.
Even debating whether the people involved are willing collaborators or just politically ignorant is a waste of resources at this point. Their motivations do not matter, they aren't allies either way.
These people aren't in a position to get the database column they created recognized as "trustworthy." Their act was one of symbolic bootlicking, so publicly opposing it likewise only has symbolic meaning.
The actual attestation framework that we will get is most likely some "login with Microsoft/Google/Apple/Canonical/Facebook¹" thing that integrates into the login process and requires no trustworthy components on the user's machine because all it does is forward a token received from one online service to another, and not presenting such a token locks you out from a lot of services, because the services themselves are at risk if they do not require such a token.
*That* is where this is headed, and *that* is what we need to oppose.
Of course open source will be left with local-only accounts. These just won't be able to connect to online fora.
¹ involvement of Facebook still unclear — no good reason for Microsoft and Google to let them have a seat at the table, really. Canonical will be there in the beginning to make it a bit less off-putting to Linux users, but at least their identity management will be merged into Google at some point.
Replying to
@jkruse_de@norden.social @SheDrivesMobility@norden.social da gab es schon mal eine Debatte drüber, als das neu war (also erst mit Ingress, dann mit Pokémon go), aber die Nerds wollten.halt unpolitisch bleiben.
Doing a small prerelease of my tarball validation tool:
https://psi5.com/~geier/tmp/git-treeof-0.0.20260411.tar.gz
The basic idea is that you can feed it a tar file, and it tells you What Would Git Do™, i.e. it is equivalent to
git init
tar -x
git add -A
git commit -m 'irrelevant'
git rev-parse HEAD^{tree}"
but it does not require a local filesystem, git installation or anything -- it's an isolated program that allocates 64 MB of RAM (that should be enough for everyone) and then activates strict SECCOMP mode before looking at any input data.
The main thing I'm unhappy with is the name, in principle it would be easy to extend this to a tar diff tool, and I also want to add some features like "if you supply the commit object via a separate channel, that can be verified too", so it's not strictly "give me the tree ID for this tarball" anymore.
Also, if you have use cases where it doesn't work, I'd like to collect these as well. I already know about submodules, and I also want to have a mode where I also parse gitignore files.
Replying to
@drs1969@mstdn.social @BaronWalschap@mastodon.nl @AkaSci@fosstodon.org @sundogplanets@mastodon.social we put large sails on our satellites to be able to deorbit them quickly, there is still enough atmospheric drag. The ones we use are smaller than what RO proposes, and attitude control for deorbiting is a lot easier than for pointing.
This isn't quite at "datacenters in space" levels of foreseeably impossible, but it's close. I suspect the target audience is the same.
Replying to
@tomjennings Raptor Computing are going as far as possible with their POWER based boards, in general the IBM CPUs are very well documented.
Modern x86 follows a similar model as POWER: the CPU core does not have much startup state beyond the initial instruction pointer and privilege level, but it isn't started until after the entire physical memory map is set up from an extra microcontroller.
The difference is that x86 integrates *that* microcontroller onto the same die and doesn't document.
We're restarting #HillHacks this year!
Since we don't have the momentum of a yearly event yet (but want to get there again) this might be a bit smaller than previous editions, but this is still an excellent opportunity to meet people in Northwestern India.
This is an "unconference", where the talk schedule is decided mostly by the participants who are there. If you have an interesting project and would like to present it, a digital rights topic you want to talk about, or are interested in what other people are doing, you are more than welcome to join us.
Other coordination before the even happens mostly on the mailing list, which can be found at https://hillhacks.in/mailinglist/ . There are plans for a school program, for an open source hackathon and for some #OpenStreetMap mapping, and I expect that more ideas will come up in the coming weeks as well.
The main event website is at https://hillhacks.in/
Warm hier, aber waren Sie schon mal in meinem Serverraum?
Running my tar parser on all the tar files in the Debian archive, to find cases I haven't handled yet. It's not done yet, but so far I've collected
2 num_read == sizeof s->buffer
3 symlink_size < LEN_LINK_TARGET
3243 unexpected character in octal
780 unhandled file type
17 unused tar header has data
8 value_len < sizeof s->filename
Basically, I've plastered a bunch of checks all over the code that are like assert(), but remain in Release code. All of these are hard errors so far, and I will need to add more checks and also make the parser less strict in places where it is overly so.
"num_read == sizeof s->buffer" means that the tar file is not a multiple of 512 bytes in size.
"symlink_size < LEN_LINK_TARGET" means that the link target field is used completely and not NUL terminated. That's probably legal, but I'm amazed someone is running into this. Also, I suspect there is an encoding for even longer symlinks.
"unexpected character in octal" seems to be mostly octal numbers that are right-justified in the field with spaces, which should be legal.
"unhandled file type" refers to the file type marker inside the archive -- I handle directories, files, symlinks and extension headers for the whole archive and for individual files. So that's a TODO item.
"unused tar header has data" happens when the first byte of the filename is NUL, but there are non-NUL bytes in that block. Need to investigate what these are.
"value_len < sizeof s->filename" means that there is a long file name with more than 255 bytes.
Getting the full statistics will take a while, and then I need to fix the program and re-run everything.
Replying to
@zzt in this case, it is really just a field in a database, because the bootlickers have not understood that they're not part of the club that gets to implement age verification.
Why should a website trust an age indication from a piece of software that might have been installed by a minor?
It was useful to see who the bootlickers are, and we should focus on getting them to understand which side of the walls they will be on when they come up.
Replying to
@regehr I think this should be a kernel feature.
There is a scheduling class, SCHED_BATCH, with low priority and very long (1.5s) slices. I think this could be improved even further, by essentially making the slices infinite, yielding CPU in execve() at the point where the process image is smallest, and on memory pressure, swap out the other batch class processes, smallest first. When there are free CPUs, resume the largest batch process.
Users would just make -j with no limit.
Replying to
Replying to
@Thayer@mastodon.social @threedaymonk@sonomu.club not sure about UK regulation — I understand they already have a work visa, they just may not be unemployed for more than three months under it.
Replying to
@AndiBarth@muenchen.social there is a book now, "Run Your Own Mail Server", that has checklists.
Other than that, I started my postfix config the same way I start other config files, by importing the manual page and then editing in place.
Replying to
@dngrs@chaos.social @CyReVolt@mastodon.social switched to ibus for now, which took a while to set up properly, and displays the preview and property windows on the wrong screen, but it will do for now, I should get back to work.
I should add a public graph "Inbox size" to my personal homepage.
Spending the rest of my "lunch" break on improving society somewhat.
Replying to
@ralphstark @kajakster @gerolf_antritt @detektorfm @cbollert ja, das ist eher was für den Stadtverkehr, wo man oben sitzen will, um über Autos drüberzuschauen.
Replying to
@benjamineskola@hachyderm.io @srtcd424@mas.to @zzt@mas.to my impression (from reading the mailing list) is that they are mostly reacting to authoritarians infringing on their freedom, but not extrapolating that to resisting any form of authoritarianism.
Replying to
@zzt@mas.to wait how is Devuan fascist-controlled?
(my feeling is more "so understaffed that they have to follow systemd's lead in more aspects than they'd like to")
Replying to
Replying to
@gsuberland @0xabad1dea The "letter to Baghdadi" was also an entertaining read.
https://en.wikipedia.org/wiki/Letter_to_Baghdadi
Replying to
@the_etrain@beige.party @quinn@social.circl.lu it should be written using a futuristic font, and flickering.
Time to order the good biryani and kebap for the fourth time in a week, I guess.
Replying to
@jenny753@indiepocalypse.social @alice@lgbtqia.space does this option insta-reject, or does it create a tarpit?
I should put the TalosII into a colo facility.
100Gbps is probably too expensive, but it would be cool to see if it can keep up.
Replying to
@azonenberg@ioc.exchange when I was working at R&S, the project leader said "ah, your feature is close to complete, I guess, you have been writing bug reports against other components."
Oh cool, updating the work laptop broke SCIM.
What does one use these days for
- Japanese text input
- ideally, Anthy tables
- needs to work with QT, GTK and XIM
?
Replying to
@mshelton@mastodon.social @quinn@social.circl.lu the cynical take is that it behaves like bodycams that cannot be turned off, and also tracks them, and that's what they are unhappy about.
Replying to
@azonenberg@ioc.exchange I kind of want to build a 16 lane OpenCAPI contraption out of a Ti375.
Replying to
@masek@infosec.exchange @floe@hci.social I doubt there is a viable business model in there, even if you manage to stop people from deploying local models, and that can be done cheaper by no longer publishing weights.
Replying to
@NanoRaptor@bitbang.social always a good feeling when they tell you to drink lots of water and stay away from small kids and pregnant people.
Replying to
@skington@glasgow.social @cstross@wandering.shop also, individual action can never correct a systemic problem, only reduce the pressure to fix it.
Replying to
@azonenberg@ioc.exchange btw, current packaging state is on my server at https://deb.simonrichter.eu/ inside the "misc" repositories. I need to redo the copyright/license summary file again because it got rejected for that last time.
Longer term, I'd like to split out the documentation and all the architecture-independent data (images, shaders, documentation) into an architecture-independent "common" package, but it is difficult to build only those files and skip building the main program.
Replying to
@david_chisnall@infosec.exchange I can even fine tune the plausibility by hydrating more or less.
[computing.png]
The rain is so loud I can't hear the CPU fans anymore.
Replying to
@Elwood_G@muenchen.social @AndiBarth@muenchen.social same, but I got "Run Your Own Mail Server" to get a checklist for things I need to do.
Replying to
@sundogplanets@mastodon.social aerobraking with an 18x18m sail still works quite well in LEO.