|
feedmegin posted:I realise this is still technically 'intel' but Itanium workstations running Windows were a thing more recently than that.  https://www.ebay.com/itm/133025780816 at least it can also run VMS rather well
|
#
?
Jan 26, 2021 22:20
|
|
|
|
| # ? Apr 24, 2024 21:00 |
|
|
eschaton posted:
the same seller prices much more ancient hppa machines significantly higher, even in obsolescence itanium is a fizzle
|
|
#
?
Jan 27, 2021 02:29
|
|
|
well of course, pa-risc had a cool rap. itanium is just lame
|
|
#
?
Jan 27, 2021 02:49
|
|
|
yeah, I got an rx2620 without the pedestal case for a couple hundred and I could spend another couple hundred at a surplus dealer I know and get the pedestal it runs the latest OpenVMS beautifully but it’s loud enough that it overpowers my noise suppressing headphones I ported the 3dfmaps “benchmark” and compiled it with as much optimization as I could muster and the best scores/times I was able to get were around 0.43
|
|
#
?
Jan 27, 2021 04:20
|
|
|
my general memory of the era is that itanium workstations were a giant wet fart, literally only useful for software devs stuck with the deadend job of developing code for hp nonstop or whatever that big (ish?) iron product line was, you know the one place itanium ever did halway well (and that mostly due to captive hp customers)
|
|
#
?
Jan 27, 2021 04:31
|
|
|
Ganoo Cyst 'em D. open sores cocks, get it? eh eh my love/hate relationship with linux is eternal
|
|
#
?
Jan 28, 2021 06:25
|
|
|
the future of init is GNU shepherd praise gnu
|
|
#
?
Jan 28, 2021 10:25
|
|
|
hell yeah, my init system definitely needs a 50 MB scripting language that depends on a bunch of external libraries praise be to saint iGNUcius for showing his holiest of holies and vanquishing the evil TCL language Poopernickel fucked around with this message at 16:41 on Jan 28, 2021 |
|
#
?
Jan 28, 2021 16:38
|
|
|
to summarize, my position is: - runit bad - shepherd bad - sysvinit bad - systemd-init good, systemd bad overall 
|
|
#
?
Jan 28, 2021 16:43
|
|
|
I just looked at the github page for sudo and apparently this thing has like 400k lines of code? For something people basically use as a "pretend to be root for a few seconds" command?
|
|
#
?
Jan 28, 2021 16:57
|
|
|
Nitrousoxide posted:I just looked at the github page for sudo and apparently this thing has like 400k lines of code? For something people basically use as a "pretend to be root for a few seconds" command? all that commented BS everyone ignores in /etc/sudoers takes a lot of code to support. also unix makes it a bitch and a half to write sensitive code that's actually secure
|
|
#
?
Jan 28, 2021 17:15
|
|
|
Poopernickel posted:hell yeah, my init system definitely needs a 50 MB scripting language that depends on a bunch of external libraries just put scheme in the kernel already you cowards, guile as pid 1 is just a stopgap.
|
|
#
?
Jan 28, 2021 17:33
|
|
|
Nomnom Cookie posted:all that commented BS everyone ignores in /etc/sudoers takes a lot of code to support. also unix makes it a bitch and a half to write sensitive code that's actually secure Hey, I like /etc/sudoers. I learned to love it at the company that used visudo tied to a script to update sudoers manually for a company with ~10k unix accounts. Frequently to grant access to 2-3 machines out of a global farm of ~20k. That poo poo is awesome if you manage it right. Or just touch it once to make a sudoers group across the whole farm and manage access by group membership.
|
|
#
?
Jan 28, 2021 17:50
|
|
|
sudo is inefficent - I just login as root and do everything from there
|
|
#
?
Jan 28, 2021 18:08
|
|
|
/etc/sudoers is also helpful in the world of embedded systems, where you maybe want one or two narrowly-defined commands that run as sudo you can use it to express something like: - user "foo" can run "/usr/local/bin/firmware-updater /tmp/update-file.tar.aes" as passwordless sudo - user "foo" can't run firmware-updater from any other path, or pointing to any other update file - the "PATH" environment variable is locked to "/bin:/sbin:/usr/bin:/usr/sbin" by sudo - the "UPDATE_MODE" environment variable is passed through, but all other variables are ignored/replaced then on engineering builds/debug builds, maybe you swap out /etc/sudoers for something with fewer restrictions
|
|
#
?
Jan 28, 2021 18:12
|
|
|
you see what I mean about Unix making it incredibly difficult to write actually secure code
|
|
#
?
Jan 28, 2021 18:42
|
|
|
Cybernetic Vermin posted:just put scheme in the kernel already you cowards, guile as pid 1 is just a stopgap. it's cowardice! They could be gnu / gnu linux
|
|
#
?
Jan 28, 2021 21:33
|
|
|
wha apparently you can use guix with the hurd kernel... fuckin seeya linux *marilyn manson voice* this is the gnu poo poo
|
|
#
?
Jan 28, 2021 22:42
|
|
|
|
|
#
?
Jan 28, 2021 23:29
|
|
|
cjs: doing lots of science at the YAML> prompt
|
|
#
?
Jan 28, 2021 23:31
|
|
|
im the mysterious black cube at the heart of it all
|
|
#
?
Jan 29, 2021 01:25
|
|
|
it turns out transparency doesn't work very well with 219.css
|
|
#
?
Jan 29, 2021 01:32
|
|
|
my stepdads beer posted:im the mysterious black cube at the heart of it all that’s the web it’s powered by NEXTSTEP
|
|
#
?
Jan 29, 2021 02:54
|
|
|
I had a volunteer job at a lab... one of the computers was a nextstation. I think it had a failing hard drive and no one knew the login. We forgot to plug it in one day and someone actually came by to complain. Apparently it ran some sort of dictionary service. Welp, that's my nextstation story.
|
|
#
?
Jan 29, 2021 11:22
|
|
|
i worked in a computer lab that had a forgotten nextstation running a print spooler forgotten until one of the other students rooted it and set up a webserver serving porn that got enough traffic to saturate the building network
|
|
#
?
Jan 29, 2021 16:52
|
|
|
i saw tim bernese-lee’s next station, literally the first ever web server, at the science museum in london
|
|
#
?
Jan 29, 2021 20:38
|
|
|
my design team is working on a custom ASIC right now. One of the design requirements is a hardware RNG. The RNG's requirement reads ""MUST RETURN A UNIQUE VALUE ON EVERY READ" in all caps. gotta love mathematically impossible requirements!
|
|
#
?
Jan 30, 2021 00:33
|
|
|
like if I had a coin-flip machine that returned a unique value on every flip, it would be pretty easy to predict
|
|
#
?
Jan 30, 2021 00:35
|
|
|
Sounds like you just need a pseudorandom permutation and some nonvolatile storage, and it just has to output enough bits based on the maximum number of times it will be used.
|
|
#
?
Jan 30, 2021 00:39
|
|
|
mystes posted:Sounds like you just need a pseudorandom permutation and some nonvolatile storage, and it just has to output enough bits based on the maximum number of times it will be used. skip the storage and just return the sequence 1,2,3,... fully unique on every read.
|
|
#
?
Jan 30, 2021 00:43
|
|
|
Poopernickel posted:my design team is working on a custom ASIC right now. One of the design requirements is a hardware RNG. The RNG's requirement reads ""MUST RETURN A UNIQUE VALUE ON EVERY READ" in all caps. point them at the NIST publications, even if you don't decide to comply directly they're excellent docs to mine for better language to use when specifying RNGs, both deterministic and non deterministic https://csrc.nist.gov/publications/sp800 https://nvlpubs.nist.gov/nistpubs/SpecialPublications/NIST.SP.800-90Ar1.pdf https://nvlpubs.nist.gov/nistpubs/SpecialPublications/NIST.SP.800-90B.pdf https://csrc.nist.gov/CSRC/media/Publications/sp/800-90c/draft/documents/sp800_90c_second_draft.pdf
|
|
#
?
Jan 30, 2021 01:31
|
|
|
p.s. if you actually care about doing RNG (either T or not-T) right, it is a deep fuckin' topic, so you may also want to advise them to perform due diligence on some ip cores and license the best one instead of trying to do it in house
|
|
#
?
Jan 30, 2021 01:43
|
|
|
Poopernickel posted:my design team is working on a custom ASIC right now. One of the design requirements is a hardware RNG. The RNG's requirement reads ""MUST RETURN A UNIQUE VALUE ON EVERY READ" in all caps. I dunno, depends on the other requirements re: device lifetime. "535660986 ... This chip will self destruct." 
|
|
#
?
Jan 30, 2021 02:21
|
|
|
Hm, cool! AMD FreeSync HDMI Patch Appearing For Their Open-Source Linux Driver
|
|
#
?
Jan 30, 2021 12:35
|
|
|
sounds like it's just for proprietary freesync and not hdmi 2.1 vrr, which is being held up by legal bs surrounding the hdmi spec
|
|
#
?
Jan 30, 2021 18:05
|
|
|
I'm really trying to like Arch, I really am. But like I'm getting core dumps from half a dozen applications every few minutes on this computer with an Arch based install, USB barely works and crashes out at least once a day entirely, bluetooth doesn't work until after login. This is even if I switch to the LTS kernal and have been looking through the Arch wiki and using the Arch distro forums to troubleshoot for over a week. I really want to like it with the AUR. My Debian distro I have installed on another drive on this computer Just Works (R) with only a handful of tweaks here and there needed.
|
|
#
?
Jan 30, 2021 18:25
|
|
|
linux fuckin sucks
|
|
#
?
Jan 30, 2021 18:45
|
|
|
Nitrousoxide posted:I'm really trying to like Arch, I really am. But like I'm getting core dumps from half a dozen applications every few minutes on this computer with an Arch based install, USB barely works and crashes out at least once a day entirely, bluetooth doesn't work until after login. This is even if I switch to the LTS kernal and have been looking through the Arch wiki and using the Arch distro forums to troubleshoot for over a week. I really want to like it with the AUR. use debian. dont torture yourself for no reason
|
|
#
?
Jan 30, 2021 18:47
|
|
|
Nitrousoxide posted:I'm really trying to like Arch, I really am. But like I'm getting core dumps from half a dozen applications every few minutes on this computer with an Arch based install, USB barely works and crashes out at least once a day entirely, bluetooth doesn't work until after login. This is even if I switch to the LTS kernal and have been looking through the Arch wiki and using the Arch distro forums to troubleshoot for over a week. I really want to like it with the AUR. I had to google this if it's copypasta or something. As a common arch user I've never experienced anything like this, covering a wide variation of hardware setups and multiple different architectures as well. I'm curious what hardware setup is doing this if you're willing to share. Apps constantly crashing sounds like some serious core issue going on with the install, and it's not "cuz it's arch"
|
|
#
?
Jan 30, 2021 22:07
|
|
|
|
| # ? Apr 24, 2024 21:00 |
|
|
It could be a driver issue if they're using worse-supported hardware, but isn't Debian more strict than Arch about having proprietary drivers in the repos? You can probably install the exact same software on Arch as debian and just didn't.
|
|
#
?
Jan 30, 2021 22:08
|
|
