2023-06-02 14:40:04 <veltas      > Yeah that locking down of stuff is a total ballache for embedded, hardware and desktop development
2023-06-02 14:40:51 <veltas      > IT lock it all down to make it easier for them to support you and then they don't support anything, they just have the passwords to make stuff happen. 'Security', gentlemen
2023-06-02 14:53:32 <thrig       > give users local admin and suddenly the network is dead because they all clicked on some .bat or .zip or whatever
2023-06-02 15:01:26 <KipIngram   > So, does linux delete your key files if you move or rename the .ssh folder?
2023-06-02 15:01:41 <KipIngram   > Or does it cause a problem because ssh is running at the time?
2023-06-02 15:04:49 <KipIngram   > Anyway, I briefly renamed .ssh so I could do an ssh login without it processing my keys.  And then when I moved it back, my keys were gone.
2023-06-02 15:05:05 <KipIngram   > I had to make new ones and upload the public key to all of my boxes.
2023-06-02 15:11:42 <thrig       > shouldn't, unless someone added a clever script that does stupid things
2023-06-02 15:47:31 <drakonis    > it does not do any of that, no.
2023-06-02 15:47:55 <drakonis    > KipIngram: its probably because you renamed it
2023-06-02 15:48:05 <drakonis    > when you moved it back, did not overwrite anything
2023-06-02 15:49:15 <drakonis    > there is a ssh flag for altering the key processing behavior
2023-06-02 15:49:29 <KipIngram   > Oh, I see.  So you're saying that the ssh code re-created that directory, and then it just abandoned the content of the other one when I re-named?
2023-06-02 15:49:40 <KipIngram   > Instead of doing literally what you'd expect?
2023-06-02 15:49:56 <drakonis    > oh no.
2023-06-02 15:50:04 <drakonis    > it created a new .ssh directory
2023-06-02 15:50:25 <drakonis    > it did not delete anything else though
2023-06-02 15:50:40 <drakonis    > did you use mv?
2023-06-02 15:50:50 <drakonis    > or cp -r?
2023-06-02 15:51:11 <drakonis    > if the former, you might've accidentally moved the original directory inside the new one
2023-06-02 15:51:34 <KipIngram   > I used mv.
2023-06-02 15:51:37 <drakonis    > i see
2023-06-02 15:51:42 <KipIngram   > mv .ssh .ssh-bak
2023-06-02 15:51:49 <drakonis    > and then to move it back?
2023-06-02 15:51:49 <KipIngram   > And then mv .ssh-bak .ssh
2023-06-02 15:51:52 <drakonis    > there it is then
2023-06-02 15:51:54 <drakonis    > its inside .ssh
2023-06-02 15:52:17 <KipIngram   > You see, though, I presumed there was no .ssh after I did the first move.
2023-06-02 15:52:22 <KipIngram   > But it must have gotten re-created.
2023-06-02 15:52:27 <KipIngram   > What you just said makes sense, though.
2023-06-02 15:52:39 <KipIngram   > I think the ssh software must have recreated .ssh behind my back.
2023-06-02 15:52:46 <KipIngram   > That's actually sensible.
2023-06-02 15:52:48 <drakonis    > it most certainly always does that though
2023-06-02 15:53:30 <KipIngram   > Yes - there it is.
2023-06-02 15:53:32 <KipIngram   > Ha ha.
2023-06-02 15:53:55 <KipIngram   > Now that I've created and distributed a new key.
2023-06-02 15:54:18 <drakonis    > woof.
2023-06-02 16:17:41 <thrig       > (ssh -o PubkeyAuthentication=no ... might be a better way to temporarily turn off the use of keys)
2023-06-02 16:30:05 <KipIngram   > Probably.  I just wouldn't have known to go do it.
2023-06-02 16:30:26 <KipIngram   > Anyway, nice to know I still have the old key if it turns out I need it at some point.
2023-06-02 18:05:14 <veltas      > thrig: That is the issue yes, there's no good answer
2023-06-02 18:05:49 <veltas      > Well IMO the answer is IT need resources to actually help, and they need IT that *can* help, which means have some devopsy people or sysadminy types who know what they're doing and are productive
2023-06-02 18:05:58 <veltas      > Lots of IT pros are just microsoft button clickers
2023-06-02 18:06:09 <veltas      > Which has a place, but it's not sufficient for this development problem
2023-06-02 18:06:39 <thrig       > productive and knows == expensive, problematic in a cost center
2023-06-02 18:27:56 <gordonjcp   > veltas: you'd be amazed how much we piss money away at work on "consultants" who will then bring in "contractors" who will then produce an Excel spreadsheet with a bit of scripting that lives on Sharepoint
2023-06-02 18:28:10 <gordonjcp   > veltas: when it's something I could knock together in an afternoon with Django
2023-06-02 18:29:21 <KipIngram   > That's the way of the corporate world.
2023-06-02 18:29:31 <KipIngram   > Such "processes" look swanky on paper.
2023-06-02 18:29:48 <KipIngram   > I've always had an affinity for small companies.
2023-06-02 18:30:01 <KipIngram   > Which makes it pretty funny that I work for IBM.
2023-06-02 18:30:42 <gordonjcp   > oh, which bit?
2023-06-02 18:30:52 <gordonjcp   > I used to do tech support at Greenock, about 20 years ago
2023-06-02 18:30:57 <KipIngram   > Flash data storage.
2023-06-02 18:31:01 <KipIngram   > "FlashSystems"
2023-06-02 18:31:18 <gordonjcp   > it's funny, I was talking to a mate of mine who did tech support for Desktop stuff in another building, about SameTime and Notes :-)
2023-06-02 18:31:49 <gordonjcp   > there was a guy call Sridhar Ayengar that was on the Classic Computer Mailing List, if my UK shifts lined up with his US ones we'd sometimes chat on Sametime
2023-06-02 18:31:58 <KipIngram   > I do performance testing, which I do think is somewhat cooler than "plain vanilla testing," but it's still a lot less high brow than the sort of stuff I used to do.
2023-06-02 18:32:28 <KipIngram   > But I endure a whole lot less stress than I did back then too.
2023-06-02 18:43:35 <KipIngram   > You know, I generally think Chuck's advice on Forth is sound, but it sure is hard for me to fully let go of the ! word.
2023-06-02 18:43:56 <KipIngram   > He's advocated replacing that with >A A!
2023-06-02 18:44:37 <KipIngram   > And that certainly is appealing from a hardware Forth perspective, because it eliminates the only operation that drops two items from the stack.
2023-06-02 18:44:55 <KipIngram   > The only atomic operation, I mean.
2023-06-02 18:45:18 <KipIngram   > There's TYPE, which I implement as a primitive for performance.
2023-06-02 18:45:25 <KipIngram   > CMOVE and friends.
2023-06-02 18:45:55 <KipIngram   > But somehow @ and ! just seem like part of the "soul" of the thing.
2023-06-02 18:46:35 <unjust      > is that an A register? or an auxiliary stack? (sorry sent this to the wrong place at first)
2023-06-02 18:46:43 <KipIngram   > A register.
2023-06-02 18:47:47 <KipIngram   > He's gone as far as to say that that register will change your coding style.
2023-06-02 18:53:14 <unjust      > how does it replace ! ?
2023-06-02 18:59:51 <crc         > : ! ( value address ) >A ( set A to address ) A! ( store value to address A points to ) ;
2023-06-02 19:09:05 <unjust      > thanks crc
2023-06-02 19:09:35 <crc         > no problem
2023-06-02 19:10:09 <unjust      > is @ then defined as : @ ( address -- value ) >A A@ ; ?
2023-06-02 19:11:04 <crc         > Yes, at least in the models I've done with an A register
2023-06-02 19:24:22 <KipIngram   > Yeah, it simplifies the stack logic a little, but probably more important is that the address persists and can be re-used, post-incremented, etc.
2023-06-02 19:26:08 <KipIngram   > Hewlett Packard calculators had the notion of a "LastX" register that is in the neighborhood - it would be a bit like your usual Forth address going into LastX when you used it.
2023-06-02 19:26:20 <KipIngram   > It's not exactly the same model, of course, but you can get the value back from there.
2023-06-02 19:27:52 <thrig       > handy in an equation where x turns up a lot
2023-06-02 19:28:48 <KipIngram   > Yes, and since that thing had only a 4-deep stack, that deepest register would just keep feeding its value down on pops.  Very handy for polynomial evaluation.
2023-06-02 19:29:52 <KipIngram   > a*x^2 + b*x + c --> x push push push a * b + * c +
2023-06-02 19:30:03 <KipIngram   > Handier for higher order polys.
2023-06-02 19:30:37 <KipIngram   > You just keep doing * coef + over and over until you're done.
2023-06-02 19:36:56 <thrig       > taylor made, you could say
2023-06-02 19:38:14 <KipIngram   > You could.  My favorite feature, though, was the way you could "store op" or "recall op" with any stack register, all of it could be indirect too.
2023-06-02 19:38:33 <KipIngram   > so, RCL/ IND Z or whatever.
2023-06-02 19:43:20 <KipIngram   > Oh, I decided that (golden ratio)^n / sqrt(5) rounding off to Fibonacci numbers wasn't as profound as I first thought.
2023-06-02 19:44:00 <KipIngram   > The place the golden ratio comes from to begin with is if you have an a by b rectangle, with a > b, then if a/b = (a+b)/a then that's the golden ratio.
2023-06-02 19:44:16 <KipIngram   > So you almost SEE the Fibonacci process showing there on the right.
2023-06-02 19:53:44 <KipIngram   > It's really kind of a catch 22 - on the one hand, I'm looking to pare the design of a system down to these nice "portable operations" that I can build primitives with, and that says "minimize."  On the other hand, once you've paid the price of finding your way to a primitive, you're incentified to make it possible for that to do as much for you as possible, which pushes you toward a primitive heavy system.
2023-06-02 19:53:55 <KipIngram   > Either way you go you're giving something up.