2022-08-26 02:31:54 <remexre     > how reasonable does it feel to people to call forth "a scripting language for building the binary you want in memory" ?
2022-08-26 02:33:03 <remexre     > (as opposed to a traditional compiled language)
2022-08-26 02:38:51 <dave0       > remexre: forth is really close to compiled
2022-08-26 02:40:21 <dave0       > i remember reading that the threading used in forth (ie. a list of subroutines to call in order) is how early cobol compilers worked
2022-08-26 02:40:39 <dave0       > and i believe B, the precursor to C, was also compiled to threaded code
2022-08-26 02:41:48 <remexre     > i'm not familiar enough with cobol to usefully comment on it, but I guess C would be the foil I'd use
2022-08-26 02:42:10 <remexre     > like the idea of back-patching a jump is just... not a thing you'd even think of being expressible in C
2022-08-26 02:42:28 <remexre     > (not that it's standard forth, i suppose)
2022-08-26 02:44:08 <remexre     > and of the languages that let you (iirc ruby and lua have notions of this?), they allow it because they're interpreted in a more... "heavy-weight" way?
2022-08-26 02:44:32 <remexre     > i guess a while(1) do(*opcode++); feels "heavier" than NEXT
2022-08-26 02:44:54 <remexre     > er, do_opcode(...) i suppose, forgot do's a keyword lol
2022-08-26 02:47:11 <dave0       > remexre: my love of assembly means i admire back patching the jumps :-)
2022-08-26 02:48:29 <dave0       > remexre: my forth `execute` uses self-modifying code to call the execution token
2022-08-26 02:49:19 <remexre     > huh, does it mess with its NEXT or w/e to make it a direct jump? and why do it that way; does it help enough with branch prediction to be worth the icache flushes?
2022-08-26 02:49:40 <remexre     > ig it wouldn't be NEXT, but the jump at the end
2022-08-26 02:49:46 <dave0       > remexre:   : execute   LIT [ here @ ] nop ! [ here @ swap ! ] nop ;
2022-08-26 02:50:41 <dave0       > remexre: it stores the address directly into the threaded code
2022-08-26 02:51:19 <remexre     > hm, I'm not understanding what the  here @  is about?
2022-08-26 02:51:41 <remexre     > wouldn't the nop not have been written yet?
2022-08-26 02:51:45 <dave0       > oh shit yeah my here is non standard
2022-08-26 02:51:50 <remexre     > ah
2022-08-26 02:52:03 <dave0       > remexre: it back-patches like a branch
2022-08-26 02:52:30 <remexre     > is your here like the standard  here 0 ,  or something?
2022-08-26 02:52:31 <dave0       > hmm i could probably use an unconditional branch to achieve it ... that's interesting
2022-08-26 02:52:51 <dave0       > remexre: oh i just made here a normal variable
2022-08-26 02:52:57 <remexre     > ah
2022-08-26 02:54:11 <remexre     > yeah, i'm not sure how well that'd work with a real cpu; iirc there's a limit on how close to the current insn ptr you can self-modify code before bad things happen
2022-08-26 02:54:25 <remexre     > on intel's impl of x86_64, at least
2022-08-26 02:55:15 <dave0       > i use that code on direct and indirect threaded, so it's not actually code, it's a pointer .. but yes it's still indirect
2022-08-26 02:55:41 <dave0       > or rather unpredictable
2022-08-26 02:57:02 <dave0       > remexre: does this make more sense for my execute?     : execute   LIT [ here 0 , ] ! [ here swap ! 0 , ] ;
2022-08-26 02:57:41 <dave0       > is it more traditional code?
2022-08-26 02:58:00 <remexre     > yeah, I think i understand it
2022-08-26 02:58:26 <remexre     > kinda funky, but saves you a primitive I guess :)
2022-08-26 02:58:56 <dave0       > :-)
2022-08-26 02:59:54 <dave0       > i've never seen it in any other language (except assembly :-)
2022-08-26 03:00:10 <dave0       > that low-level control
2022-08-26 03:00:20 <remexre     > yeah
2022-08-26 11:33:46 <KipIngram   > That's pretty funky, dave0, but I guess "whatever works" applies here.
2022-08-26 11:34:28 <KipIngram   > A few weeks ago I was trying to write a boot sector Forth that could be used for a full system bootstrap.
2022-08-26 11:34:35 <KipIngram   > That meant it needed to offer BIOS access.
2022-08-26 11:35:00 <KipIngram   > A "clean" implementation would have had separate functions for int 10h, int 13h, etc. etc., but there wasn't room.
2022-08-26 11:35:13 <thrig       > EFI support might be more complicated
2022-08-26 11:35:19 <KipIngram   > So it uses self modifying code there - the byte that specifies the interrupt level has to be plugged before you call it.
2022-08-26 11:35:32 <thrig       > (some systems ain't even shipping with legacy bios no more)
2022-08-26 11:35:44 <KipIngram   > yeah, EFI did look more compicated, but more functional too (network boots, etc.)
2022-08-26 11:35:54 <KipIngram   > I stuck with what I was familiar with from the past.
2022-08-26 11:36:17 <thrig       > why we got not-invented-here EFI instead of openfirmware...
2022-08-26 12:08:12 <KipIngram   > Corporate agendas.
2022-08-26 12:08:19 <KipIngram   > "Control."
2022-08-26 12:08:56 <KipIngram   > We spent all that time fretting over how to avoid tyrannical governments, and while we weren't looking we got tyrannical business instead.
2022-08-26 12:11:41 <veltas      > Quite simply blame Intel
2022-08-26 12:12:16 <veltas      > UEFI is pretty much DOS, it's actually far more complicated than DOS ever was
2022-08-26 12:12:40 <veltas      > Your OS boots from an OS
2022-08-26 12:13:15 <KipIngram   > Yeah.  The general phenomenon is across the board, though.  Eisenhower tried to warn us as he left office about "the military industrial complex," but apparently we didn't pay attention.
2022-08-26 12:13:49 <veltas      > Target the UEFI Shell for your Forth, not much harder to make it into an OS at that point
2022-08-26 12:14:07 <thrig       > it's hard to make someone pay attention when hey hey money!!
2022-08-26 12:14:11 <veltas      > But do we really need another OS?
2022-08-26 12:14:34 <KipIngram   > We probably actually don't "need" one - we've got a good  FOSS OS.
2022-08-26 12:14:42 <veltas      > I actually think UEFI exists because firmware developers have nothing better to do, but that's not a good reason either
2022-08-26 12:14:52 <veltas      > That's why systemd exists, it's not a good thing
2022-08-26 12:14:56 <KipIngram   > oh, but it IS the reason for a lot of things.
2022-08-26 12:15:15 <thrig       > systemd is brilliant (if you run a company that makes money off of service contracts)
2022-08-26 12:15:18 <KipIngram   > I've thought for a long time that the MS Office user interface hit a "best point" decades ago, but they just keep right on dicking with it.
2022-08-26 12:15:28 <KipIngram   > Every little manager has to figure out how to keep his team in work.
2022-08-26 12:15:33 <KipIngram   > Job's never "done."
2022-08-26 12:16:07 <KipIngram   > No one's ever going to say "We've completed our work - this team can be dissolved now.'
2022-08-26 12:16:28 <KipIngram   > because you might just get "dissolved" right onto the street.
2022-08-26 12:16:48 <veltas      > I know people that say this
2022-08-26 12:16:59 <veltas      > And they're the ones that understand the job market and will walk into another job
2022-08-26 12:17:07 <KipIngram   > I think corporations are just packed with little "self serving groups" that aren't at all tuned in to any "global mission."
2022-08-26 12:17:14 <KipIngram   > Sure - there are exceptions to every rule, and bless them.
2022-08-26 12:17:25 <KipIngram   > Good people - still got a few around.
2022-08-26 12:18:06 <veltas      > UEFI Shell is pretty much a FOSS OS
2022-08-26 12:18:34 <veltas      > Most of the shell is open source code, then the rest is all firmware baked into your PC that wasn't open source regardless
2022-08-26 12:18:42 <KipIngram   > I really don't know enough about it to comment.
2022-08-26 12:18:55 <veltas      > I've written programs for it at my last job
2022-08-26 12:19:00 <KipIngram   > I used BIOS when it was current, and I still remember fairly well how that game worked.
2022-08-26 12:19:02 <veltas      > Like a ROM flashing utility
2022-08-26 12:19:19 <veltas      > BIOS was a very thin interface for firmware
2022-08-26 12:19:25 <KipIngram   > Yes.
2022-08-26 12:19:28 <veltas      > And DOS was playing the role UEFI Shell plays
2022-08-26 12:19:41 <veltas      > UEFI Shell plus all the 'boot services' in UEFI
2022-08-26 12:20:21 <veltas      > Some of the boot services are just replacements for BIOS functions
2022-08-26 12:21:27 <KipIngram   > Well, my goal was to make everything needed for a bootstrap fit into 512 bytes, and I was able to do that (though I haven't tested it in any way, yet).  I tinkered with qemu enough to decide that it looked like it was a viable way to test it, but I haven't gone any further.
2022-08-26 12:21:47 <KipIngram   > Bad habit of mine - I dive into things pretty immersively, then after a while my interest gets pulled off to something else.
2022-08-26 12:22:11 <KipIngram   > At the moment I'm brushing up my old chemistry learning.
2022-08-26 12:22:31 <KipIngram   > My dad was a chemistry professor - I decided it was a little shameful for me to have forgotten as much of it as I had.
2022-08-26 12:22:59 <KipIngram   > I'm pretty on top of the "physics part" - electron structures, etc., but beyond that it had largely faded away.
2022-08-26 12:24:45 <KipIngram   > It's pretty cool how the shape of the periodic table pops right out of that physiscs.  Schrodinger equation in a central potential --> perioic table.
2022-08-26 12:26:17 <KipIngram   > Oh, I discovered this morning that my brand new work Mac wasn't charging through it's magnetic port.
2022-08-26 12:26:30 <KipIngram   > Fortunately it will also charge via USB C, and that does work.
2022-08-26 12:26:53 <KipIngram   > Don't know if it's a bad cable or a dirty connection or what yet - I'll figure that out later.
2022-08-26 12:27:43 <thrig       > the power supply is the most unreliable part of this here 2009 macbook
2022-08-26 12:34:07 <KipIngram   > This is just the second Mac I've ever used.  Had a 13" MacBook Air before this one.
2022-08-26 12:35:07 <KipIngram   > I just set it and my personal notebook side by side on the sofa in front of me during the day - end of day I just close the Mac and slip it under the Linux unit.
2022-08-26 12:35:17 <KipIngram   > That's my 10 second commute these days.
2022-08-26 12:39:34 <thrig       > besides this 2009 macbook the old apple //e was good
2022-08-26 12:40:25 <KipIngram   > I've been pretty happy with the general hardware quality of everything Apple I've ever used.  I did a brief stint with an iPad, and I've had a couple of iPods.
2022-08-26 12:40:31 <KipIngram   > All seemed like solid stuff.