2024-08-29 01:05:26 <xentrac     > I just walked through some of F83 in the screen editor with my cousin
2024-08-29 01:06:06 <xentrac     > he pointed out that in CGA 2-color mode or on a Hercules, the IBM PC had enough horizontal pixels to display 128 columns of text, so you could display shadow screens side by side
2024-08-29 01:11:11 <veltas      > Yeah I guess in a different colour?
2024-08-29 01:11:30 <veltas      > Or else it would be hard to read, with no gap between
2024-08-29 01:28:12 <xentrac     > yes, he suggested inverse video
2024-08-29 01:28:31 <xentrac     > I think it would *still* be hard to read but maybe less so than typing a ENTER a ENTER a ENTER
2024-08-29 01:29:26 <xentrac     > I did successfully use fix c@ to jump the editor to the definition of c@
2024-08-29 01:29:33 <xentrac     > which is pretty goddamned cool
2024-08-29 01:30:42 <xentrac     > but there were some other things that broke often.  it's getting some kind of error when it reads utility.blk; I think some stray bytes found their way into the file
2024-08-29 01:30:49 <xentrac     > hmm, no, that's not it
2024-08-29 01:31:21 <xentrac     > maybe it's a DOSBox bug.  It can open kernel86.blk fine
2024-08-29 01:34:13 <veltas      > I don't mind a ENTER etc
2024-08-29 01:34:24 <veltas      > I agree inverse video makes most sense
2024-08-29 01:34:39 <veltas      > But might be harsh on the eyes
2024-08-29 01:35:59 <xentrac     > you could set the CGA background to blue so it would be white on blue vs. blue on white
2024-08-29 01:36:08 <veltas      > 128 cols is a bit harsh on the eyes too
2024-08-29 01:36:21 <veltas      > I mean, would you want that on a real period-correct CRT?
2024-08-29 01:37:14 <veltas      > Definitely display shadow block in grey or blue or something, don't change background or invert
2024-08-29 01:37:33 <xentrac     > yeah, blue on white and white on blue were pretty common, and nice
2024-08-29 01:37:58 <xentrac     > in 640×200 CGA mode you couldn't change the foreground color at all, let alone in the middle of a line
2024-08-29 01:38:34 <xentrac     > I think the uncomfortably condensed font would be the bigger problem than the colors.  that and the lack of margins
2024-08-29 01:38:38 <veltas      > Grey background then maybe
2024-08-29 01:38:45 <xentrac     > yeah, that would work
2024-08-29 01:38:59 <xentrac     > I'd probably like blue better, though
2024-08-29 01:39:43 <xentrac     > it puts all the contrast in the red and green channels and eliminates the blue contrast that we're both bad at seeing and bad at focusing on our retinas
2024-08-29 01:40:20 <veltas      > Yeah don't get me started on blue stuff on OLED screens
2024-08-29 01:40:26 <xentrac     > heh
2024-08-29 01:40:48 <veltas      > I didn't notice blue focuses weird until I got an OLED screen
2024-08-29 01:40:52 <xentrac     > a kind of overarching problem I'm having with F83 is that it doesn't worry about error handling much
2024-08-29 01:41:09 <veltas      > 1366x768 LCD ought to be enough for anyone
2024-08-29 01:41:24 <xentrac     > like, there's a " word, which is ANS 94 s"
2024-08-29 01:41:37 <xentrac     > what does it do in interpret state?
2024-08-29 01:41:52 <xentrac     > nothing, it just discards the parsed string
2024-08-29 01:41:59 <amby        > in my experience 1366*768 is a little cramped
2024-08-29 01:42:48 <amby        > but physical screen size is important too
2024-08-29 01:42:57 <amby        > 1920*1080 isn't much good if it's only a 10 inch screen
2024-08-29 01:43:20 <xentrac     > I try open utility.blk and it says, "Open error in A:UTILITY.BLK"
2024-08-29 01:43:36 <xentrac     > hmm, maybe actually that *is* telling me the problem...
2024-08-29 01:44:02 <xentrac     > yes.  I had to mount the directory as drive A: as well
2024-08-29 01:44:06 <xentrac     > not just drive C:
2024-08-29 01:44:12 <unjust      > btw, dosbox-debug is pretty useful, if your build of dosbox includes it
2024-08-29 01:44:25 <xentrac     > thanks!
2024-08-29 01:46:18 <xentrac     > but I would have liked it to tell me "File not found"
2024-08-29 01:48:33 <xentrac     > how does dosbox-debug decide when to update its register/data/code overview windows?
2024-08-29 01:49:03 <unjust      > when you tell it to debug something
2024-08-29 01:49:20 <xentrac     > how do you do that?
2024-08-29 01:49:28 <unjust      > debug program
2024-08-29 01:49:56 <xentrac     > aha, I see
2024-08-29 01:50:21 <unjust      > the navigation is a bit weird
2024-08-29 01:50:27 <unjust      > as you can probably see by those "Scroll:" labels above each panel
2024-08-29 01:50:42 <xentrac     > maybe it has a manual
2024-08-29 01:51:37 <unjust      > well if you type help in the debugger it gives you a command listing
2024-08-29 01:51:51 <xentrac     > I tried that, it didn't
2024-08-29 01:51:57 <unjust      > but it took me a while to realize that i needed to have a terminal with large enough dimensions for it to be displayed
2024-08-29 01:52:41 <unjust      > as in, you have to invoke dosbox-debug on a tty with a large size before you'll see a panel labelled "OutPut/Input     Scroll: home/end"
2024-08-29 01:52:48 <xentrac     > oh, I see, yes
2024-08-29 01:53:44 <xentrac     > this is actually freaking awesome
2024-08-29 01:54:01 <unjust      > 49 lines / rows seems to be the minimal tty height required for that
2024-08-29 01:56:34 <xentrac     > once I've continued with F5, how do I get back to the prompt if I don't have any breakpoints?
2024-08-29 01:57:34 <unjust      > alt+pause
2024-08-29 01:59:59 <xentrac     > aha, Alt-Fn-P on my Lenovo E495, as helpfully explained in https://en.wikipedia.org/wiki/Break_key
2024-08-29 02:00:42 <xentrac     > see, this is what my wife should have used for her assembly language programming class instead of Emu8086
2024-08-29 02:01:36 <unjust      > emu8086 is probably a little bit more friendly for people completely new to assembly and the concept of debuggers though, at least from what i remember of its interface
2024-08-29 02:04:00 <unjust      > another cool dos debugger was insight
2024-08-29 02:04:08 <unjust      > and i'm sure you played with softice
2024-08-29 02:04:32 <xentrac     > no, I never had softice
2024-08-29 02:05:19 <unjust      > https://en.wikipedia.org/wiki/SoftICE / https://archive.org/details/manualzilla-id-5714182
2024-08-29 02:05:24 <xentrac     > I got access to the internet in 01992, see
2024-08-29 02:05:51 <xentrac     > and from that moment forward, MS-DOS had very little interest to me
2024-08-29 02:06:13 <unjust      > earlier than i did, i only got a modem in 1997
2024-08-29 02:06:20 <xentrac     > although I did write, for example, http://canonical.org/~kragen/sw/dev3/morecircles.s
2024-08-29 02:07:00 <xentrac     > yeah, my dad started calling BBSes in 01984
2024-08-29 02:07:29 <xentrac     > so I was regularly participating in online discussions on FidoNet "echoes" by 01988
2024-08-29 02:07:33 <unjust      > nice
2024-08-29 02:08:10 <unjust      > i never accessed fidonet
2024-08-29 02:08:19 <unjust      > did you ever connect to any X.25 networks?
2024-08-29 02:08:49 <xentrac     > yes, I set up an X.25 network in 01996
2024-08-29 02:09:01 <xentrac     > it went from the testing lab to the server room
2024-08-29 02:09:03 <xentrac     > ;)
2024-08-29 02:09:35 <unjust      > was it on telenet (or sprintnet i think it might have become later?)
2024-08-29 02:09:52 <xentrac     > no, it just went from one room to anothe
2024-08-29 02:09:54 <xentrac     > *r
2024-08-29 02:10:02 <xentrac     > on the same floor
2024-08-29 02:10:22 <unjust      > ah ok
2024-08-29 02:10:32 <xentrac     > we were working on a network management system which was somehow embroiled in X.25 networks, I forget how
2024-08-29 02:17:39 <xentrac     > I think it might have needed X.25 to talk to the craft port of one of the pieces of equipment it had to configure
2024-08-29 03:48:37 <zelgomer    > from cmforth:  : U< ( u u - t)   - 2/  0< ; ( 3)
2024-08-29 03:50:00 <zelgomer    > i can't figure out how that works. i guess 2/ must bring in a borrow bit?
2024-08-29 03:57:31 <crc         > https://www.forth.org/OffeteStore/4001-footstepsFinal.pdf page 77 has a few comments on this
2024-08-29 04:16:34 <zelgomer    > lol not much
2024-08-29 04:18:22 <zelgomer    > but reading page 44, i don't fully understand it yet but i think it confirms that there is more going on here than just right shift
2024-08-29 05:06:34 <xentrac     > unsigned less than.  Subtraction is the same for signed and unsigned; 2/ is presumably the signed version, so it will sign-extend
2024-08-29 05:07:40 <xentrac     > perhaps?  No, I think your guess about a borrow bit is the only thing that makes sense
2024-08-29 05:09:59 <xentrac     > subtracting two n-bit quantities to get an n-bit result loses the information about which was greater; in 16 bits, 0 65535 u<, but 1 0 u>.  but the 16-bit subtraction result is 1 in both cases
2024-08-29 05:10:22 <xentrac     > so you need the borrow bit
2024-08-29 09:40:37 <hello-operator> Yeah it must shift through borrow
2024-08-29 10:04:53 <xentrac     > probably worth mentioning, though likely for you it's stating the obvious, but the MuP21 and F21 had 20-bit memory words but 21-bit operand stack registers (and I think return stack too) in order to have space for the carry bit
2024-08-29 10:05:16 <xentrac     > so maybe Chuck was doing a similar trick already on the Novix 4000
2024-08-29 14:41:09 <veltas      > I think DO parameters are in the wrong order
2024-08-29 14:41:43 <veltas      > Pretty rare I don't have to SWAP right before the DO / ?DO
2024-08-29 14:42:20 <veltas      > Unless I'm specifying manually, in which case e.g. 0 10 would make more sense than 10 0
2024-08-29 15:16:59 <zelgomer    > xentrac: didn't realize it had 21-bit stack elements, thanks.
2024-08-29 15:17:25 <zelgomer    > i guess that's where the 21 comes from
2024-08-29 15:17:29 <zelgomer    > in MuP21
2024-08-29 20:03:15 <xentrac     > veltas: plausible.  from time to time I do write a word like : foo 0 do ... loop ;
2024-08-29 20:04:06 <xentrac     > zelgomer: yup!  sure!
2024-08-29 20:48:14 <unjust      > crc: what does this bot do?
2024-08-29 20:48:42 <thrig       > tries to take over the world?
2024-08-29 20:48:45 <crc         > lookup documentation on ans forth words
2024-08-29 20:48:55 <crc         > dpans: .s
2024-08-29 20:48:55 <[bot]dpans  > 15.6.1.0220   .S                           "dot-s"                                   TOOLS
2024-08-29 20:48:55 <[bot]dpans  > ( -- )
2024-08-29 20:48:55 <[bot]dpans  > Copy and display the values currently on the data stack. The format of
2024-08-29 20:48:56 <[bot]dpans  > the display is implementation-dependent.
2024-08-29 20:48:56 <[bot]dpans  > .S may be implemented using pictured numeric output words.
2024-08-29 20:48:56 <[bot]dpans  > Consequently, its use may corrupt the transient region identified by #>.
2024-08-29 20:48:57 <[bot]dpans  > See: 3.3.3.6 Other transient regions.
2024-08-29 20:49:21 <unjust      > dpans: does>
2024-08-29 20:49:21 <[bot]dpans  > 6.1.1250   DOES>                              "does"                                  CORE
2024-08-29 20:49:21 <[bot]dpans  > Interpretation: Interpretation semantics for this word are undefined.
2024-08-29 20:49:22 <[bot]dpans  > Compilation: ( C: colon-sys1 -- colon-sys2 )
2024-08-29 20:49:22 <[bot]dpans  > Append the run-time semantics below to the current definition.  Whether
2024-08-29 20:49:22 <[bot]dpans  > or not the current definition is rendered findable in the dictionary by
2024-08-29 20:49:23 <[bot]dpans  > the compilation of DOES> is implementation defined.  Consume colon-sys1
2024-08-29 20:49:23 <[bot]dpans  > and produce colon-sys2.  Append the initiation semantics given below to
2024-08-29 20:49:23 <[bot]dpans  > the current definition.
2024-08-29 20:49:23 <[bot]dpans  > Run-time: ( -- ) ( R: nest-sys1 -- )
2024-08-29 20:49:24 <[bot]dpans  > Replace the execution semantics of the most recent definition, referred
2024-08-29 20:49:24 <[bot]dpans  > to as name, with the name execution semantics given below.  Return
2024-08-29 20:49:24 <[bot]dpans  > control to the calling definition specified by nest-sys1.  An ambiguous
2024-08-29 20:49:25 <[bot]dpans  > condition exists if name was not defined with CREATE or a user-defined
2024-08-29 20:49:25 <[bot]dpans  > 51
2024-08-29 20:49:26 <[bot]dpans  > ANSI X3.215-1994
2024-08-29 20:49:27 <[bot]dpans  > word that calls CREATE.
2024-08-29 20:49:27 <[bot]dpans  > Initiation: ( i*x -- i*x a-addr )  ( R:  -- nest-sys2 )
2024-08-29 20:49:28 <[bot]dpans  > Save implementation-dependent information nest-sys2 about the calling
2024-08-29 20:49:28 <[bot]dpans  > definition.  Place name's data field address on the stack.  The stack
2024-08-29 20:49:29 <[bot]dpans  > effects i*x represent arguments to name.
2024-08-29 20:49:29 <[bot]dpans  > name Execution: ( i*x -- j*x )
2024-08-29 20:49:30 <[bot]dpans  > Execute the portion of the definition that begins with the initiation
2024-08-29 20:49:30 <[bot]dpans  > semantics appended by the DOES> which modified name.  The stack effects
2024-08-29 20:49:31 <[bot]dpans  > i*x and j*x represent arguments to and results from name, respectively.
2024-08-29 20:49:31 <[bot]dpans  > See: 6.1.1000 CREATE.
2024-08-29 20:49:42 <thrig       > bit spammy there, maybe a link for the longer stuff?
2024-08-29 20:50:21 <crc         > I'll work on that
2024-08-29 20:52:36 <vms14       > tou could try to replace newlines for a space and have some kind of limit to decide whether to flood or link
2024-08-29 20:52:41 <vms14       > you*
2024-08-29 20:53:03 <vms14       > but it's actually a cool bot
2024-08-29 22:13:34 <veltas      > xentrac: I guess that's the intention behind the order
2024-08-29 22:17:15 <veltas      > In my blind drunken rage trying to reinstall Windows 11 it didn't occur to me