The shell remembers what was typed before it

Up walks back through the last eight lines and Down forward again. It exists
only because the keys reach the system now: until A1 and A2 there was nothing to
press Up at, and the line was assembled somewhere the shell could not see.

A RING RATHER THAN A LIST. A ninth line pushes the oldest out by moving where
the ring starts, not by moving any of the lines - so keeping a line costs a copy
of that line and nothing else, however full the history is. Eight is a power of
two, so which slot an entry lives in is an AND. The ISA had the awkward part
already: A and B are a sixteen bit shift register, so one SHR with B empty turns
a slot number into the offset of a 128 byte slot, high byte and low, ready for
DPUW.

A NINTH SLOT HOLDS WHAT WAS BEING TYPED when Up left it, and Down brings it
back. Losing a half written line to a keypress is the sort of small rudeness
that makes a thing unpleasant to use, and it costs one slot to avoid.

An empty line is not kept, and neither is one the same as the line already at the
top. The test proves the second by looking one further back: if a repeated
command were kept twice, the line behind the newest would be the same line
again.

The redraw had to learn to rub out. One space was enough while the only thing
that shortened a line was taking one character out of it; a recalled line
replaces the whole of it, and a short line over a long one left the tail of the
long one on screen looking like part of what you were typing. It now covers
exactly what was lost - which turned out to be one space fewer than before in
the cases that GREW, so two lines of cosmosEditKeys lost a trailing space that
was never doing anything.

Costs 1157 bytes of Data Memory, taking CosmOS to 6220 of the 8192 it has before
a loaded program's data begins. Worth writing down: that is the budget, and this
is the largest single thing in it.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01E2JrLzFvuFX9fgi1LDRjrW
This commit is contained in:
Anachronaut
2026-09-01 12:24:23 -04:00
co-authored by Claude Opus 5
parent 81e544eb3d
commit 71f6e215f9
7 changed files with 409 additions and 9 deletions
+328 -6
View File
@@ -453,6 +453,8 @@ editLine:
STA.1
SETD.1 EditAt
STA.1
SETD.1 EditDrawn
STA.1
SETD.1 ConsoleEndOfInput
STA.1
@@ -531,6 +533,12 @@ editNotEnd:
INIB 0x85
XOR
BRQ editEnd
INIB 0x80
XOR
BRQ editUp
INIB 0x81
XOR
BRQ editDown
; Anything else is a character if it is printable and nothing at all if it is not. The
; console's own keys that this shell has no use for land here and are ignored rather than
@@ -600,6 +608,10 @@ editInsertPut:
SETD.1 EditChar
LDA.1
OUTA 0x00
SETD.1 EditLength
LDA.1
SETD.1 EditDrawn
STA.1
BRI editKey
editInsertRedraw:
CALL editRedraw
@@ -636,6 +648,10 @@ editBack:
OUTA 0x00
INIA 0x08
OUTA 0x00
SETD.1 EditLength
LDA.1
SETD.1 EditDrawn
STA.1
BRI editKey
editBackMiddle:
CALL editTakeOut
@@ -722,6 +738,64 @@ editEnd:
LDA.1
SETD.1 EditAt
STA.1
BRI editShow
; ---- Backwards and forwards through what was typed before ----
;
; The line on the screen is replaced outright, so both of these redraw. That is also what
; makes the padding below matter: a long line replaced by a short one leaves the tail of the
; long one behind unless something rubs it out.
editUp:
SETD.1 HistoryPick
LDA.1
BRA editKey ; Already at the oldest one kept.
; Leaving the line being typed for the first time, so it is put somewhere. Down brings it
; back, and a person who pressed Up to look at something gets their line returned rather
; than taken.
SETD.1 HistoryCount
LDB.1
CCF
SUB
BNQ editUpMoving
SETD.3 HistoryTyped
CALL historyPut
editUpMoving:
SETD.1 HistoryPick
LDA.1
DECA
STA.1
CALL historySlot
BRI editRecall
editDown:
SETD.1 HistoryPick
LDA.1
SETD.1 HistoryCount
LDB.1
CCF
SUB
BRQ editKey ; Already back at the line being typed.
SETD.1 HistoryPick
LDA.1
INCA
STA.1
SETD.1 HistoryCount
LDB.1
CCF
SUB
BRQ editDownTyped
SETD.1 HistoryPick
LDA.1
CALL historySlot
BRI editRecall
editDownTyped:
SETD.3 HistoryTyped
editRecall:
CALL historyTake
CALL editRedraw
BRI editKey
editShow:
SETD.1 EditAt
LDA.1
@@ -752,6 +826,10 @@ editDone:
CALL editPlace
editDoneEnd:
CALL newLine
; Kept before the buffer is handed over and while EditLength still says how long it is.
; Only on this path: a console that ran out was not somebody finishing a line.
CALL historyAdd
editFinish:
; Line mode, with the cursor still on: that is how boot left the console and how every
@@ -799,10 +877,48 @@ editRedrawNext:
DECB
BRI editRedrawNext
editRedrawTail:
; One space after it. A line that has just got shorter has a character left over on the
; end, and printing over it costs less than working out whether there is one.
INIA 0x20
OUTA 0x00
; ---- Rubbing out what is no longer there ----
;
; One space was enough while the only thing that ever shortened a line was taking one
; character out of it. Pressing Up replaces the whole line, and a long one replaced by a
; short one leaves the tail of the long one sitting on the screen looking like part of
; what you are typing.
;
; So the spaces go on for exactly as far as the line has shrunk, which is one after a
; character was taken out and none at all when one was put in. printSpaces takes none for
; an answer and prints nothing, which is what makes that case cost nothing to allow.
SETD.1 EditDrawn
LDA.1
SETD.1 EditLength
LDB.1
CCF
SUB
BRC editRedrawGrew ; Borrowed, so the line is longer than what is up there.
MVQA
BRI editRedrawPad
editRedrawGrew:
RSTA
editRedrawPad:
SETD.1 EditPrinted
STA.1
CALL printSpaces
; How much went to the screen altogether, which is how far back the start of the line is.
SETD.1 EditPrinted
LDA.1
SETD.1 EditLength
LDB.1
CCF
ADD
MVQA
SETD.1 EditPrinted
STA.1
; And what is up there now is exactly the line.
SETD.1 EditLength
LDA.1
SETD.1 EditDrawn
STA.1
CALL editAnchor
SETD.1 EditAt
@@ -823,9 +939,8 @@ editAnchor:
INA 0x04
SETD.1 EditWalkColumn
STA.1
SETD.1 EditLength
SETD.1 EditPrinted
LDA.1
INCA ; The space was printed too.
SETD.1 EditWalkBack
STA.1
editAnchorWalk:
@@ -919,6 +1034,183 @@ editPlaceDone:
OUTA 0x04
RET
; ---- What was typed before ----
;
; Eight lines, oldest first, in a RING: a ninth pushes the oldest out by moving where the
; ring starts rather than by moving any of the lines. Eight is a power of two, so which slot
; an entry lives in is an AND rather than a division, and nothing is ever copied to make
; room.
;
; A NINTH SLOT HOLDS WHAT WAS BEING TYPED when Up was first pressed, and Down walks back into
; it. Losing a half written line to a keypress is exactly the sort of small rudeness that
; makes a thing unpleasant to use, and it costs one slot to avoid.
;
; The history belongs to the SHELL rather than to the console, which is the whole reason it
; can exist at all: until the keys arrived here, there was nothing to press Up at.
; DP3 to history entry A, where entry nought is the oldest one still kept.
historySlot:
SETD.1 HistoryStart
LDB.1
CCF
ADD
MVQA
INIB 0d7
AND
MVQA
SETD.3 HistoryLines
; A slot is 128 bytes, and A and B are a sixteen bit shift register: one rotation right
; with B empty turns the slot number into the offset of the slot, high byte and low.
RSTB
SHR
DPUW.3
RET
; Copies the line being edited into the slot DP3 names.
historyPut:
SETD.1 EditBase
LDD.0.1
SETD.1 EditLength
LDB.1
historyPutNext:
BRB historyPutEnd
LDA.0
STA.3
INCD.0
INCD.3
DECB
BRI historyPutNext
historyPutEnd:
RSTA
STA.3
RET
; Copies the line in the slot DP3 names into the buffer, and puts the cursor at the end of
; it, which is where somebody who has just recalled a line wants to be.
;
; A LINE TOO LONG FOR THE BUFFER IS CUT SHORT rather than written past the end of it. The
; history is shared with the monitor, which reads into forty bytes where the shell reads
; into a hundred and twenty seven, so this is a real case and not a defensive one.
historyTake:
SETD.1 EditBase
LDD.0.1
RSTA
SETD.1 EditLength
STA.1
historyTakeNext:
SETD.1 EditLength
LDA.1
SETD.1 EditRoom
LDB.1
CCF
SUB
BNC historyTakeDone ; No room for another character.
LDA.3
BRA historyTakeDone
STA.0
INCD.0
INCD.3
SETD.1 EditLength
LDA.1
INCA
STA.1
BRI historyTakeNext
historyTakeDone:
RSTA
STA.0
SETD.1 EditLength
LDA.1
SETD.1 EditAt
STA.1
RET
; DP0 and DP3 name strings ending in zero bytes. Q is zero if they are the same.
;
; textSame does this already and takes its two strings in DP0 and DP1, and getting a pointer
; from DP3 into DP1 costs a store, a scratch word and a load. This is shorter than that and
; says what it is doing.
historySameAs:
LDA.0
LDB.3
XOR
BNQ historySameNo
LDA.0
BRA historySameYes ; They matched and they both ended.
INCD.0
INCD.3
BRI historySameAs
historySameNo:
INIA 0x01
RSTB
CCF
ADD ; Q is one.
RET
historySameYes:
RSTA
RSTB
CCF
ADD ; Q is zero.
RET
; The line just finished, kept. An empty one is not - that is somebody pressing Return - and
; neither is one the same as the newest already there, because running a command twice
; should not put it in twice.
historyAdd:
SETD.1 EditLength
LDA.1
BRA historyAddDone
SETD.1 HistoryCount
LDA.1
BRA historyAddPut ; Nothing kept yet, so nothing to be the same as.
DECA
CALL historySlot
SETD.1 EditBase
LDD.0.1
CALL historySameAs
BRQ historyAddDone
historyAddPut:
SETD.1 HistoryCount
LDA.1
INIB 0d8
CCF
SUB
BRC historyAddRoom ; Fewer than eight, so there is a slot on the end.
; Full. The new line goes over the oldest and the ring starts one further along, which
; forgets the oldest without moving any of the others.
INIA 0d8
CALL historySlot
CALL historyPut
SETD.1 HistoryStart
LDA.1
INCA
INIB 0d7
AND
MVQA
STA.1
BRI historyAddDone
historyAddRoom:
SETD.1 HistoryCount
LDA.1
CALL historySlot
CALL historyPut
SETD.1 HistoryCount
LDA.1
INCA
STA.1
historyAddDone:
; However that went, the next Up starts from the newest again.
SETD.1 HistoryCount
LDA.1
SETD.1 HistoryPick
STA.1
RET
; ---- A command that did not work ----
;
; The one place a failure is recorded, so that the thing reading lines out of a file can
@@ -5337,6 +5629,36 @@ EditWalkColumn:
EditWalkBack:
0x00
; How many characters of the line are on the screen, which is not always how many are in it:
; between a line getting shorter and the screen being put right, the screen still has the
; old one on it. That gap is one character wide for a Delete and a whole line wide for a
; recalled one.
EditDrawn:
0x00
; How much a redraw sent to the screen, line and rubbing out together. The start of the line
; is that far back from where the cursor ended up.
EditPrinted:
0x00
; Eight lines of a hundred and twenty eight bytes, and one more holding whatever was being
; typed when somebody pressed Up.
HistoryLines:
#Reserve 0d1024
HistoryTyped:
#Reserve 0d128
; Which slot holds the oldest, which is what moves when a ninth line pushes one out.
HistoryStart:
0x00
; How many of the eight hold anything.
HistoryCount:
0x00
; Which one is on the screen, counting the oldest as nought. HistoryCount means none of them
; is: what is on the screen is the line being typed.
HistoryPick:
0x00
; A hundred and twenty seven characters and the zero byte that ends them. It was sixty three
; until the shell learned to edit a line, which is when the limit started to be felt: a copy
; between two disks with a directory on each is most of the way there before anything has