The shell runs what it reads, not what a name ends in

A typed word had ".sbx" pasted on the end of it before anything went
looking, which is why "notes.txt" sent the shell after a notes.txt.sbx
that was never going to exist, and why a script could only be started
with "do". The extension was what made a file reachable by name.

Now the word as typed is asked for first and the word with the extension
on it only after that misses. What comes back is dispatched on what is
inside it: SBEX loads and starts, "#!" is read as lines. The loader
already refuses anything that is not SBEX and the script reader already
refuses anything without the shebang, so the two kinds of runnable file
turn each other away and neither has to know the other exists.

The suffix can only ever be a second guess, so a file that is really
there always beats one that would have to be invented and every program
already on a disk still starts by the short name people type for it. It
costs a second walk of each directory when a word is not found in it.

Tab completion offers a file under the name it actually has, and the
suffix stripper is gone: a first word can now be any file at all, and
offering only the ones ending .sbx would hide the scripts.

The video capture windows moved out a hundred thousand cycles, because
starting Lander now walks four names where it walked two.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01E2JrLzFvuFX9fgi1LDRjrW
This commit is contained in:
Anachronaut
2026-09-06 13:23:14 -04:00
co-authored by Claude Opus 5
parent 4e61158b11
commit fd962f0084
7 changed files with 203 additions and 121 deletions
+45 -23
View File
@@ -28,8 +28,9 @@ for itself.
- Loadable Applications: Validate SBEX files, copy their Program and Data segments into
the addresses for which they were assembled, and start them at their declared entry
point.
- Invocation By Name: A word the shell has no command for is looked for on the disk as
`<name>.sbx`, and loaded and started if it is there. Built-in commands are tried first.
- Invocation By Name: A word the shell has no command for is looked for on the disk, and
what is found is run according to what is inside it - a program is loaded and started, a
script is read as lines. Built-in commands are tried first.
- Resident Services: Applications can print strings and numbers, read lines, receive their
command arguments, read and write files, and return to the shell through named software
interrupts.
@@ -211,6 +212,10 @@ A `$` with nothing name-like after it is just a `$`.
the same way it does at the prompt, because the only thing a script changes is where the next
line comes from - the shell splits it, matches it and runs it without knowing the difference.
**Typing a script's name runs it too**, the same way typing a program's name does; `do` is how
a file is reached by its whole name, the way `load` is for a program. See Starting An
Application By Name.
```
#! script
; Build the system and put it where the machine will find it.
@@ -327,7 +332,7 @@ CosmOS currently provides these built-in commands:
| `cd [path]` | Go to a directory, or to the root with nothing after it. |
| `mkdir <path>` | Make a directory. |
| `rmdir <path>` | Remove one, if it is empty. |
| `<name> [words]` | Any word the shell does not recognise is looked for on the disk as `<name>.sbx`, and loaded and started if it is there. |
| `<name> [words]` | Any word the shell does not recognise is looked for on the disk, and started if it is found: loaded if it is a program, read as lines if it is a script. |
| `delete <file>` | Remove a file from the filesystem and release its blocks. |
| `rename <file> <to>` | Give a file a different name without moving its contents. |
| `set [name [value]]` | Give a name a value, or say what the names are. See Names For Things. |
@@ -422,18 +427,23 @@ program; anything after it is a file.
| Where | What is offered |
| -- | -- |
| The first word | the fifteen commands, and programs - under the name you would type, with the extension taken off |
| The first word | the fifteen commands, and every file, under the name it actually has |
| After it | anything on the disk, under its real name |
| Either, with a `/` in it | whatever is in the directory the word names |
A directory answers with a separator on the end instead of a space, which says what it is and
lets the next part be typed straight away.
**Programs are looked for in the three places the shell would look to run one**: where you
are, `/Apps` on the disk you are on, and `/Apps` on drive 0. Offering something the shell
would not find would be finishing a word into a thing that then does not work. A program that
is in two of those places is offered twice and so is never the only match, which costs it the
space after it - the shell is comparing names, not deciding which file it would have run.
**Files are looked for in the three places the shell would look to run one**: where you are,
`/Apps` on the disk you are on, and `/Apps` on drive 0. Offering something the shell would not
find would be finishing a word into a thing that then does not work. A file that is in two of
those places is offered twice and so is never the only match, which costs it the space after
it - the shell is comparing names, not deciding which file it would have run.
A first word used to be offered with `.sbx` taken off, and anything without that ending was
not offered at all. That was right while the extension was what made a file reachable; now
that the shell decides by reading, hiding the files without one would hide scripts - which is
much of what a first word is.
Nothing typed and nothing matching both do nothing, quietly.
@@ -640,22 +650,34 @@ directory's unused fields hold, the version, the free count.
### Starting An Application By Name:
A word the shell has no command for is not immediately an error. Before saying so, the
shell adds `.sbx` to it unless it is already there, looks for a file of that name, and if
one is there loads it and starts it exactly as `load` and `run` would. Whatever followed
the word reaches the program through `osArgument`, the same way and by the same route as
whatever follows `run`.
shell looks for a file of that name and, if one is there, decides what to do with it by
what is inside it. Whatever followed the word reaches the program through `osArgument`, the
same way and by the same route as whatever follows `run`.
Three properties of this are deliberate:
**The shell runs what it reads.** A file beginning `SBEX` is loaded and started exactly as
`load` and `run` would; a file beginning `#!` is read as lines exactly as `do` would. The
loader already refuses anything that is not `SBEX` and the script reader already refuses
anything without the shebang, so between them the two kinds of runnable thing turn each
other away and neither has to know the other exists. **The name says nothing about which
happened**, which is what makes `.sbx` a convention rather than a rule.
Four properties of this are deliberate:
**Built-in commands are tried first and always win.** The search happens only after the
whole dispatch chain has failed to match, so a file named `dir.sbx` cannot become `dir`.
The commands that are worth trusting when the disk is the thing being doubted stay
trustworthy.
**The extension is what makes a file reachable by name.** Typing `notes` looks for
`notes.sbx`, and typing `notes.txt` looks for `notes.txt.sbx`. A text file therefore
cannot be started by typing what it is called, whatever happens to be inside it. Only
`load` reaches a file by its literal name.
**The word as typed is tried first, and only then the word with `.sbx` on the end.** The
first is what makes a file reachable under the name it actually has - a script, a program
somebody dropped the extension from, or a launcher. The second is what keeps every program
already on a disk startable by the short name people have always typed for it. The suffix
can only ever be a second guess, so a file that is really there always wins over one that
would have to be invented, and typing `notes.txt` no longer sends the shell after a
`notes.txt.sbx` that was never going to exist.
It costs a second walk of each directory when a word is not found in it, which is the price
of the exact name being asked for first.
A path works here too, so `/Apps/Say hello` starts `/Apps/Say.sbx` and gives it `hello`.
@@ -666,10 +688,10 @@ said where to look, so only that place is tried. Neither is stored anywhere, so
nothing to configure and nothing to go stale - a search path somebody could set would need
somewhere to live between one boot and the next, and there is no such place yet.
**A file that is found but is broken says so.** If `notes.sbx` exists and is not an SBEX
program, typing `notes` reports `not a program` rather than `I do not know: notes`.
Reporting an unknown command about a file that is sitting on the disk would send somebody
looking in the wrong place.
**A file that is found but is neither says so.** If `notes.txt` exists and begins with
neither `SBEX` nor `#!`, typing `notes.txt` reports `that is not a program or a script`
rather than `I do not know: notes.txt`. Reporting an unknown command about a file that is
sitting on the disk would send somebody looking in the wrong place.
Names are matched exactly, including case, because every other name on the filesystem is.
What limits the typed word is the buffer it is built in rather than the format: each name
@@ -688,7 +710,7 @@ laid out in three directories:
| Where | What |
| -- | -- |
| `/Apps` | The programs. The second place the shell looks for a word it does not recognise, so anything here starts by name from anywhere on the disk. |
| `/Apps` | What you run. The second place the shell looks for a word it does not recognise, so anything here starts by name from anywhere on the disk - programs, and scripts too. |
| `/Source` | The things you name to the assembler: CosmOS itself, the assembler itself, and small programs to read. |
| `/Lib` | The things those include. Everything here is named by an `#Include` somewhere and by nothing else, which is what makes it a library rather than a source. |
+111 -86
View File
@@ -1744,12 +1744,11 @@ tabIsFirst:
; The shell's own commands, only at the front of a line and only when no directory has been
; named: nothing with a separator in it was ever going to be "dir".
;
; And on the disk: programs at the front of a line, since that is all a first word can
; usefully be, and anything at all after it.
; And on the disk: everything, wherever in the line it is. A first word used to be
; narrowed to programs, back when a program was a file ending ".sbx"; now that the shell
; runs what it reads, a first word can be any file at all.
SETD.1 TabFirst
LDA.1
SETD.1 TabPrograms
STA.1
CALL tabRunSources
SETD.1 TabCount
@@ -2261,22 +2260,15 @@ tabEntry:
INCD.3
RSTA
STA.3
BRI tabEntryOffer
tabEntryFile:
; A file. When the first word of a line is being finished it can only usefully be a
; program, and the name to type for one is the name without its extension - so anything
; else is not offered at all rather than offered and useless.
SETD.1 TabPrograms
LDA.1
BRA tabEntryOffer
CALL tabStripSbx
BNQ tabEntryDone
tabEntryOffer:
; A file, offered under the name it actually has - so nothing happens to it here at all.
; A first word used to be offered with ".sbx" taken off, and anything without that ending
; not offered, which was right while the extension was what made a file reachable by name.
; THE SHELL NOW RUNS WHAT IT READS, so a script and a launcher start by name too, and
; hiding them would hide exactly the things a first word is most likely to be.
SETD.3 TabCandidate
CALL tabOffer
tabEntryDone:
RET
; DP0 names a string and DP1 where to put it. Copies up to what a name can be.
@@ -2308,62 +2300,6 @@ tabEnd:
tabEndDone:
RET
; Takes ".sbx" off the end of TabCandidate. Q is zero if it was there, and one if it was not
; - which is how a file that is not a program is passed over.
tabStripSbx:
; How long the name is, counted rather than measured from a pointer, because what is
; wanted next is a NUMBER: four characters back from the end is only a place if there are
; four of them to go back over.
SETD.0 TabCandidate
RSTB
tabStripCount:
LDA.0
BRA tabStripCounted
INCD.0
INCB
BRI tabStripCount
tabStripCounted:
; The count is in B and wanted in A, and there is no instruction that moves one to the
; other: adding it to nothing is how a value crosses from B to A on this machine.
RSTA
CCF
ADD
MVQA
INIB 0d4
CCF
SUB
BRC tabStripNo ; Shorter than the extension, so it cannot end in it.
MVQA
SETD.3 TabCandidate
DPUA.3 ; Four from the end, which is where the extension would begin.
SETD.0 SbxText
tabStripMatch:
LDA.0
BRA tabStripYes
LDB.3
XOR
BNQ tabStripNo
INCD.0
INCD.3
BRI tabStripMatch
tabStripYes:
; It ended in the extension, so the name to type is what comes before it. DP3 was left on
; the zero by the comparison, so four back is where the dot was.
INIA 0d4
DPDA.3
RSTA
STA.3
RSTB
CCF
ADD ; Q is zero: it is a program.
RET
tabStripNo:
INIA 0x01
RSTB
CCF
ADD ; Q is one: it is not.
RET
; ---- The shell's own words ----
;
@@ -3501,10 +3437,25 @@ promptUnknown:
STA.0
promptSearch:
; ---- The name as typed, and only then the name with ".sbx" on it ----
;
; The exact word first is what makes a file reachable under the name it actually has: a
; script, a launcher, or a program somebody dropped the extension from. The word with the
; extension after it is what keeps every program already on a disk startable by the short
; name people have always typed for it.
;
; THE SUFFIX CAN ONLY EVER BE A SECOND GUESS. A file that is really there always wins over
; one that would have to be invented, so "notes.txt" stops sending the shell after a
; notes.txt.sbx that was never going to exist.
RSTA
SETD.0 NameSuffix
STA.0
promptSearchName:
CALL nameProgram
SETD.0 NameOk
LDA.0
BRA promptElsewhere ; Too long to be a path, so it is not the name of one.
BRA promptNextName ; Too long to be a path, so it is not the name of one.
CALL loadProgram
SETD.0 LoadStatus
@@ -3526,6 +3477,18 @@ promptNotLoaded:
SETD.0 LoadStatus
LDA.0
; ---- Not a program is not the end of it ----
;
; The file is there and it does not say SBEX. A script says what it is on its own first
; line, so the thing to do is read that rather than decide from the name that this could
; never have been runnable. THE SHELL RUNS WHAT IT READS, which is the whole point: the
; loader turns away anything that is not SBEX and scriptOpen turns away anything without
; the shebang, and between them nothing has to know what an extension means.
INIB 0d4
CCF
SUB
BRQ promptAsScript
; No file of that name is not a fault. It is the ordinary case of a word this shell does
; not know, and it is the only answer worth looking somewhere else for. Anything else
; means a file of that name IS there and something is wrong with it, and answering
@@ -3536,6 +3499,52 @@ promptNotLoaded:
SUB
BNQ loadFailed
promptNextName:
; Nothing of that name in this place. The same place with the extension on, and then the
; next place.
SETD.0 NameSuffix
LDA.0
BNA promptElsewhere
INCA
STA.0
BRI promptSearchName
promptAsScript:
; ---- The drive goes back before a script opens ----
;
; The same handing back that a loaded program gets, and here it has to happen BEFORE the
; file is opened rather than after: a program is wholly in memory by then and a script is
; read a block at a time as it runs, so the lines in it must run on the disk the person was
; standing on. The name itself still says where the script lives, and reading it does not
; move anybody - see scriptKeepDrive.
SETD.0 SearchDrive
LDA.0
CALL sbfsUse
SETD.0 ProgramName
CALL scriptOpen
BRQ prompt ; It is open, and the next line read will come from it.
; Which of the two went wrong. There is a file, so "there is no such file" cannot happen.
MVQA
INIB 0x02
CCF
SUB
BRQ promptNotRunnable
SETD.0 ScriptTooDeep
BRI fileComplain
promptNotRunnable:
; A file of that name IS there, and it is neither kind of runnable thing. Saying "I do not
; know" about it would send somebody looking for a file that is sitting right in front of
; them; what is wrong is the file, not the word.
SETD.0 NotRunnable
CALL printString
SETD.0 CommandLine
CALL printString
CALL newLine
BRI commandFailed
promptElsewhere:
; Was that the last place there is?
SETD.0 NamePrefix
@@ -3814,13 +3823,18 @@ widthDone:
; ---- Making a file name out of a typed word ----
;
; Turns what was typed into the name of a file to go and look for. ".sbx" goes on the end
; unless it is already there, so that "Snake" and "Snake.sbx" both find the same file.
; Turns what was typed into the name of a file to go and look for. Which place to look in is
; NamePrefix, and whether to put ".sbx" on the end is NameSuffix - the search calls this once
; for each combination and the two answers are two different files to go and ask about.
;
; THE EXTENSION IS WHAT MAKES A FILE REACHABLE BY NAME. Typing "notes" looks for notes.sbx
; and typing "notes.txt" looks for notes.txt.sbx, so a text file cannot be started by
; typing what it is called, whatever is inside it. Only load reaches a file by its whole
; name, which is why the monitor can still put any file at all in front of itself.
; THE EXTENSION USED TO BE WHAT MADE A FILE REACHABLE BY NAME. It went on unconditionally, so
; typing "notes" looked for notes.sbx and typing "notes.txt" looked for notes.txt.sbx, and no
; text file, script or launcher could be started by typing what it is called. Now the word as
; typed is tried first and the extension is only a fallback for the programs that carry one.
;
; A word that already ends in ".sbx" is refused on the suffix pass rather than named twice:
; the exact pass has been there and looked, and asking the disk the same question again is a
; walk of a directory for an answer already known.
;
; NameOk is one if there is a path in ProgramName and zero if the word could not be made
; into one. What limits it is the buffer, not the format: each NAME along a path is still
@@ -3879,8 +3893,16 @@ nameCopied:
SETD.0 NameLeft
STB.0
; The word exactly as typed, which is the first thing asked for and the only thing asked
; for when the extension is not wanted.
SETD.0 NameSuffix
LDA.0
BRA nameMade
; Only a word of four characters or more can already end in ".sbx". Stepping back four to
; look at a shorter one would read whatever happens to sit in front of the buffer.
SETD.0 NameLeft
LDB.0
INIA 0d55
CCF
SUB
@@ -3892,7 +3914,7 @@ nameCopied:
SETD.0 SbxSuffix
nameSuffixSame:
LDA.0
BRA nameMade ; The suffix ran out with all of it matched.
BRA nameDone ; All of it matched, so the exact pass already asked for this.
LDB.2
CCF
SUB
@@ -7513,6 +7535,10 @@ ScriptStopped:
"stopped: that line did not work"
Unknown:
"I do not know: "
; A file of that name is there and it is neither a program nor a script, which is a different
; thing from a word this shell does not know and reads differently to whoever typed it.
NotRunnable:
"that is not a program or a script: "
Farewell:
"halted"
DirectoryText:
@@ -7594,8 +7620,8 @@ DataWord:
ExecMagic:
"SBEX"
; What every program on the disk is called, and the four characters that make one
; reachable by typing its name.
; What every program on the disk is called. It is a convention now rather than a rule: the
; shell tries the word as typed first and only puts these four characters on afterwards.
SbxSuffix:
".sbx"
LoadWhat:
@@ -7781,8 +7807,6 @@ ForName:
ShellNameCount:
0d22
SbxText:
".sbx"
AppsPath:
"/Apps"
VarNoSuch:
@@ -7864,6 +7888,9 @@ NameOk:
0x00
NamePrefix:
0x00
; Nought for the word exactly as typed, one for the word with ".sbx" put on the end.
NameSuffix:
0x00
NameLeft:
0x00
LoadedEntry:
@@ -8206,8 +8233,6 @@ TabCopyAt:
0x00
TabFirst:
0x00
TabPrograms:
0x00
; Whether the walk is looking for the answer or showing what the answers were.
TabShowing:
0x00
+16 -4
View File
@@ -11,15 +11,27 @@ finished
> dir
Say.sbx 156
dir.sbx 156
Greet 156
notes.txt 21
notes.sbx 21
4 files
hello.sh 38
Apps <dir>
6 files, 1 directory
> Greet a program with no extension
it says: a program with no extension
finished
> hello.sh
> echo a script, started by its name
a script, started by its name
> away.sh
> echo a script from the system place
a script from the system place
> notes
not a program
that is not a program or a script: notes
> notes.sbx
not a program
that is not a program or a script: notes.sbx
> notes.txt
I do not know: notes.txt
that is not a program or a script: notes.txt
> nosuchprogram
I do not know: nosuchprogram
> abcdefghijklmnopqr
+3 -3
View File
@@ -1,9 +1,9 @@
CosmOS
> CrCrash zz
> CrCrash.sbx zz
Crash opcode | service | bank | device | blind
finished
> SaSayzz
I do not know: Sayzz
> SaSay.sbxzz
I do not know: Say.sbxzz
> cd Apcd Apps/
/Apps> echo where am I
where am I
+3
View File
@@ -2,6 +2,9 @@ Say hello there
Say.sbx spelled out in full
run once more
dir
Greet a program with no extension
hello.sh
away.sh
notes
notes.sbx
notes.txt
+16 -3
View File
@@ -487,8 +487,8 @@ printf 'and these are the very different notes in B\n' > notesB.txt
"$ROOT/Programs/CosmOS/Apps/Wander.asm" -o "$WORK/Wander.sbx" >/dev/null
"$TOOL" put "$DISKS/cwd.img" "$WORK/Wander.sbx" /Apps/Wander.sbx >/dev/null
# A disk for invoking a program by typing its name. Four files, each there to say one
# thing about how a typed word turns into a file name.
# A disk for invoking a program by typing its name. Every file on it is there to say one
# thing about how a typed word turns into a file that runs.
#
# dir.sbx is a working program under the name of a built-in command, which is the only way
# to check that the built-ins really are tried first. If the search ever moved ahead of the
@@ -496,15 +496,28 @@ printf 'and these are the very different notes in B\n' > notesB.txt
#
# notes.sbx is text under a program's name, so that a file that IS found and IS NOT a
# program can be told apart from a word that names nothing. notes.txt is the same text
# under its own name, which no typed word can reach: "notes.txt" looks for notes.txt.sbx.
# under its own name - which the shell could not reach at all until it began trying the
# word as typed, because "notes.txt" only ever looked for notes.txt.sbx.
#
# Greet is a program with NO extension, and hello.sh is a script. Neither could be started
# by typing its name before the shell decided what to run by reading it, and between them
# they are the whole of that claim: SBEX loads, #! is read as lines, and the name says
# nothing about which happened. /Apps/away.sh is the same for a script found somewhere
# other than where you are standing.
"$TOOL" format "$DISKS/invoke.img" 64 2 >/dev/null
"$ROOT/Assembler" -I "$ROOT/Programs/CosmOS/Source" \
"$ROOT/Programs/CosmOS/Apps/Say.asm" -o "$WORK/Say.sbx" >/dev/null
"$TOOL" put "$DISKS/invoke.img" "$WORK/Say.sbx" >/dev/null
"$TOOL" put "$DISKS/invoke.img" "$WORK/Say.sbx" dir.sbx >/dev/null
"$TOOL" put "$DISKS/invoke.img" "$WORK/Say.sbx" Greet >/dev/null
printf 'this is not a program' > notes.txt
"$TOOL" put "$DISKS/invoke.img" notes.txt >/dev/null
"$TOOL" put "$DISKS/invoke.img" notes.txt notes.sbx >/dev/null
printf '#!\necho a script, started by its name\n' > "$WORK/hello.sh"
"$TOOL" put "$DISKS/invoke.img" "$WORK/hello.sh" hello.sh >/dev/null
"$TOOL" mkdir "$DISKS/invoke.img" /Apps >/dev/null
printf '#!\necho a script from the system place\n' > "$WORK/away.sh"
"$TOOL" put "$DISKS/invoke.img" "$WORK/away.sh" /Apps/away.sh >/dev/null
# A disk for the assembler that runs on the machine. It holds a source file and the two
# programs that check the front end - the reader on its own and the tokenizer on its own -
+9 -2
View File
@@ -1690,8 +1690,15 @@ print(sum(1 for o in range(0, len(px), 3) if px[o:o + 3] == moon))
# What is checked is that it is HIGHER LATER: two captures, the second further on, and the
# lander nearer the top in the second. A thruster that only fired once for a held button
# would let it fall between them instead, which is exactly what the console does.
#
# ONE AND 1.8 MILLION, and they were 800 thousand and 1.6 million. The shell tries a typed
# word as the name of a file before it tries it with ".sbx" on the end, so starting Lander
# now walks four names where it walked two, and at 800 thousand the program had not drawn
# yet - which reads as row -1 and fails, correctly, for a reason that has nothing to do with
# thrusters. The gap between the two captures is what the check is made of, and it is the
# same gap.
python3 -c "open('$BUILD/lander.pad','wb').write(b'\x00' * 40 + b'\x08' * 400)"
for when in 800000 1600000; do
for when in 1000000 1800000; do
timeout 30 "$EMU" --fast --cycles $when --keyboard "$BUILD/lander.keys" \
--pad "$BUILD/lander.pad" --screen "$BUILD/held$when.ppm" \
--disk "$ROOT/Tests/build/disks/cosmos.img" --ram-disk 2048 \
@@ -1705,7 +1712,7 @@ def top(path):
ship = bytes.fromhex('d8c048')
ys = [i // 320 for i in range(len(px) // 3) if px[i * 3:i * 3 + 3] == ship]
return min(ys) if ys else -1
print(top('$BUILD/held800000.ppm'), top('$BUILD/held1600000.ppm'))
print(top('$BUILD/held1000000.ppm'), top('$BUILD/held1800000.ppm'))
" 2>/dev/null || echo "-1 -1")
EOT
[ "$EARLY" -gt 0 ] && [ "$LATE" -gt 0 ] && [ "$LATE" -lt "$EARLY" ] \