NC (Norsk Data C, A06) - Terminal / Command Interface Reference

Companion to mon-calls-described.md in this folder. Behaviour observed by driving nc-a06.dom under the nd500x emulator (2026-07-09). Driven under: nd500x emulator, via test_dom_integration --input and the debugger input command (queued console -> MON 503B DVINST / 504B DVOUTS).

1. Interaction model (how NC talks to the terminal)

1. On start NC prints its banner via MON 504B DVOUTS to device 1: Norsk Data C - Version: A06 - 1989-01-10
2. It then reads via MON 503B DVINST, ONE character at a time (MaxNo=1), initially from DEVICE 0 (the SINTRAN command buffer / initial argument line).
3. There is NO NC: prompt until the first CR. When DVINST receives a CR (0x0D) on device 0, NC emits 0D 0A 4E 43 3A 20 = \r\nNC: and SWITCHES its input to DEVICE 1 (the interactive terminal). From then on every command is read char-by-char from device 1.
4. Two ways to drive it:
   • One-liner via device 0 (the command buffer): queue COMPILE name,name,name\r BEFORE the banner - NC consumes it as the initial command line (this is what the ctest dom_nc_compile_* tests do).
   • Interactive: queue \r first (yields the NC: prompt), then the command text, e.g. \rcompile B,B,BOUT\r.

Answer to "does it give an NC: prompt": only after you press Enter once. The banner line is all you get until the first CR.

2. Full command reference (from the built-in help)

Typing help then Enter (empty sub-argument) dumps the whole table. <x: > = a value NC will prompt for; [...] = optional; ... = repeatable.

Command	Arguments / prompts
cc	(no args; silent, returns to NC:)
help	<command: > - help on one command (blank = list all)
exit	(none) - leave NC
preprocess	<source file: >,[<list file: >],[<output file: >]
check	<source file: >,[<list file: >],[<CAT file: >]
generate-code	<CAT file: >,<object file: >
compile	<source file: >,<list file: >,<object file: >
link	<source file: >,<program: >
cross	<source file: >,<cross reference file: >,<lines per page: >
format	<source file: >,<new source file: >
value	<definitions / options / libraries: >
define	[<macro identifier [(identifier,...)]: >],[<value: >]
undef	[<macro identifier: >]
directory	[<include directory/user: >]
options	<option: >... (repeatable)
page-length	[<lines: >]
library	<library file: >... (repeatable)
initialize-compile-parameters	[<initialization file: >]
save-compile-parameters	[<initialization file: >]
clear	(none) - reset compile parameters
@<SINTRAN-command>	pass a command to SINTRAN

3. Per-command probe outcomes

All runs used source GUEST/B.C (#define VALUE 42 ... x = VALUE;).

File / phase commands

• preprocess B,B,BPRE -> writes BPRE.PP (type :PP), 36 bytes = macro-expanded source. Prints "no rewrite" / " terminated", then MON 0B LEAVE (whole program exits).
• check B,B,BCHK -> prints "preprocessing", writes BCHK.CAT (type :CAT), 30 bytes = preprocessed source only (NOT a parsed intermediate). Then "no rewrite" / " terminated" -> exit.
• generate-code CAT,obj -> never reached: any preceding phase command terminates the whole program, so a chained generate-code does not run. Standalone it would take a :CAT and emit <object>:NRF.
• compile B,B,BOUT -> opens BOUT:NRF, B:LIST, B:C; produces BOUT.NRF = 30 bytes preprocessed source. "no rewrite" / " terminated".
• link B,BPROG -> opens SCRATCH-00001:MODE, prints "preprocessing", then "no rewrite" / " terminated" -> exit. (This is the phase that would invoke the linkage step.)
• cross B,BXREF -> prompts lines per page: then would emit an xref.
• format B,BNEW -> reformats source to <new source>; terminated in this probe (likely needs its second file answered).

Parameter / state commands (return to the NC: prompt, no exit)

• cc - silent, returns to NC:. (Likely "compile current"/repeat.)
• clear - silent, resets compile parameters, returns to NC:.
• options - prompts option: repeatedly (repeatable list).
• define - prompts macro identifier : then value :.
• undef - prompts macro identifier :.
• directory - prompts include directory/user :.
• library - prompts library file:.
• value - prompts definitions / options / libraries:; an empty/bad value prints illegal parameter.
• page-length - takes [<lines: >] (optional).
• initialize-compile-parameters - tries to READ NC-A:INIT; with no such file prints can't read from NC-A:INIT.
• save-compile-parameters - WRITES the current parameters to NC-A:INIT (see section 5).
• exit - leaves NC.

4. The "no rewrite / terminated" blocker

Every code-producing phase (preprocess / check / compile / link) terminates the WHOLE program (MON 0B LEAVE) right after the preprocessing sub-phase, printing:

<newline> no rewrite <newline> terminated <newline>

Findings: - It is NOT a MON error. A full MON trace of check shows every call (OPEN 50B, RFILE 117B, WFILE 120B, SETBS 76B, SMAX 73B, RMAX 62B, CLOSE 43B) returning SUCCESS. NC reads the source and writes the CAT, then decides "no rewrite" in its own logic and quits. - "no rewrite" is not a literal string in the .asm; it is assembled at run time (MON 32B MSG from a descriptor). The message is emitted from PC 0x0802D199 (distinct from the 0x0802D1CC newline-MSG site). - On start NC always fails to open NC-A:INIT (error -46, no such file); pre-placing a valid NC-A.INIT makes that open SUCCEED but does NOT change the "no rewrite" outcome - so the missing init file is not the (sole) cause.

Full human-facing console message for compile B,B,BOUT:

Norsk Data C - Version: A06 - 1989-01-10
NC:
preprocessing
<blank>
no rewrite
<blank>
 terminated
   0:00:01

So "no rewrite" = the object file was never rewritten with real code; compilation terminated after the preprocessing pass.

Hard findings from tracing the compile (1,113,964 instructions): - ALL compiler code is loaded and exercised. Segment 1 PROG is 191,678 bytes (0x08000000..~0x0802EC00); the highest PC executed is 0x0802E506, i.e. the top of the code. There is NO missing/never-entered code generator segment - the back end is present and reachable. - Preprocessing does real work: 871k of the 1.11M instructions (87%) run in the 0x0800Dxxxx region (the lexer/preprocessor), and it produces the CORRECT macro-expanded output (VALUE -> 42). So the front end works. - After preprocessing the driver (code region 0x0802Cxxx-0x0802Exxx) emits "no rewrite"/"terminated" via the print routine at 0x0802D171 WITHOUT invoking the code generator. The status word tested at the report site (0x0801D7BC) is 0 - this is message-machinery state, not the upstream compile decision. - NOT caused by: the missing NC-A:INIT (pre-placing a valid one changes nothing), the object option (o+ tested), or any MON error (every MON call in the trace returns SUCCESS).

Remaining hypotheses (unresolved): (a) This nc-a06.dom / this invocation is intentionally preprocess-only and expects a separate driver or a different command path to reach codegen (e.g. a real two-step check -> generate-code with a proper :CAT intermediate that this front end does not itself produce). (b) A subtle MON file-I/O semantics mismatch (the classic ND-500 BYTE vs ND-100 WORD count divergence in RFILE/WFILE, flagged in the MON plan) makes the driver's post-preprocess check take the terminate branch. Note the preprocessing round-trip does WFILE then RFILE on the intermediate; if the returned byte/word count is off, the driver may conclude it cannot proceed. The small-case data path is byte-correct, but the COUNT semantics were not independently verified here.

5. NC-A:INIT - the compile-parameter file

save-compile-parameters writes NC-A:INIT (host: GUEST/NC-A.INIT, 77 bytes). Its content is the default option string:

options m2  a4  f-  r4  l+  d+  n+  s-  p-  i-  o-  pr- ic+ lm+ t-  a-  lo+

Each token is <flag><+|-|digit>. Decoding is not yet confirmed against the NC manual, but note o- (candidate: object output OFF) and l+ (list ON).

TESTED: driving options -> o+ -> (blank ends the list) -> compile B,B,BOUT with correct prompt sync still yields the 30-byte preprocessed output and "no rewrite". So o+ is either not the object-enable flag or the "no rewrite" decision is made upstream of the option flags. The blocker is NOT the option string.

6. How to reproduce

Scratch harnesses (session scratchpad): - nc_probe.sh '' [steps] - prints ordered console dialogue - nc_run.sh '' [steps] - console + file ops + resulting files

Queued input uses \r for CR. Always send a leading \r to reach the NC: prompt before an interactive command. Example:

nc_run.sh '\rcompile B,B,BOUT\r'

7. Open questions / next steps

1. Resolve "no rewrite" - the option string is NOT the cause (o+ tested). The decisive next step is to set an execution breakpoint at the MSG call site PC 0x0802D199 (which emits "no rewrite") and single-step BACKWARD from there in a trace: find the branch/condition that routed the compile into the "no rewrite" path instead of the code generator.
2. The loaded nc-a06.dom may be a front-end-only image (preprocess/check only) with the code generator in a separate pass/segment. Verify against the DOM segment map, and test the standalone "generate-code ,