TP3-comp

Recreating Turbo Pascal 3.0 (shell, editor, compiler, interpreter) in GNU Modula-2 (-fiso). Faithful to the original: the screen layout, keys and behaviours are reconstructed from the disassembled TP3.0 source (Resources/) and the reference manual.

State

Where things are: SUMMARY.md. It holds the milestone table, the list of checks and what each one cannot see, and the numbers behind every claim. The short version:

• The shell, WordStar-style editor, compiler, linker and in-process 8086 interpreter all build and run.
• The compiler is split into focused modules: CompilerShared, CompilerUtil, CompilerSource, CompilerSymbols, CompilerLex, CompilerExpr, CompilerIO, CompilerEmit, CompilerStmt (statements), CompilerDecl (declarations) and CompilerCode (the address-forming and small emitters the parsers call); Compiler.mod is the driver. The runtime is split the same way: RuntimeEmit (state and instruction emitters), RuntimeText, RuntimeReal, RuntimeFile, RuntimeData and a small Runtime.mod driver.
• The surviving TP3 corpus compiles 71/71; CALC.PAS compiles to its final .COM and runs through the welcome screen, grid, menu and a formula.
• Emitted .COM images run under both qemu-system-i386 and Exec86; the shell's R key now runs the linked image in-process.
• Current measured runtime: 11,476 bytes / 86 entries. CALC: 39,359-byte .COM (code 25,945 + data 19,694), and its full Help path allocates and reads the 16,281-byte Pages array, then pages through it with N/P/Esc.
• Data and heap live in a separate DS segment (DataStart = 5000h), as in the original TP3; the .COM is code plus a small initialised-data image.
• A call takes up to 64 arguments (MaxArgs); t168_manyargs runs 20 and t169_arglimit pins the 65th. A malformed program-header list such as program p(1;) is an error, not an infinite loop.
• Latest code checkpoint: v-TP3-ABS-INLINE; conformance details are in TP3-CONFORMANCE.md.

The remaining work is TP3-conformance and verification, not basic compilation: interrupt services beyond the ones the harnesses answer, heap coalescing, and the accepted no-op runtime/editor options. SUMMARY.md lists these in priority order.

Build

Requires GNU Modula-2 (gm2, tested with 16.0.1) and a C compiler for the FOR "C" bridge.

cd shell
make
./tpshell

The shell runs in raw mode: it needs a real ANSI terminal, no wrappers.

To run the checks: cd shell && bash tests/run_all.sh. It is not part of make, because it needs qemu and takes a few minutes. DOS file tests use dosbox-x when available and skip cleanly when it is not. The mutation suite is bash tests/nonvacuity.sh.

New here? Start with TUTORIAL.md. It walks through the build, the shell, editing and compiling a program, running the emitted image, and the test harnesses.

CALC launcher

tests/calc_cli.py builds CALC.COM in a temporary include directory, boots it under qemu and prints its serial output:

cd shell
python3 tests/calc_cli.py                  # build + default welcome/quit keys
python3 tests/calc_cli.py --com CALC.PAS.COM --keys '\r\r/Q'
python3 tests/calc_cli.py --keys-file keys.bin --raw > screen.bin

Options: --keys (Python escapes such as \r, \n, \x1b), --keys-file, --com, --timeout, --raw, --quiet, --verbose, and --no-keypress (make AH=0Bh report no queued key; useful for scripted CALC input). Input is a static boot-time key stream, not a live terminal. The automated smoke check is python3 tests/calc_smoke.py, run by tests/run_all.sh. For a manual FreeDOS/qemu setup, see FREEDOS-CALC.md; CALC is verified under the FreeDOS 1.4 LiveCD.

Layout

• shell/Term.def|mod — raw termios keyboard + ANSI screen primitives (blocking GetCh, non-blocking Avail/GetKey via poll, GotoXY, highlight Marked/Normal, Beep).
• shell/Posix.def|Posix.c — DEFINITION MODULE FOR "C" bridge to libc: read/write/open/close/unlink/rename/getcwd/chdir/opendir/readdir/closedir/ statvfs/poll and termios (tcgetattr/tcsetattr/cfmakeraw).
• shell/TextBuf.def|mod — the shared TP3-style text buffer: a 62903-char array, lines terminated by CR. Text is kept in memory so the editor and the shell's S (save) operate on the same buffer.
• shell/Editor.def|mod — WordStar-style full-screen editor.
• shell/Shell.mod — TP3.0 main-menu dispatcher (kmenu/kcmdtab), work-file load/save with .BAK rotation, directory listing, options menu.
• shell/Compiler.def|mod — the compiler's public interface and the statement/declaration/driver core.
• shell/CompilerShared.def — shared compiler state, constants and types.
• shell/CompilerUtil.def|mod — pure character/word/bit/string helpers.
• shell/CompilerSource.def|mod — source cursor and {$I} expansion.
• shell/CompilerSymbols.def|mod — symbol table, lookup, insertion and scope hiding.
• shell/CompilerLex.def|mod — keyword table, scanning, constants and the string-literal pool.
• shell/CompilerExpr.def|mod — expression parser and value construction.
• shell/CompilerIO.def|mod — standard-procedure dispatch.
• shell/CompilerEmit.def|mod — primitive 8086 emitters and patch machinery. Compiler.CodeByteAt exists so a test can read the emitted image.
• shell/Runtime.def|mod — the 8086 runtime library, built as a byte blob at run time and copied to the front of every image. Its entries are addressed by name from Runtime.def; the emitter procedures are audited name-against-decode.
• shell/Linker.def|mod — assembles [JMP rel16][runtime][header][program] and writes a real DOS .COM when the Destination option is COM.
• shell/Makefile — the authoritative build recipe; two-phase link, see SUMMARY.md.
• shell/tests/ — the harnesses. run_all.sh runs them, nonvacuity.sh proves each one can go red. tests/fixtures/expected.tsv is the live fixture matrix and is asserted against, not copied into prose.
• SUMMARY.md — the state document: milestones, checks, numbers, and what is still missing. It also carries the prioritized next milestones.
• TUTORIAL.md — a hands-on walkthrough from build to compile/run and tests.
• TP3-CONFORMANCE.md — directive, type and library conformance matrix.
• TP3-COMPILER.md — how the compiler works, the build recipe, and the codegen bug classes with their fixes.
• TP3-EDITOR.md — the editor's key handling and screen layout.
• RESUME-TP3.md — book-derived reference for the 8086 code the TP3 compiler generates (source of the compiler phase).
• Resources/ — OCR of "Au coeur de Turbo Pascal", TP3.0 manual and the disassembled TPSRC1-10. The disassembly is the ground truth: when our behaviour and the book disagree, the disassembly wins.

Shell (TP3.0 main menu)

Exact TP3.0 screen: Logged drive, Active directory, Work file, Main file, Edit Compile Run Save / Dir Quit compiler Options, Text: n bytes, Free: n bytes, > prompt. Command letters are drawn bold (TP3 uses high-bit characters).

Key	Action
L	new drive / directory
A	new active directory
W	set work file (auto-.PAS), Loading/New File
M	set main file
E	edit work file
C	compile (wired: TP3 error report + editor jumps to the error)
R	run the linked image in-process through Exec86 and report its exit status
S	save with ^Z EOF; old version → .BAK
D	directory listing, DOS *.* mask, k bytes free
O	options submenu (compile target, Code/Data/Stack, params)
Q	quit (prompts to save when text changed)

Any other key redraws the menu, like TP3.

Editor

WordStar-style full-screen editing over TextBuf. Status line: Line n Col n Insert/Overwrite Indent X:FILENAME. Text rows are screen rows 2–24; row 1 carries status/prompt/messages.

Highlights:

• Movement — ^S/^D/^E/^X/^A/^F/^R/^C/^W/^Z and ^Q S/D/E/X/R/C/B/K/P, arrow keys / PgUp / PgDn / Home / End (ANSI escape sequences).
• Editing — insert/overwrite (^V), ^G delete char, backspace joins lines, ^T word, ^Y line, ^Q-Y to end of line, ^N / CR line break, TAB auto-indent to word start on the line above.
• Block — ^K B/K mark, ^K T word, ^K H show/hide, ^K C/V/Y copy/move/delete, ^K R/W read/write block to/from a file.
• Search — ^Q-F find, ^Q-A replace, ^L repeat last; options B/G/n/U/W, and N (no-confirm) for replace; ^A matches any char, CR LF searches a line break.
• ^P prefix inserts a literal control character; ^U / ESC abort prompts.
• ^K-D returns to the shell; the text stays in memory, changed is reported and S (or the quit prompt) writes it to disk.

Disk format matches TP3: CRLF line endings and a trailing ^Z; load normalizes CRLF/LF, save re-expands.

Porting notes (gm2)

• Foreign modules (DEFINITION MODULE FOR "C") must use ADDRESS, not Modula-2 record/pointer types — cast at the call site.
• gm2 ICEs on FOR ... EXIT; use WHILE+flag (or LOOP+EXIT).
• gm2 only emits main when the program module source is on the link line: gm2 -fiso -o tpshell Shell.mod Term.o Posix.o TextBuf.o Editor.o.
• termios is used in raw mode, so every newline is explicit CR LF.
• EXIT is only legal inside LOOP (ISO Modula-2), not WHILE.