src_load reads a SEQ file from device 8 once into a bank-1 cache (D118).
It enforces the 80-byte line, 599-line, and 18,999-byte limits as bytes
arrive and reports stable codes with exact positions. src_rewind and
src_next_line replay the cache one logical line at a time for every pass.
diskio.s wraps the KERNAL file calls; channels close on every path.
Verified in x128 with a 1581: eight fixtures, including the exact limits,
each limit error, a missing file (DOS 62), and a missing final $0D, all
match host-computed line counts, Fletcher-16 checksums, and positions.
Adds a shared PowerShell test library for x128 scripts.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Add cfg/compiler.cfg, the production mmu.s, c128.inc, and the test mailbox,
proven by a capacity-checking skeleton in x128:
- 8K bottom common RAM, shared code at $1300, private zero page at bank 0
page $1A via P0, bank-1 tables at $2000-$BFFF, test mailbox at $1BF0
- ld65 rejects overflows of the ZP, COMMON, MAIN, SWSTACK, and TABLES areas
- a CIA #2 timer NMI taken under bank 1 with the private zero page returns
cleanly through the NMI guard (closes the E01-S03 NMI risk)
- decision D117 records the map; RESTORE is ignored while compiling
Also pin newLang/ to LF in .gitattributes. A branch switch under
core.autocrlf had rewritten the byte-exact fixtures with CRLF.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
vim-user-guide.pdf has no NUL bytes, so git auto-detected it as text and
core.autocrlf rewrote its line endings on checkout, corrupting it. Mark
PDF, PNG, PRG, BIN, and D64/D71/D81 files binary so they are never converted.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Nova128 is an object-oriented language with an on-machine C128 compiler.
This adds the spec, decisions, architecture, backlog, and story claims,
plus the grammar fixtures (E00-S02), disk-image builder (E07-S01), and
the x128-verified platform probes:
- E01-S01 minimal native PRG
- E01-S02 MMU/common-RAM probe. With bottom common on, a P0/P1 pointer
aimed at bank 1 lands in bank 0 at the translated page. The verifier
traces every MMU write in the probe window.
- E01-S03 KERNAL wrapper (kcall). CHROUT, drive status, and a
device-not-present failure work from bank 0 and bank 1 mappings.
A/X/Y/C/ST are preserved and the MMU state is restored exactly.
- E01-S04 40/80-column display and clock probe
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
NOTES records the dd/p, A/I/O/o, :q/:q! and missing-file checks that pass
on the work disk, and that \x03 in -keybuf arrives as RUN/STOP so a script
can leave insert mode.
Generated with Codebuff 🤖
Co-Authored-By: Codebuff <noreply@codebuff.com>
The single-row painter became vml.s at $C000: one sys() paints all 24
rows, memmove and line-index fix-ups run in ML, and the keyboard buffer
is popped with IRQs masked. dall() repaints only the cursor row or the
whole screen, so j went from 17 to 0-3 jiffies. Loading now checks the
drive error channel and splits lines over 80 columns. The temporary
jiffy counters used to measure the speedup are removed; the numbers are
in NOTES.md.
Generated with Codebuff 🤖
Co-Authored-By: Codebuff <noreply@codebuff.com>
Power C on the C64 calls ML as sys(addr, &a, &x, &y); the bank argument
from the manual's C128 library page made sys() jump to $000F and crash.
The ML row painter now runs. Quitting resets through $FCE2 because Power C
owns BASIC's zero-page pointers. Notes and Power C skill updated.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Daniel Covington
Daniel Covington