Add a '$EXTERN NAME(ARGTYPES) AS RET pragma so compiled BASIC can call C
functions directly, the natural follow-up to Level 1 (--emit-obj /
--main-name). The pragma is an apostrophe comment, so the interpreter
ignores it while the compiler registers it.
Map INTEGER/SINGLE/DOUBLE/STRING to int16_t/float/double/const char* at the
boundary: a string argument crosses as a temporary C copy that is freed
after the call, and a string return is copied into the pool. The call name
is matched case-insensitively but emitted as the C symbol with the case
written in the pragma. Names are recognized before parse_var() truncates
identifiers to two significant characters, so multi-character C function
names work.
A string return that aliases a char* argument is copied before the argument
temporaries are freed, which avoids a use-after-free. Over-supplied
arguments are consumed without desyncing the token stream and warn on arity
mismatch.
Docs: getting-started.md "Foreign Functions from BASIC". Test:
tests/run_ffi_test.sh, wired into CI. 63/63 compiler, 72/72 interpreter,
68/68 compat still pass.
Also refile the roadmap "Next Up" backlog as git-bug issues and prune
docs/roadmap.md to point at git-bug as the source of truth for planned work.
Co-authored-by: Eremey Valetov <evvaletov@users.noreply.github.com>
Four roadmap items:
- codegen: fix parenthesized string comparison. emit_atom didn't
consume the body of a string-literal token (`"`), so for
PRINT (A$+B$ < "ZZZ") it emitted a 0 placeholder, advanced one byte,
and left "ZZZ" to be reparsed as a variable + extra trailing tokens
-- the binary then failed to link with `var_ZZ_sng` undeclared.
emit_atom now skips to the closing quote. Separately, the
left_type tracking in emit_num_prec dropped VT_STR after a string +
string concat (becoming VT_SNG), so the string-comparison codepath
skipped when the relational operator arrived. Preserve VT_STR
through TOK_PLUS when both operands are strings. Verified: paren
string-cmp now compiles and produces the same -1 / 0 result as the
interpreter.
- compiler: --no-gc-check and --fast-math optimization flags.
--no-gc-check skips the per-line gwrt_check_line() (no string-pool
GC, no Ctrl+Break trap). --fast-math drops the divide-by-zero
guard on `/`; the divisor still goes through (double) so 10/0
produces inf rather than SIGFPE. Both threaded through
codegen_opts_t and exposed in --help. --inline-arrays from the
roadmap deferred -- larger refactor.
- interp: raise static caps on 32-bit / Linux builds. vars 256
-> 1024, arrays 64 -> 256, MAX_FOR_DEPTH 16 -> 64, MAX_GOSUB_DEPTH
24 -> 128, MAX_WHILE_DEPTH 16 -> 64. Codegen FOR_STACK_MAX 16
-> 64. Analysis-pass caps: MAX_LINES 4096 -> 8192, MAX_VARS 256
-> 1024, MAX_GOTOS 256 -> 1024, MAX_DATA 1024 -> 4096,
MAX_GOSUB_RET 256 -> 1024. 16-bit DOS keeps the original modest
caps via #ifdef _M_I86 -- the MEDIUM model has a single 64KB
DGROUP for all static data and the bumped sizes broke runtime
startup under DOSBox-X. 16-bit binary grew from 128KB to 132KB
from the offset_secs field plus DATE$/TIME$ shift code, well
within the FreeDOS budget.
- interp + codegen: DATE$ / TIME$ assignment via process-local
clock offset. Was a no-op accept-and-ignore. Now sets
gw.time_offset_secs (long), and DATE$ / TIME$ / TIMER readers
apply it to time(NULL) before formatting. The OS clock is
unaffected (would need root). Compiled-binary readers also
reference gw.time_offset_secs since libgwrt shares the gw
struct. Verified: PRINT DATE$; DATE$="12-31-1999"; PRINT DATE$
shows the expected before/after in both interpreter and AOT
paths.
After these changes: 72/72 interpreter tests, 68/68 compat, 63/63
compiler tests, DOS smoke under DOSBox-X all pass. Build clean on
both Linux (cmake) and 16-bit DOS (build_dos.sh 16).
