Files
ww/test/wcc/962_opaque_assign_cast_run.c
Hojun-Cho 3a0c7442d4 wcc: opaque assignability sink + reinterpret-cast verify (#108)
#108 sub-fold (c): opaque as a type-erasure sink. Two implicit
assignability rules + the reinterpret casts sort's impl relies on.

  rule 1  `*T -> *opaque`  IMPLICIT — any pointer is the universal
          void-pointer. harec type_is_assignable pointer arm
          (ref/harec/src/types.c:1053: `case STORAGE_OPAQUE: break;`
          — the referent need not match).
  rule 2  `[]T -> []opaque` IMPLICIT — any slice is the erased slice;
          {ptr,len,cap} header is normal, byte stride supplied at
          runtime. harec slice arm (types.c:1094).

Both fire only when the destination element is opaque, so they are
inert on the opaque-free selfhost corpus.

Rule-10 (per-rule, empirical): rules 1 & 2 are CSTAGE-ONLY. cstage
type_assignable gains the sink; the wwstage check.ww isassignable is a
resolve-only AST approximation that returns "can't tell, stay quiet"
(confident=false) for a ptr/slice whose element it cannot match, so it
already ACCEPTS every form (let-init AND call-arg). Verified: w6c_ww
compiles each probe source exit 0, byte-identically to w6c. cstage
rejected these before this change; no ww twin is needed (same
align-down precedent as 960/961's cstage-only arms).

Casts: N_CAST is validation-free in BOTH stages (the checker never
checks cast legality), so `[]opaque -> *u8` / `*opaque -> *u8`/`*i32`
are already legal. The reinterpret CGEN needed NO change: cgexpr
leaves the pointer in AX for both a slice (so slice->ptr naturally
takes .ptr) and a pointer (ptr->ptr is a no-op). drew described the
Hare idiom as `*[*]u8`; ww has no unbounded-array `[*]`, so the
ww-faithful reinterpret target is `*u8` + uintptr stride arithmetic.
cs==ww byte-id proven on every probe row.

Array->[]opaque (harec array->slice decay, types.c:1080-1099) is
deliberately EXCLUDED: ww has no implicit array->slice for any element
type (`let s: []i32 = a` is rejected too — a slice is built only via
an explicit `a[0:n]`), so there is no array->slice-header cgen.
Accepting array->[]opaque alone would assign a fat array local into a
24-byte slot with no decay: a silent miscompile (rule 7). sort's
caller passes a slice, so slice->[]opaque suffices.

opaque is unused by the bootstrap → INERT → 990-997 stay
byte-identical; combined.ww unchanged (no embedded source touched).
New probe 962_opaque_assign_cast_run carries both dimensions per row
(cstage build+run asserting type-erasure round-trips, AND a w6c-vs-
w6c_ww .s byte-id gate — the 990-997 gates never exercise opaque, so
the test pins rule-10 symmetry itself): rule1_implicit_ptr,
rule2_implicit_slice, and sort_pattern (byte-swap via uintptr stride
through []opaque, read back through the *opaque path and the original
[]i32 view). Probe binds call results before comparing to dodge a
pre-existing inline-call-result-in-comparison cgen bug (#116 family,
reproduces with zero opaque) — same dodge 960 uses.
2026-05-26 10:29:04 +09:00

283 lines
10 KiB
C

/*
* 962_opaque_assign_cast_run — runtime + byte-id proof of #108 sub-fold
* (c): opaque as a type-erasure sink. Two assignability rules + the
* reinterpret casts sort's implementation relies on.
*
* rule 1 `*T -> *opaque` IMPLICIT (no cast). Any pointer is the
* universal void-pointer. harec type_is_assignable, pointer
* arm: ref/harec/src/types.c:1053 (`case STORAGE_OPAQUE:
* break;` — the referent need not match).
* rule 2 `[]T -> []opaque` IMPLICIT (no cast). Any slice is the
* type-erased slice; the {ptr,len,cap} header is normal,
* the byte stride is supplied at runtime (itemsz). harec
* slice arm: types.c:1094 (`if (to_secondary->storage ==
* STORAGE_OPAQUE) return true;`).
* casts `[]opaque -> *u8` (slice -> byte ptr; cgexpr leaves the
* ptr in AX, so the cast naturally takes .ptr) and
* `*opaque -> *u8` / `*opaque -> *i32` (ptr->ptr reinterpret,
* a no-op). drew described the Hare idiom as `*[*]u8`; ww has
* no unbounded-array `[*]`, so the ww-faithful reinterpret
* target is `*u8` + uintptr stride arithmetic.
*
* Rule-10 placement (per-rule, empirical):
* rules 1 & 2 are CSTAGE-ONLY. cstage type_assignable (cmd/wcc/
* type.c) gained the opaque sink; the wwstage check.ww isassignable
* is a resolve-only AST approximation that returns "can't tell, stay
* quiet" (confident=false) for a ptr/slice whose element it cannot
* match, so it already ACCEPTS every form here (let-init AND call-
* arg). Verified empirically: w6c_ww compiles each row's source with
* exit 0, byte-identically to w6c (the cs==ww gate below). cstage
* rejected these before the type.c change. No ww twin is needed
* (same align-down precedent as 960/961's cstage-only arms).
* The casts are validation-free in BOTH stages (N_CAST never checks
* legality) and the reinterpret cgen needed no change — proven by
* the cs==ww byte-id gate.
*
* Array->[]opaque (harec's array->slice decay, types.c:1080-1099) is
* deliberately EXCLUDED: ww has no implicit array->slice conversion for
* any element type (`let s: []i32 = a` is rejected too — a slice is
* built only via an explicit `a[0:n]`), so there is no array->slice-
* header cgen. Accepting array->[]opaque alone would assign a fat
* array local into a 24-byte slot with no decay: a silent miscompile
* (rule 7). sort's caller passes a slice, so slice->[]opaque suffices.
*
* opaque is unused by the bootstrap, so the new rules fire only on
* opaque-typed operands — INERT on the selfhost corpus, 990-997 stay
* byte-identical. But that same inertness means the 990-997 gates
* never exercise opaque cs==ww; this test carries its own w6c-vs-w6c_ww
* byte-id gate (dimension (b)) to cover the rule-10 symmetry directly.
*
* Each row carries BOTH dimensions, like 953_f64crossmod_run:
* (a) cstage `ww build` + run, asserting the exit code — pins that
* the converged asm is runtime-correct (the type erasure round-
* trips: a value written/read through the opaque path reads back
* intact).
* (b) w6c vs w6c_ww `.s` cmp — FAILS if the stages diverge.
*
* NOTE — call results are bound to locals before any comparison, never
* compared inline (`if (f(x) != k)`). That inline-call-result-in-
* comparison shape is mis-compiled by a PRE-EXISTING cgen bug (#116
* family: reproduced with zero opaque — a fn-call result compared
* inline when its pointer arg was produced by a prior call doing
* uintptr arithmetic). Binding first is the same dodge 960 uses; it is
* NOT a workaround for the opaque feature, which is exercised in full.
*/
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include <unistd.h>
#include <sys/stat.h>
#include <sys/wait.h>
static int
runwait(const char *cmd)
{
int rc = system(cmd);
if (rc == -1) return -1;
if (WIFEXITED(rc)) return WEXITSTATUS(rc);
return -1;
}
static int
slurp_eq(const char *a, const char *b)
{
FILE *fa = fopen(a, "rb");
FILE *fb = fopen(b, "rb");
if (!fa || !fb) { if (fa) fclose(fa); if (fb) fclose(fb); return -1; }
int rc = 0;
for (;;) {
int ca = fgetc(fa);
int cb = fgetc(fb);
if (ca != cb) { rc = -1; break; }
if (ca == EOF) break;
}
fclose(fa); fclose(fb);
return rc;
}
struct row { const char *label; const char *src; int want_exit; };
static const struct row rows[] = {
/* rule 1: `*T -> *opaque` IMPLICIT (no cast) at a let-init AND a
* call-arg. Round-trip a real *i32 through *opaque and back, deref. */
{ "rule1_implicit_ptr",
"package main;\n"
"fn readi32(p: *opaque) i32 = { let pi: *i32 = p: *i32; return *pi; };\n"
"export fn main() i32 = {\n"
" let n: i32 = 42;\n"
" let po: *opaque = &n;\n" /* let-init *i32 -> *opaque */
" let v: i32 = readi32(po);\n"
" let v2: i32 = readi32(&n);\n" /* call-arg *i32 -> *opaque */
" if (v != 42) { return 1; };\n"
" if (v2 != 42) { return 2; };\n"
" return v;\n"
"};\n", 42 },
/* rule 2: `[]T -> []opaque` IMPLICIT (no cast) at a let-init AND a
* call-arg. Read .len, round-trip .ptr (a *opaque) back to *i32. */
{ "rule2_implicit_slice",
"package main;\n"
"fn slen(o: []opaque) i32 = { return o.len: i32; };\n"
"fn first(o: []opaque) i32 = { let p: *opaque = o.ptr; let pi: *i32 = p: *i32; return *pi; };\n"
"export fn main() i32 = {\n"
" let a: [4]i32 = [11, 22, 33, 44];\n"
" let s: []i32 = a[0:4];\n"
" let o: []opaque = s;\n" /* let-init []i32 -> []opaque */
" let n: i32 = slen(o);\n"
" let n2: i32 = slen(s);\n" /* call-arg []i32 -> []opaque */
" if (n != 4) { return 1; };\n"
" if (n2 != 4) { return 2; };\n"
" let f: i32 = first(s);\n"
" if (f != 11) { return 3; };\n"
" return n;\n"
"};\n", 4 },
/* sort's actual usage end-to-end: a `fn(items: []opaque, itemsz:
* size)` called with a []i32; inside, reinterpret the slice as a
* byte base (`items: *u8`), uintptr-arith two element addresses,
* byte-swap them by itemsz. Then read back through the *opaque
* element path and through the original []i32 view — the type
* erasure round-trips iff both agree. Also a *opaque arg straight
* from a *i32 (rule 1). exit 0 == every assertion held. */
{ "sort_pattern",
"package main;\n"
"fn elemptr(items: []opaque, i: size, itemsz: size) *opaque = {\n"
" let base: *u8 = items: *u8;\n" /* []opaque -> *u8 (takes .ptr) */
" let off: uintptr = (i * itemsz): uintptr;\n"
" return ((base: uintptr) + off): *opaque;\n"
"};\n"
"fn swap(items: []opaque, x: size, y: size, itemsz: size) void = {\n"
" let pa: *u8 = elemptr(items, x, itemsz): *u8;\n"
" let pb: *u8 = elemptr(items, y, itemsz): *u8;\n"
" let k: size = 0;\n"
" for (k < itemsz) {\n"
" let qa: *u8 = ((pa: uintptr) + k: uintptr): *u8;\n"
" let qb: *u8 = ((pb: uintptr) + k: uintptr): *u8;\n"
" let t: u8 = *qa;\n"
" *qa = *qb;\n"
" *qb = t;\n"
" k = k + 1;\n"
" };\n"
"};\n"
"fn readi32(p: *opaque) i32 = { let pi: *i32 = p: *i32; return *pi; };\n"
"export fn main() i32 = {\n"
" let a: [4]i32 = [10, 20, 30, 40];\n"
" let s: []i32 = a[0:4];\n"
" swap(s, 0: size, 3: size, size(i32));\n" /* []i32 -> []opaque call-arg */
" if (a[0] != 40) { return 1; };\n" /* erased swap round-trips */
" if (a[3] != 10) { return 2; };\n"
" let p0: *opaque = elemptr(s, 0: size, size(i32));\n"
" let v0: i32 = readi32(p0);\n"
" if (v0 != 40) { return 3; };\n"
" let nn: i32 = 77;\n"
" let pn: *opaque = &nn;\n" /* *i32 -> *opaque let-init */
" let vn: i32 = readi32(pn);\n"
" if (vn != 77) { return 4; };\n"
" let o: []opaque = s;\n"
" let ol: i32 = o.len: i32;\n"
" if (ol != 4) { return 5; };\n"
" return 0;\n"
"};\n", 0 },
{ NULL, NULL, 0 }
};
int
main(void)
{
const char *bin = getenv("BIN");
if (!bin) bin = "out/bin";
char absbin[1024];
if (bin[0] != '/') {
char cwd[1024];
if (getcwd(cwd, sizeof cwd) == NULL) return 1;
snprintf(absbin, sizeof absbin, "%s/%s", cwd, bin);
bin = absbin;
}
char w6c[1100], w6c_ww[1100];
snprintf(w6c, sizeof w6c, "%s/w6c", bin);
snprintf(w6c_ww, sizeof w6c_ww, "%s/w6c_ww", bin);
if (access(w6c_ww, X_OK) != 0) {
fprintf(stderr, "opaque_assign_cast: w6c_ww missing — cannot "
"run the cs==ww byte-id gate (the rule-10 proof)\n");
return 1;
}
int n = 0, fail = 0;
for (int i = 0; rows[i].src; i++, n++) {
char src[64];
snprintf(src, sizeof src, "/tmp/wwopqc_%d_%d.ww", getpid(), i);
FILE *f = fopen(src, "wb");
if (f == NULL) { fail++; continue; }
fputs(rows[i].src, f);
fclose(f);
/* (a) cstage build + run in a scratch dir. */
char tmpdir[64];
snprintf(tmpdir, sizeof tmpdir, "/tmp/wwopqc_%d_d_%d",
getpid(), i);
mkdir(tmpdir, 0755);
char cmd[2048];
snprintf(cmd, sizeof cmd, "cd %s && %s/ww build %s",
tmpdir, bin, src);
if (runwait(cmd) != 0) {
fprintf(stderr, "row[%s]: cstage build failed\n",
rows[i].label);
fail++;
unlink(src); rmdir(tmpdir);
continue;
}
char outbin[128];
const char *base = strrchr(src, '/');
base = base ? base + 1 : src;
snprintf(outbin, sizeof outbin, "%s/%s", tmpdir, base);
char *dot = strrchr(outbin, '.');
if (dot && strcmp(dot, ".ww") == 0) *dot = '\0';
int got = runwait(outbin);
if (got != rows[i].want_exit) {
fprintf(stderr, "row[%s]: cstage exit %d, want %d\n",
rows[i].label, got, rows[i].want_exit);
fail++;
}
unlink(outbin); rmdir(tmpdir);
/* (b) cs==ww byte-id gate: emit .s from both stages, cmp. */
char cs_s[64], ws_s[64];
snprintf(cs_s, sizeof cs_s, "/tmp/wwopqc_%d_%d_cs.s",
getpid(), i);
snprintf(ws_s, sizeof ws_s, "/tmp/wwopqc_%d_%d_ww.s",
getpid(), i);
snprintf(cmd, sizeof cmd, "%s -o %s %s 2>/dev/null",
w6c, cs_s, src);
if (runwait(cmd) != 0) {
fprintf(stderr, "row[%s]: w6c failed\n", rows[i].label);
fail++; unlink(src); continue;
}
snprintf(cmd, sizeof cmd, "%s -o %s %s 2>/dev/null",
w6c_ww, ws_s, src);
if (runwait(cmd) != 0) {
fprintf(stderr, "row[%s]: w6c_ww failed\n",
rows[i].label);
fail++; unlink(src); unlink(cs_s); continue;
}
if (slurp_eq(cs_s, ws_s) != 0) {
fprintf(stderr,
"row[%s]: cstage/wwstage .s DIFFER (rule-10 "
"byte-id violation)\n", rows[i].label);
fail++;
}
unlink(src); unlink(cs_s); unlink(ws_s);
}
if (fail) {
fprintf(stderr, "%d/%d opaque assign/cast tests failed\n",
fail, n);
return 1;
}
printf("opaque_assign_cast: %d/%d ok (cstage run + cs==ww byte-id)\n",
n, n);
return 0;
}