/*
 * 741_dotbase_chained — sentinel for #28 + #30. Pins wwstage's
 * `indexvaluetnode` to handle N_DOT base (graduates the helper #24
 * introduced for chained N_INDEX so the N_DOT base shape rides the
 * same path).
 *
 * Two sister latents, one helper graduation:
 *
 *   Read (#28): `obj.mat[i][k]` where obj is a struct with field
 *   `mat: **u8`. The outer N_INDEX's base is the inner N_INDEX,
 *   whose base is N_DOT. cgindex routes through `indexvaluetnode`
 *   for the inner base type; pre-fix `indexvaluetnode` only handled
 *   N_IDENT + N_INDEX bases, so the recursion bottomed out at the
 *   N_DOT base with bt=nil. esz fell through to 8 and signed_elem
 *   to false — wwstage emitted a stray outer `MOVQ $8, CX; IMULQ
 *   CX, AX` plus `MOVQ (AX), AX` (8-byte read over a 1-byte u8)
 *   instead of cstage's bare `MOVZBQ (AX), AX`.
 *
 *   Write (#30): `obj.arr[i] = v` where obj is a struct with field
 *   `arr: [N]Tagged` (e.g. (i64|str)). cgassign's N_DOT-base arm
 *   computed esz via `indexbaseesz` but never set `elemtn`, so the
 *   tagged-element store gate (keyed on elemtn) missed and the
 *   24-byte tagged slot was overwritten by a single MOVQ scalar
 *   store — wrong-width store + tag/payload junk in the upper
 *   16 bytes. The matching scanlocals pre-pass arm (cgendecl.ww)
 *   was also missing N_DOT base, so post-elemtn-fix the @tagscr
 *   slot allocated past the frame boundary and clobbered live
 *   locals (s, i).
 *
 * Cstage walks `n->lhs->type` directly via the typed AST (cmd/w6c/
 * cgen.c idx_eff + the N_INDEX-lhs N_ASSIGN branch). Wwstage
 * mirrors via `indexvaluetnode`, which #24 introduced for the
 * N_INDEX-base case and #28/#30 now graduate for N_DOT base via
 * `dotfieldtnode` lookup.
 *
 * Class A wwstage cgen UNDER. No in-tree consumer; sister latents
 * filed during #24 + #27 reviews (commits aa8ca47, 3ba1922). This
 * sentinel is the sole exerciser of the shape.
 *
 * Per row: in the `probe` body, assert the dispatch-defining MOV
 * mnemonic + scaling is present, plus cstage vs wwstage asm is
 * byte-identical.
 */
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include <unistd.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;
}

/* `needle` = a substring that must appear in the probe body and
 * uniquely identifies the post-fix dispatch (narrow MOV mnemonic
 * for read, 24B element scale for tagged write).
 * `antineedle` = a substring that MUST NOT appear — pre-fix wwstage
 * regression marker. Empty string skips the anti-check (write row
 * relies on cmp -s byte-id to catch divergence, since the tagged-
 * store byte-copy legitimately contains MOVQ-store fragments). */
struct row {
	const char *label;
	const char *src;
	const char *needle;
	const char *antineedle;
};

static const struct row rows[] = {
	/* Read: chained `obj.mat[i][k]` where mat: **u8. Post-fix wwstage
	 * routes the inner N_INDEX's N_DOT base through indexvaluetnode →
	 * dotfieldtnode, esz collapses to 1, outer load becomes MOVZBQ.
	 * Pre-fix had stray `MOVQ (AX), AX` (8-byte read over u8). */
	{ "dotbase_chained_read",
	  "type S = struct{ pad: i64, mat: **u8 };\n"
	  "fn probe(s: *S) u8 = {\n"
	  "    let i: i32 = 0;\n"
	  "    let k: u64 = 0u64;\n"
	  "    return s.mat[i][k];\n"
	  "};\n"
	  "export fn main() i32 = { return 0; };\n",
	  "\tMOVZBQ\t(AX), AX\n",
	  "" },
	/* Write: `obj.arr[i] = v` where arr: [N](i64|str) — tagged element.
	 * Post-fix wwstage: cgassign N_DOT arm sets elemtn → tagged-store
	 * path → IMULQ (stride) + byte-copy from scratch. Scanlocals N_DOT
	 * arm pre-reserves @tagscr in the frame. Pre-fix: scalar `MOVQ AX,
	 * (BX)` over the slot.
	 * #1/Phase 3: str IS []u8 (24B), so the (i64|str) slot is
	 * 8(tag)+24(str payload)=32B — stride is $32, not the 16B-world
	 * $24. Byte-identical across stages. */
	{ "dotbase_array_tagged_write",
	  "type T = (i64 | str);\n"
	  "type S = struct{ pad: i64, arr: [4]T };\n"
	  "fn probe(s: *S) void = {\n"
	  "    let i: i32 = 0;\n"
	  "    s.arr[i] = 42i64;\n"
	  "};\n"
	  "export fn main() i32 = { return 0; };\n",
	  "\tMOVQ\t$32, CX\n",
	  "" },
};

static int
slurp(const char *path, char *buf, size_t cap)
{
	FILE *f = fopen(path, "rb");
	if (!f) return -1;
	size_t n = fread(buf, 1, cap - 1, f);
	fclose(f);
	buf[n] = '\0';
	return (int)n;
}

static int
emit_s(const char *w6c, const struct row *r, int i, char *out_s, size_t cap)
{
	char src[96], cmd[1024];
	snprintf(src, sizeof src, "/tmp/dotbase_%d_%d.ww", getpid(), i);
	snprintf(out_s, cap, "/tmp/dotbase_%d_%d_%s.s",
	    getpid(), i, w6c[strlen(w6c) - 1] == 'w' ? "ww" : "c");

	FILE *f = fopen(src, "wb");
	if (!f) return -1;
	fputs(r->src, f);
	fclose(f);

	snprintf(cmd, sizeof cmd, "%s -o %s %s 2>/dev/null", w6c, out_s, src);
	int rc = runwait(cmd);
	unlink(src);
	return rc;
}

/* Inside `TEXT probe`, the needle must appear and the antineedle
 * must NOT appear (anti-regression on pre-fix wwstage emit). */
static int
check_probe(const char *spath, const struct row *r, const char *stage)
{
	char buf[1 << 14];
	if (slurp(spath, buf, sizeof buf) < 0) {
		fprintf(stderr, "row[%s][%s]: cannot read %s\n",
		    r->label, stage, spath);
		return -1;
	}
	const char *fn = strstr(buf, "TEXT probe");
	if (!fn) {
		fprintf(stderr,
		    "row[%s][%s]: no TEXT probe in %s\n",
		    r->label, stage, spath);
		return -1;
	}
	const char *ret = strstr(fn, "\tRET\n");
	const char *m = strstr(fn, r->needle);
	if (!m || (ret && m > ret)) {
		fprintf(stderr,
		    "row[%s][%s]: expected `%s` in probe body\n",
		    r->label, stage, r->needle);
		return -1;
	}
	if (r->antineedle[0] != '\0') {
		const char *bad = strstr(fn, r->antineedle);
		if (bad && (!ret || bad < ret)) {
			fprintf(stderr,
			    "row[%s][%s]: stray `%s` in probe body — "
			    "pre-fix wwstage regression\n",
			    r->label, stage, r->antineedle);
			return -1;
		};
	}
	return 0;
}

int
main(void)
{
	const char *bin = getenv("BIN");
	if (!bin) bin = "out/bin";
	char absbin[512];
	if (bin[0] != '/') {
		char cwd[256];
		if (getcwd(cwd, sizeof cwd) == NULL) return 1;
		snprintf(absbin, sizeof absbin, "%s/%s", cwd, bin);
		bin = absbin;
	}

	char w6c[640], w6c_ww[640];
	snprintf(w6c, sizeof w6c, "%s/w6c", bin);
	snprintf(w6c_ww, sizeof w6c_ww, "%s/w6c_ww", bin);

	int have_ww = (access(w6c_ww, X_OK) == 0);
	int n = (int)(sizeof rows / sizeof rows[0]);
	int total = 0, fail = 0;

	for (int i = 0; i < n; i++) {
		char cs_path[128], ws_path[128];

		if (emit_s(w6c, &rows[i], i, cs_path, sizeof cs_path) != 0) {
			fprintf(stderr,
			    "dotbase_chained[cstage][%s]: w6c failed\n",
			    rows[i].label);
			fail++; total++; continue;
		}
		total++;
		if (check_probe(cs_path, &rows[i], "cstage") != 0) fail++;

		if (!have_ww) { unlink(cs_path); continue; }

		if (emit_s(w6c_ww, &rows[i], i, ws_path, sizeof ws_path) != 0) {
			fprintf(stderr,
			    "dotbase_chained[wwstage][%s]: w6c_ww failed\n",
			    rows[i].label);
			fail++; total++;
			unlink(cs_path); continue;
		}
		total++;
		if (check_probe(ws_path, &rows[i], "wwstage") != 0) fail++;

		total++;
		char cmd[512];
		snprintf(cmd, sizeof cmd, "cmp -s %s %s", cs_path, ws_path);
		if (runwait(cmd) != 0) {
			fprintf(stderr,
			    "dotbase_chained[%s]: cstage vs wwstage asm differs\n",
			    rows[i].label);
			fail++;
		}

		unlink(cs_path); unlink(ws_path);
	}

	if (fail) {
		fprintf(stderr,
		    "dotbase_chained: %d/%d fixtures failed\n", fail, total);
		return 1;
	}
	printf("dotbase_chained: %d/%d ok\n", total, total);
	return 0;
}