Files
ww/test/wcc/809_delete_range.c
Hojun-Cho 1bcf2726cf wcc+w6c+w6c_ww: delete() range form delete(xs[lo:hi]) (fold-5a P2)
Hare's delete also takes a slicing place (harec check.c:1981-2027
EXPR_SLICE; Hare spells it delete(xs[i..j])): remove [lo, hi) — shift
[hi..len) down count = hi-lo strides, len -= count, cap unchanged; lo
defaults 0, hi defaults len, so delete(xs[:]) clears the slice with
storage retained. Checker accepts N_SLICE next to N_INDEX (object must
chase to a slice, harec :2024); the old range-unimplemented reject and
its #35 cite drop.

Lowering (both stages, converged byte-identical by construction) is the
single-element arm's same-slice whole-stride word-copy loop with a
DYNAMIC src offset (count*esz via a src register) instead of the
constant one-stride. Base shapes: local slice ident, deref-of-local,
plus NEW indexed local-slice base xs[g][lo:hi] — the fold-5a consumer
shape (regex.ha:333 delete(jump_idxs[group_level][..]); outer stride
off the type table). Bounds stay implicit, inheriting the documented
single-element posture (no index checks anywhere in cgen). Operands
evaluate left-to-right, exactly once, before the shift (harec order);
only the header ADDRESS is taken before operand eval, so a bound
expression's writes through the slice land before the copy.

test/809: 64 fixtures — full/explicit/re-clear/head/mid/tail/empty
a:a/end-boundary len:len/explicit 0:0 on a never-appended (nil-ptr)
slice, single-vs-range equivalence, cap preservation, esz 1/2/4/8/16/24
copy tails against the dynamic src, operand order-of-eval (lo/hi CALLs
fire once each, in order) + aliasing-visibility pins, the EXACT
[][]size regex consumer shape, deref base, 2 reject rows w/ diagnostic
text; every accept row cs==ww asm byte-id. test/804: reject_range row
retired (form now accepted), reject_nonindex text follows the widened
message.
2026-06-04 23:37:08 +09:00

608 lines
19 KiB
C

/*
* 809_delete_range — cstage and wwstage agree, byte-for-byte and at
* runtime, that the range form `delete(xs[lo:hi])` removes elements
* [lo, hi) from a slice: the tail [hi..len) shifts down count = hi-lo
* strides, len -= count, cap unchanged; lo defaults 0, hi defaults
* len (so delete(xs[:]) clears the slice, storage retained). The
* range half of the delete() builtin — fold-5a prereq P2 for the
* regex.ha:333 consumer `delete(jump_idxs[group_level][..])` (harec
* ref/harec/src/check.c:1981-2027 accepts EXPR_SLICE alongside the
* 804-covered ACCESS_INDEX form).
*
* Lowering (BOTH stages, converged byte-identical by construction):
* push &hdr, lo and count = hi-lo, then an ascending word-copy loop
* moves element j+count onto element j until j+count >= len, then
* hdr.len -= count. The per-element move is 804's same-slice
* whole-stride word copy with a DYNAMIC src offset (count*esz via a
* src register) instead of the constant one-stride; the esz spread
* below (1/2/4/8/16/24) re-proves every copy-tail arm against the
* dynamic src. Base shapes: local slice ident, deref-of-local
* ptr-to-slice, and — new for the range arm — an INDEXED local slice
* base xs[g][lo:hi] (the regex fold-5a consumer shape). Bounds are
* implicit (no range check), matching the 804 single-element arm and
* the rest of cgen — so no out-of-bounds abort rows here; a
* violating range is as undefined as a violating index.
*
* row | shape | want
* -----------------+----------------------------------------+------
* full_range | [7,11,13], delete(xs[:]) -> len 0 | 7
* full_explicit | [7,11,13], delete(xs[0:3]) -> len 0 | 8
* full_twice | delete(xs[:]) twice — second is a | 9
* | no-op on the emptied slice (count 0) |
* head | [1,2,3], delete(xs[:2]) — lo default 0 | 13
* mid | [1..5], delete(xs[1:3]) | 148
* tail_open | [1,2,3], delete(xs[1:]) — hi default | 11
* | len, copy loop never entered |
* empty_range | [1,2], delete(xs[1:1]) — count 0, | 23
* | self-copy loop, len unchanged |
* end_boundary | [1,2], delete(xs[len:len]) — lo==hi== | 212
* | len, loop exits at entry, no deref |
* | past the end |
* empty_explicit | delete(xs[0:0]) on a never-appended | 31
* | empty slice (ptr nil) — count 0 and |
* | len 0, nothing dereferenced |
* single_via_range | delete(xs[1:2]) == delete(ys[1]) | 42
* | element-for-element |
* cap_preserved | manual {ptr,len=3,cap=8} header, | 80
* | delete(xs[:]); cap must still read 8 |
* u8_narrow | []u8 esz=1 — MOVB tail, dynamic src | 215
* u16_narrow | []u16 esz=2 — MOVW tail, dynamic src | 25
* i32_narrow | []i32 esz=4 — MOVL tail, dynamic src | 50
* struct16 | []pair esz=16 — 2 whole qwords, no | 216
* | sub-word tail (804's struct esz twin) |
* str_elem | []str esz=24 — 3-qword headers move | 235
* | whole across a 2-stride shift |
* regex_exact | [][]size, delete(jj[lvl][:]) — the | 10
* | EXACT regex.ha:333 consumer shape |
* | (indexed base, size-typed index) |
* deref_shape | delete((*p)[:]) behind *[]u8 — the | 17
* | 804 regex_shape twin |
* order_of_eval | delete(xs[lof():hif()]) — lo then hi, | 16
* | each CALLED exactly once (trace must |
* | read 12; harec evaluates the slicing |
* | operands left-to-right, once) |
* alias_visible | delete(xs[poke(&xs):]) — lo's write | 91
* | through &xs lands BEFORE the shift |
* | (loop reads ptr/len/elems fresh from |
* | the header, after operand eval) |
* reject_array | delete(t[0:2]) on [3]i64 — "delete | BUILD_FAIL
* | must operate on a slice" |
* reject_str | delete(s[:]) on str — slicing a str | BUILD_FAIL
* | yields str, not a slice |
*
* BUILD_FAIL rows also assert the diagnostic TEXT (stderr substring,
* both stages) — a build that fails for any other reason (parse error,
* crash) is a vacuous reject and fails the row.
*
* Every non-BUILD_FAIL row also asserts cstage/wwstage asm byte-id,
* which subsumes the frame canary (TEXT main,$N) and the rdl_l/rdl_e
* label-counter symmetry.
*/
#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;
}
/* want == BUILD_FAIL: the row must FAIL to build on both stages AND
* emit expect_err on stderr (the checker reject set — message text
* included — is part of the contract: rule 7, never a silent
* acceptance; without the message check a row would pass vacuously on
* any unrelated build failure). */
#define BUILD_FAIL (-2147483647 - 1)
struct row {
const char *label;
const char *src;
int want;
const char *expect_err; /* BUILD_FAIL rows: required stderr substring */
};
static const struct row rows[] = {
{ "full_range",
"package main;\n"
"export fn main() i32 = {\n"
"\tlet xs: []i64 = [];\n"
"\tappend(xs, 7);\n"
"\tappend(xs, 11);\n"
"\tappend(xs, 13);\n"
"\tdelete(xs[:]);\n"
"\treturn len(xs): i32 + 7;\n"
"};\n",
7, NULL },
{ "full_explicit",
"package main;\n"
"export fn main() i32 = {\n"
"\tlet xs: []i64 = [];\n"
"\tappend(xs, 7);\n"
"\tappend(xs, 11);\n"
"\tappend(xs, 13);\n"
"\tdelete(xs[0:3]);\n"
"\treturn len(xs): i32 + 8;\n"
"};\n",
8, NULL },
/* Second full-range delete sees len == 0: count = len - 0 = 0,
* the copy loop exits at entry, len -= 0. Pins that an emptied
* slice survives a re-clear (the regex `)` arm clears
* jump_idxs[group_level] whether or not `|` populated it). */
{ "full_twice",
"package main;\n"
"export fn main() i32 = {\n"
"\tlet xs: []i64 = [];\n"
"\tappend(xs, 7);\n"
"\tappend(xs, 11);\n"
"\tdelete(xs[:]);\n"
"\tdelete(xs[:]);\n"
"\treturn len(xs): i32 + 9;\n"
"};\n",
9, NULL },
{ "head",
"package main;\n"
"export fn main() i32 = {\n"
"\tlet xs: []i64 = [];\n"
"\tappend(xs, 1);\n"
"\tappend(xs, 2);\n"
"\tappend(xs, 3);\n"
"\tdelete(xs[:2]);\n"
"\treturn xs[0]: i32 + len(xs)*10;\n"
"};\n",
13, NULL },
{ "mid",
"package main;\n"
"export fn main() i32 = {\n"
"\tlet xs: []i64 = [];\n"
"\tappend(xs, 1);\n"
"\tappend(xs, 2);\n"
"\tappend(xs, 3);\n"
"\tappend(xs, 4);\n"
"\tappend(xs, 5);\n"
"\tdelete(xs[1:3]);\n"
"\treturn (xs[0]*100 + xs[1]*10 + xs[2]): i32 + len(xs);\n"
"};\n",
148, NULL },
{ "tail_open",
"package main;\n"
"export fn main() i32 = {\n"
"\tlet xs: []i64 = [];\n"
"\tappend(xs, 1);\n"
"\tappend(xs, 2);\n"
"\tappend(xs, 3);\n"
"\tdelete(xs[1:]);\n"
"\treturn xs[0]: i32 + len(xs)*10;\n"
"};\n",
11, NULL },
{ "empty_range",
"package main;\n"
"export fn main() i32 = {\n"
"\tlet xs: []i64 = [];\n"
"\tappend(xs, 1);\n"
"\tappend(xs, 2);\n"
"\tdelete(xs[1:1]);\n"
"\treturn (xs[0] + xs[1]): i32 + len(xs)*10;\n"
"};\n",
23, NULL },
/* lo == hi == len: the loop guard j+count >= len holds at entry
* with j = len, so nothing past the end is ever addressed. */
{ "end_boundary",
"package main;\n"
"export fn main() i32 = {\n"
"\tlet xs: []i64 = [];\n"
"\tappend(xs, 1);\n"
"\tappend(xs, 2);\n"
"\tdelete(xs[len(xs):len(xs)]);\n"
"\treturn (xs[0]*10 + xs[1]): i32 + len(xs)*100;\n"
"};\n",
212, NULL },
/* Never-appended empty slice: ptr is nil. count 0 and len 0 mean
* the loop body (the only dereferencing code) never runs. */
{ "empty_explicit",
"package main;\n"
"export fn main() i32 = {\n"
"\tlet xs: []i64 = [];\n"
"\tdelete(xs[0:0]);\n"
"\treturn len(xs): i32 + 31;\n"
"};\n",
31, NULL },
{ "single_via_range",
"package main;\n"
"export fn main() i32 = {\n"
"\tlet xs: []i64 = [];\n"
"\tappend(xs, 7);\n"
"\tappend(xs, 11);\n"
"\tappend(xs, 13);\n"
"\tdelete(xs[1:2]);\n"
"\tlet ys: []i64 = [];\n"
"\tappend(ys, 7);\n"
"\tappend(ys, 11);\n"
"\tappend(ys, 13);\n"
"\tdelete(ys[1]);\n"
"\tif (len(xs) != len(ys)) { return 99; };\n"
"\tif (xs[0] != ys[0]) { return 98; };\n"
"\tif (xs[1] != ys[1]) { return 97; };\n"
"\treturn 42;\n"
"};\n",
42, NULL },
{ "cap_preserved",
"package main;\n"
"export fn main() i32 = {\n"
"\tlet hb: [8]u8; hb[0] = 4u8; hb[1] = 6u8; hb[2] = 9u8;\n"
"\tlet xs: []u8; xs.ptr = &hb[0]; xs.len = 3; xs.cap = 8;\n"
"\tdelete(xs[:]);\n"
"\treturn (xs.cap*10 + xs.len): i32;\n"
"};\n",
80, NULL },
{ "u8_narrow",
"package main;\n"
"export fn main() i32 = {\n"
"\tlet xs: []u8 = [];\n"
"\tappend(xs, 1u8);\n"
"\tappend(xs, 2u8);\n"
"\tappend(xs, 3u8);\n"
"\tappend(xs, 4u8);\n"
"\tappend(xs, 5u8);\n"
"\tdelete(xs[1:4]);\n"
"\treturn (xs[0]*10u8 + xs[1]): i32 + len(xs)*100;\n"
"};\n",
215, NULL },
{ "u16_narrow",
"package main;\n"
"export fn main() i32 = {\n"
"\tlet xs: []u16 = [];\n"
"\tappend(xs, 1u16);\n"
"\tappend(xs, 2u16);\n"
"\tappend(xs, 3u16);\n"
"\tappend(xs, 4u16);\n"
"\tdelete(xs[1:3]);\n"
"\treturn (xs[0] + xs[1]): i32 + len(xs)*10;\n"
"};\n",
25, NULL },
{ "i32_narrow",
"package main;\n"
"export fn main() i32 = {\n"
"\tlet xs: []i32 = [];\n"
"\tappend(xs, 10i32);\n"
"\tappend(xs, 20i32);\n"
"\tappend(xs, 30i32);\n"
"\tappend(xs, 40i32);\n"
"\tdelete(xs[2:4]);\n"
"\treturn xs[0] + xs[1] + len(xs)*10;\n"
"};\n",
50, NULL },
/* esz 16: exactly two whole qwords per element — the only spread
* point with multiple MOVQs and NO sub-word tail. */
{ "struct16",
"package main;\n"
"type pair = struct { a: i64, b: i64 };\n"
"export fn main() i32 = {\n"
"\tlet xs: []pair = [];\n"
"\tappend(xs, pair{ a = 1, b = 2 });\n"
"\tappend(xs, pair{ a = 3, b = 4 });\n"
"\tappend(xs, pair{ a = 5, b = 6 });\n"
"\tdelete(xs[1:2]);\n"
"\treturn (xs[0].a*10 + xs[1].b): i32 + len(xs)*100;\n"
"};\n",
216, NULL },
{ "str_elem",
"package main;\n"
"export fn main() i32 = {\n"
"\tlet xs: []str = [];\n"
"\tlet a: str = \"abc\";\n"
"\tlet b: str = \"de\";\n"
"\tlet g: str = \"fghi\";\n"
"\tlet h: str = \"jklmn\";\n"
"\tappend(xs, a);\n"
"\tappend(xs, b);\n"
"\tappend(xs, g);\n"
"\tappend(xs, h);\n"
"\tdelete(xs[1:3]);\n"
"\treturn (xs[0].len*10 + xs[1].len): i32 + len(xs)*100;\n"
"};\n",
235, NULL },
/* The fold-5a consumer, shape-EXACT: indexed local-slice base,
* size-typed outer index, defaulted full range (regex.ha:333
* `delete(jump_idxs[group_level][..])`; ww spells `[..]` as
* `[:]`, D7). The header lives at jj.ptr + g*size([]size) —
* outer stride off the type table, never a literal. */
{ "regex_exact",
"package main;\n"
"export fn main() i32 = {\n"
"\tlet jump_idxs: [][]size = [];\n"
"\tlet lvl0: []size = [];\n"
"\tlet a: size = 3;\n"
"\tlet b: size = 9;\n"
"\tappend(lvl0, a);\n"
"\tappend(lvl0, b);\n"
"\tappend(jump_idxs, lvl0);\n"
"\tlet group_level: size = 0;\n"
"\tdelete(jump_idxs[group_level][:]);\n"
"\treturn len(jump_idxs[0]): i32 + len(jump_idxs): i32 * 10;\n"
"};\n",
10, NULL },
{ "deref_shape",
"package main;\n"
"fn clear(p: *[]u8) void = {\n"
"\tdelete((*p)[:]);\n"
"};\n"
"export fn main() i32 = {\n"
"\tlet xs: []u8 = [];\n"
"\tappend(xs, 5u8);\n"
"\tappend(xs, 6u8);\n"
"\tclear(&xs);\n"
"\treturn len(xs): i32 + 17;\n"
"};\n",
17, NULL },
/* lo and hi as CALLs: each side effect fires exactly once, lo
* before hi (trace 12, never 21/121/122) — pins that neither
* bound expression is re-evaluated by the count computation or
* the copy loop. */
{ "order_of_eval",
"package main;\n"
"let trace: i64 = 0;\n"
"fn lof() size = {\n"
"\ttrace = trace*10 + 1;\n"
"\treturn 1;\n"
"};\n"
"fn hif() size = {\n"
"\ttrace = trace*10 + 2;\n"
"\treturn 3;\n"
"};\n"
"export fn main() i32 = {\n"
"\tlet xs: []i64 = [];\n"
"\tappend(xs, 1);\n"
"\tappend(xs, 2);\n"
"\tappend(xs, 3);\n"
"\tappend(xs, 4);\n"
"\tdelete(xs[lof():hif()]);\n"
"\tif (trace != 12) { return 90; };\n"
"\treturn (xs[0]*10 + xs[1]): i32 + len(xs);\n"
"};\n",
16, NULL },
/* lo mutates the slice it is deleting from: only the header
* ADDRESS is taken before operand eval; ptr/len/elements are read
* after, so poke's write to xs[0] survives into the result. */
{ "alias_visible",
"package main;\n"
"fn poke(p: *[]i64) size = {\n"
"\t(*p)[0] = 9;\n"
"\treturn 1;\n"
"};\n"
"export fn main() i32 = {\n"
"\tlet xs: []i64 = [];\n"
"\tappend(xs, 1);\n"
"\tappend(xs, 2);\n"
"\tappend(xs, 3);\n"
"\tdelete(xs[poke(&xs):]);\n"
"\treturn xs[0]: i32 * 10 + len(xs);\n"
"};\n",
91, NULL },
{ "reject_array",
"package main;\n"
"export fn main() i32 = {\n"
"\tlet t: [3]i64 = [1, 2, 3];\n"
"\tdelete(t[0:2]);\n"
"\treturn 0;\n"
"};\n",
BUILD_FAIL, "delete must operate on a slice" },
{ "reject_str",
"package main;\n"
"export fn main() i32 = {\n"
"\tlet s: str = \"abc\";\n"
"\tdelete(s[:]);\n"
"\treturn 0;\n"
"};\n",
BUILD_FAIL, "delete must operate on a slice" },
};
/* errlog_has — the build-failure stderr must carry the row's expected
* diagnostic; any other failure (parse error, crash) is a vacuous
* reject and must not pass. */
static int
errlog_has(const char *path, const char *needle)
{
FILE *f = fopen(path, "rb");
if (!f) return 0;
char buf[8192];
size_t got = fread(buf, 1, sizeof buf - 1, f);
fclose(f);
buf[got] = '\0';
return strstr(buf, needle) != NULL;
}
static int
run_driver(const char *driver, const struct row *r, int i)
{
char src[64], tmpdir[64], errlog[80], cmd[1200];
snprintf(src, sizeof src, "/tmp/delr_%d_%d.ww", getpid(), i);
snprintf(tmpdir, sizeof tmpdir, "/tmp/delr_%d_d_%d", getpid(), i);
snprintf(errlog, sizeof errlog, "%s.err", src);
FILE *f = fopen(src, "wb");
if (!f) return -1;
fputs(r->src, f);
fclose(f);
mkdir(tmpdir, 0755);
snprintf(cmd, sizeof cmd, "cd %s && %s build %s 2>%s",
tmpdir, driver, src, errlog);
if (runwait(cmd) != 0) {
int rc = -1;
if (r->want != BUILD_FAIL) {
fprintf(stderr, "row[%s]: build via %s failed\n",
r->label, driver);
} else if (r->expect_err &&
!errlog_has(errlog, r->expect_err)) {
fprintf(stderr, "row[%s]: %s build failed without "
"expected diagnostic \"%s\"\n",
r->label, driver, r->expect_err);
rc = -3; /* failed, but for the wrong reason */
}
unlink(src); unlink(errlog); rmdir(tmpdir);
return rc;
}
const char *base = strrchr(src, '/');
base = base ? base + 1 : src;
char outbin[128];
snprintf(outbin, sizeof outbin, "%s/%s", tmpdir, base);
char *dot = strrchr(outbin, '.');
if (dot && strcmp(dot, ".ww") == 0) *dot = '\0';
int got = runwait(outbin);
unlink(src); unlink(errlog); unlink(outbin); rmdir(tmpdir);
return got;
}
/* asm_byte_identical — compile the row's source standalone with w6c and
* w6c_ww and diff. Both range-delete lowerings are written fresh, so
* this is the converged-by-construction gate: any drift in the shift
* loop, the rdl_l/rdl_e label sequence, the dynamic-src copy, or the
* indexed-base header address shows here. */
static int
asm_byte_identical(const char *bin, const struct row *r, int i)
{
char src[64], cs[64], ws[64], cmd[1024];
snprintf(src, sizeof src, "/tmp/delr_asm_%d_%d.ww", getpid(), i);
snprintf(cs, sizeof cs, "/tmp/delr_asm_%d_%d_c.s", getpid(), i);
snprintf(ws, sizeof ws, "/tmp/delr_asm_%d_%d_w.s", getpid(), i);
FILE *f = fopen(src, "wb");
if (!f) return -1;
fputs(r->src, f);
fclose(f);
snprintf(cmd, sizeof cmd, "%s/w6c -o %s %s 2>/dev/null", bin, cs, src);
if (runwait(cmd) != 0) {
fprintf(stderr, "row[%s]: w6c errored\n", r->label);
unlink(src);
return -1;
}
snprintf(cmd, sizeof cmd, "%s/w6c_ww -o %s %s 2>/dev/null",
bin, ws, src);
if (runwait(cmd) != 0) {
fprintf(stderr, "row[%s]: w6c_ww errored\n", r->label);
unlink(src); unlink(cs);
return -1;
}
FILE *fc = fopen(cs, "rb");
FILE *fw = fopen(ws, "rb");
int rc = 0;
if (!fc || !fw) {
rc = -1;
} else {
for (;;) {
int a = fgetc(fc);
int b = fgetc(fw);
if (a != b) { rc = -1; break; }
if (a == EOF) break;
}
}
if (fc) fclose(fc);
if (fw) fclose(fw);
if (rc != 0)
fprintf(stderr, "row[%s]: cstage vs wwstage asm differs\n",
r->label);
unlink(src); unlink(cs); unlink(ws);
return rc;
}
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 cdrv[1024];
snprintf(cdrv, sizeof cdrv, "%s/ww", bin);
char wdrv[1024];
snprintf(wdrv, sizeof wdrv, "%s/ww_ww", bin);
struct { const char *name; const char *path; int gated_on_existence; }
drivers[] = {
{ "cstage", cdrv, 0 },
{ "wwstage", wdrv, 1 },
{ NULL, NULL, 0 },
};
int n = (int)(sizeof rows / sizeof rows[0]);
int total = 0, fail = 0;
for (int d = 0; drivers[d].name; d++) {
if (drivers[d].gated_on_existence
&& access(drivers[d].path, X_OK) != 0) {
fprintf(stderr, "delete_range: skip %s (no %s)\n",
drivers[d].name, drivers[d].path);
continue;
}
for (int i = 0; i < n; i++) {
int got = run_driver(drivers[d].path, &rows[i], i);
total++;
int bad = rows[i].want == BUILD_FAIL
? (got != -1) : (got != rows[i].want);
if (bad) {
fprintf(stderr,
"delete_range[%s][%s]: exit=%d want=%d\n",
drivers[d].name, rows[i].label,
got, rows[i].want);
fail++;
}
}
}
if (access(wdrv, X_OK) == 0) {
for (int i = 0; i < n; i++) {
if (rows[i].want == BUILD_FAIL)
continue;
total++;
if (asm_byte_identical(bin, &rows[i], i) != 0)
fail++;
}
}
if (fail) {
fprintf(stderr,
"delete_range: %d/%d fixtures failed\n", fail, total);
return 1;
}
printf("delete_range: %d fixtures passed\n", total);
return 0;
}