Hojun-Cho
a9dcea70ed
prev()'s walk-back decremented offs (i32) past 0 to -1 and returned `more`; a subsequent next() then passed the signed `-1 < len` guard and read d.src[-1] — a silent OOB decode of a garbage rune (no runtime bounds net). Hare's decoder.offs is `size`: the underflow wraps to SIZE_MAX so every `offs < len` guard exits safely (next returns more, not a rune). Change offs to size and spell prev's loop as the Hare-form `offs < len` guard; index sites take an i32 temp (ww's slice index is i32 and `[...]` reads ':' as the slice separator). No-runtime-net residual: remaining() would silently build a ptr-1/len+1 OOB view when called in the post-`more` state; guard it with a loud abort (caller contract: don't call after `more`). The offs type ripples into strings.ww's iterator<->decoder bridge (move/slice) — cast at the four sites, safe on the rune-return path where offs is in range. utf8/strings embed into all five selfhost combined.ww snapshots plus the smoke.combined.ww test amalgamation; all regen'd. utf8test gains prev_more_then_next_no_oob pinning the closed OOB.
643 lines
20 KiB
Plaintext
643 lines
20 KiB
Plaintext
// utf8test — exercises lib/encoding/utf8. Run with
|
|
// `out/bin/ww run lib/encoding/utf8/utf8test.ww`. A failing row aborts
|
|
// via the assert/abort builtin (task #5 @test conversion).
|
|
|
|
package utf8_test;
|
|
|
|
import bytes;
|
|
import encoding.utf8;
|
|
|
|
|
|
fn streq(a: str, b: str) bool = {
|
|
if (a.len != b.len) { return false; };
|
|
let i: i32 = 0;
|
|
for (i < a.len) {
|
|
if (a[i] != b[i]) { return false; };
|
|
i += 1;
|
|
};
|
|
return true;
|
|
};
|
|
|
|
// ---- runesz: byte length per range ------------------------------------
|
|
// ref/hare/encoding/utf8/rune.ha:5. Boundaries: 0x7F → 1, 0x80 → 2,
|
|
// 0x7FF → 2, 0x800 → 3, 0xFFFF → 3, 0x10000 → 4, 0x10FFFF → 4.
|
|
|
|
@test fn runesz_ranges() void = {
|
|
assert(!(utf8.runesz(0u32: rune) != 1));
|
|
assert(!(utf8.runesz(0x7Fu32: rune) != 1));
|
|
assert(!(utf8.runesz(0x80u32: rune) != 2));
|
|
assert(!(utf8.runesz(0x7FFu32: rune) != 2));
|
|
assert(!(utf8.runesz(0x800u32: rune) != 3));
|
|
assert(!(utf8.runesz(0xFFFFu32: rune) != 3));
|
|
assert(!(utf8.runesz(0x10000u32: rune) != 4));
|
|
assert(!(utf8.runesz(0x10FFFFu32: rune) != 4));
|
|
};
|
|
|
|
// ---- utf8sz: start-byte classification --------------------------------
|
|
// ref/hare/encoding/utf8/rune.ha:15. ASCII → 1; legal multibyte
|
|
// leads → 2/3/4; continuation and >0xF7 → invalid.
|
|
|
|
@test fn utf8sz_classify() void = {
|
|
match (utf8.utf8sz(0u8)) {
|
|
case let n: i32 => { assert(!(n != 1)); };
|
|
case let e: utf8.invalid => { abort(); };
|
|
};
|
|
match (utf8.utf8sz(0x7Fu8)) {
|
|
case let n: i32 => { assert(!(n != 1)); };
|
|
case let e: utf8.invalid => { abort(); };
|
|
};
|
|
match (utf8.utf8sz(0x80u8)) { // continuation
|
|
case let n: i32 => { abort(); };
|
|
case let e: utf8.invalid => void;
|
|
};
|
|
match (utf8.utf8sz(0xC1u8)) { // overlong 2-byte lead
|
|
case let n: i32 => { abort(); };
|
|
case let e: utf8.invalid => void;
|
|
};
|
|
match (utf8.utf8sz(0xC2u8)) {
|
|
case let n: i32 => { assert(!(n != 2)); };
|
|
case let e: utf8.invalid => { abort(); };
|
|
};
|
|
match (utf8.utf8sz(0xE0u8)) {
|
|
case let n: i32 => { assert(!(n != 3)); };
|
|
case let e: utf8.invalid => { abort(); };
|
|
};
|
|
match (utf8.utf8sz(0xF0u8)) {
|
|
case let n: i32 => { assert(!(n != 4)); };
|
|
case let e: utf8.invalid => { abort(); };
|
|
};
|
|
match (utf8.utf8sz(0xF8u8)) { // 5-byte lead — illegal in modern UTF-8
|
|
case let n: i32 => { abort(); };
|
|
case let e: utf8.invalid => void;
|
|
};
|
|
match (utf8.utf8sz(0xFFu8)) {
|
|
case let n: i32 => { abort(); };
|
|
case let e: utf8.invalid => void;
|
|
};
|
|
};
|
|
|
|
// ---- encoderune: all four widths --------------------------------------
|
|
// ref/hare/encoding/utf8/encode.ha:7. Byte vectors come from the
|
|
// Unicode specification (UAX standard examples).
|
|
|
|
@test fn encode_ascii() void = {
|
|
let out: [4]u8;
|
|
let n: i32 = utf8.encoderune(out[0:4], 0x41u32: rune); // 'A'
|
|
assert(!(n != 1));
|
|
assert(!(out[0] != 0x41u8));
|
|
};
|
|
|
|
@test fn encode_two_byte() void = {
|
|
let out: [4]u8;
|
|
let n: i32 = utf8.encoderune(out[0:4], 0xE9u32: rune); // 'é' U+00E9
|
|
assert(!(n != 2));
|
|
assert(!(out[0] != 0xC3u8));
|
|
assert(!(out[1] != 0xA9u8));
|
|
};
|
|
|
|
@test fn encode_three_byte() void = {
|
|
let out: [4]u8;
|
|
let n: i32 = utf8.encoderune(out[0:4], 0x20ACu32: rune); // '€' U+20AC
|
|
assert(!(n != 3));
|
|
assert(!(out[0] != 0xE2u8));
|
|
assert(!(out[1] != 0x82u8));
|
|
assert(!(out[2] != 0xACu8));
|
|
};
|
|
|
|
@test fn encode_four_byte() void = {
|
|
let out: [4]u8;
|
|
let n: i32 = utf8.encoderune(out[0:4], 0x1F980u32: rune); // '🦀' U+1F980
|
|
assert(!(n != 4));
|
|
assert(!(out[0] != 0xF0u8));
|
|
assert(!(out[1] != 0x9Fu8));
|
|
assert(!(out[2] != 0xA6u8));
|
|
assert(!(out[3] != 0x80u8));
|
|
};
|
|
|
|
// ---- decoder.next: valid 1/2/3/4-byte ---------------------------------
|
|
// ref/hare/encoding/utf8/decode.ha:27. Round-trip via the same byte
|
|
// vectors used in encode.
|
|
|
|
@test fn decode_one_byte() void = {
|
|
let src: [1]u8;
|
|
src[0] = 0x41u8;
|
|
let d: utf8.decoder = utf8.decode(src[0:1]);
|
|
match (utf8.next(&d)) {
|
|
case utf8.done => { abort(); };
|
|
case utf8.more => { abort(); };
|
|
case let r: utf8.invalid => { abort(); };
|
|
case let r: rune => { assert(!(r != 0x41u32: rune)); };
|
|
};
|
|
};
|
|
|
|
@test fn decode_two_byte() void = {
|
|
let src: [2]u8;
|
|
src[0] = 0xC3u8; src[1] = 0xA9u8;
|
|
let d: utf8.decoder = utf8.decode(src[0:2]);
|
|
match (utf8.next(&d)) {
|
|
case utf8.done => { abort(); };
|
|
case utf8.more => { abort(); };
|
|
case let r: utf8.invalid => { abort(); };
|
|
case let r: rune => { assert(!(r != 0xE9u32: rune)); };
|
|
};
|
|
};
|
|
|
|
@test fn decode_three_byte() void = {
|
|
let src: [3]u8;
|
|
src[0] = 0xE2u8; src[1] = 0x82u8; src[2] = 0xACu8;
|
|
let d: utf8.decoder = utf8.decode(src[0:3]);
|
|
match (utf8.next(&d)) {
|
|
case utf8.done => { abort(); };
|
|
case utf8.more => { abort(); };
|
|
case let r: utf8.invalid => { abort(); };
|
|
case let r: rune => { assert(!(r != 0x20ACu32: rune)); };
|
|
};
|
|
};
|
|
|
|
@test fn decode_four_byte() void = {
|
|
let src: [4]u8;
|
|
src[0] = 0xF0u8; src[1] = 0x9Fu8; src[2] = 0xA6u8; src[3] = 0x80u8;
|
|
let d: utf8.decoder = utf8.decode(src[0:4]);
|
|
match (utf8.next(&d)) {
|
|
case utf8.done => { abort(); };
|
|
case utf8.more => { abort(); };
|
|
case let r: utf8.invalid => { abort(); };
|
|
case let r: rune => { assert(!(r != 0x1F980u32: rune)); };
|
|
};
|
|
};
|
|
|
|
// ---- decoder.next: invalid inputs -------------------------------------
|
|
// Cases mirror ref/hare/encoding/utf8/decode.ha:127-167 (surrogate /
|
|
// overlong / out-of-range / bad-continuation).
|
|
|
|
@test fn decode_surrogate() void = {
|
|
let src: [3]u8;
|
|
src[0] = 0xEDu8; src[1] = 0xA0u8; src[2] = 0x80u8; // U+D800
|
|
let d: utf8.decoder = utf8.decode(src[0:3]);
|
|
match (utf8.next(&d)) {
|
|
case let r: utf8.invalid => void;
|
|
case let r: rune => { abort(); };
|
|
case utf8.done => { abort(); };
|
|
case utf8.more => { abort(); };
|
|
};
|
|
};
|
|
|
|
@test fn decode_overlong() void = {
|
|
let src: [4]u8;
|
|
src[0] = 0xF0u8; src[1] = 0x82u8; src[2] = 0x82u8; src[3] = 0xACu8;
|
|
let d: utf8.decoder = utf8.decode(src[0:4]);
|
|
match (utf8.next(&d)) {
|
|
case let r: utf8.invalid => void;
|
|
case let r: rune => { abort(); };
|
|
case utf8.done => { abort(); };
|
|
case utf8.more => { abort(); };
|
|
};
|
|
};
|
|
|
|
@test fn decode_out_of_range() void = {
|
|
let src: [4]u8;
|
|
src[0] = 0xF5u8; src[1] = 0x94u8; src[2] = 0x80u8; src[3] = 0x80u8;
|
|
let d: utf8.decoder = utf8.decode(src[0:4]);
|
|
match (utf8.next(&d)) {
|
|
case let r: utf8.invalid => void;
|
|
case let r: rune => { abort(); };
|
|
case utf8.done => { abort(); };
|
|
case utf8.more => { abort(); };
|
|
};
|
|
};
|
|
|
|
// ref/hare/encoding/utf8/decode.ha:141 — `[0xC2, 0xFF]`: legal 2-byte
|
|
// lead followed by non-continuation. Pins next-table cell (state 1,
|
|
// byte 0xFF) returning -1; distinct from overlong (which is filtered
|
|
// in state 3/5/7 by lead-byte-aware sub-states).
|
|
@test fn decode_bad_continuation() void = {
|
|
let src: [2]u8;
|
|
src[0] = 0xC2u8; src[1] = 0xFFu8;
|
|
let d: utf8.decoder = utf8.decode(src[0:2]);
|
|
match (utf8.next(&d)) {
|
|
case let r: utf8.invalid => void;
|
|
case let r: rune => { abort(); };
|
|
case utf8.done => { abort(); };
|
|
case utf8.more => { abort(); };
|
|
};
|
|
};
|
|
|
|
// ref/hare/encoding/utf8/decode.ha:151 — `[0xF4, 0x8F, 0xBF, 0xBF]` =
|
|
// U+10FFFF, the largest legal codepoint. Pins the upper boundary;
|
|
// pairs with the existing `0xF5…` out-of-range row.
|
|
@test fn decode_max_in_range() void = {
|
|
let src: [4]u8;
|
|
src[0] = 0xF4u8; src[1] = 0x8Fu8; src[2] = 0xBFu8; src[3] = 0xBFu8;
|
|
let d: utf8.decoder = utf8.decode(src[0:4]);
|
|
match (utf8.next(&d)) {
|
|
case let r: rune => { assert(!(r != 0x10FFFFu32: rune)); };
|
|
case let r: utf8.invalid => { abort(); };
|
|
case utf8.done => { abort(); };
|
|
case utf8.more => { abort(); };
|
|
};
|
|
};
|
|
|
|
// ---- decoder.next: truncated → more -----------------------------------
|
|
|
|
@test fn decode_truncated() void = {
|
|
let src: [2]u8;
|
|
src[0] = 0xE3u8; src[1] = 0x81u8; // missing 3rd byte
|
|
let d: utf8.decoder = utf8.decode(src[0:2]);
|
|
match (utf8.next(&d)) {
|
|
case utf8.more => void;
|
|
case let r: rune => { abort(); };
|
|
case utf8.done => { abort(); };
|
|
case let r: utf8.invalid => { abort(); };
|
|
};
|
|
};
|
|
|
|
// ---- decoder.next: done at EOI ----------------------------------------
|
|
|
|
@test fn decode_done() void = {
|
|
let src: [1]u8;
|
|
let d: utf8.decoder = utf8.decode(src[0:0]);
|
|
match (utf8.next(&d)) {
|
|
case utf8.done => void;
|
|
case let r: rune => { abort(); };
|
|
case utf8.more => { abort(); };
|
|
case let r: utf8.invalid => { abort(); };
|
|
};
|
|
};
|
|
|
|
// ---- validate: well-formed mixed-width vs malformed -------------------
|
|
// ref/hare/encoding/utf8/decode.ha:207. Mixed input mirrors Hare's
|
|
// decode @test ('こんにちは' + NUL).
|
|
|
|
@test fn validate_mixed_ok() void = {
|
|
let src: [16]u8;
|
|
src[0] = 0xE3u8; src[1] = 0x81u8; src[2] = 0x93u8; // こ
|
|
src[3] = 0xE3u8; src[4] = 0x82u8; src[5] = 0x93u8; // ん
|
|
src[6] = 0xE3u8; src[7] = 0x81u8; src[8] = 0xABu8; // に
|
|
src[9] = 0xE3u8; src[10] = 0x81u8; src[11] = 0xA1u8; // ち
|
|
src[12] = 0xE3u8; src[13] = 0x81u8; src[14] = 0xAFu8; // は
|
|
src[15] = 0u8;
|
|
match (utf8.validate(src[0:16])) {
|
|
case let e: utf8.invalid => { abort(); };
|
|
case void => void;
|
|
};
|
|
};
|
|
|
|
@test fn validate_malformed() void = {
|
|
let src: [3]u8;
|
|
src[0] = 0xEDu8; src[1] = 0xA0u8; src[2] = 0x80u8; // surrogate
|
|
match (utf8.validate(src[0:3])) {
|
|
case let e: utf8.invalid => void;
|
|
case void => { abort(); };
|
|
};
|
|
};
|
|
|
|
@test fn validate_empty_ok() void = {
|
|
let src: [1]u8;
|
|
match (utf8.validate(src[0:0])) {
|
|
case let e: utf8.invalid => { abort(); };
|
|
case void => void;
|
|
};
|
|
};
|
|
|
|
// ---- decoder.prev: done at start-of-input -----------------------------
|
|
// ref/hare/encoding/utf8/decode.ha:53-55. Cursor at offs=0 has no prior
|
|
// codepoint.
|
|
|
|
@test fn prev_done_at_start() void = {
|
|
let src: [4]u8;
|
|
src[0] = 0x41u8;
|
|
let d: utf8.decoder = utf8.decode(src[0:1]);
|
|
match (utf8.prev(&d)) {
|
|
case utf8.done => void;
|
|
case let r: rune => { abort(); };
|
|
case utf8.more => { abort(); };
|
|
case let e: utf8.invalid => { abort(); };
|
|
};
|
|
};
|
|
|
|
// ---- decoder.prev: each codepoint width round-trips -------------------
|
|
// ref/hare/encoding/utf8/decode.ha:56-71. Forward-decode one rune,
|
|
// reverse-decode back to the same rune.
|
|
|
|
@test fn prev_one_byte() void = {
|
|
let src: [1]u8;
|
|
src[0] = 0x41u8;
|
|
let d: utf8.decoder = utf8.decode(src[0:1]);
|
|
match (utf8.next(&d)) {
|
|
case let r: rune => void;
|
|
case => { abort(); };
|
|
};
|
|
match (utf8.prev(&d)) {
|
|
case let r: rune => { assert(!(r != 0x41u32: rune)); };
|
|
case => { abort(); };
|
|
};
|
|
};
|
|
|
|
@test fn prev_two_byte() void = {
|
|
let src: [2]u8;
|
|
src[0] = 0xC3u8; src[1] = 0xA9u8; // 'é'
|
|
let d: utf8.decoder = utf8.decode(src[0:2]);
|
|
match (utf8.next(&d)) {
|
|
case let r: rune => void;
|
|
case => { abort(); };
|
|
};
|
|
match (utf8.prev(&d)) {
|
|
case let r: rune => { assert(!(r != 0xE9u32: rune)); };
|
|
case => { abort(); };
|
|
};
|
|
};
|
|
|
|
@test fn prev_three_byte() void = {
|
|
let src: [3]u8;
|
|
src[0] = 0xE2u8; src[1] = 0x82u8; src[2] = 0xACu8; // '€'
|
|
let d: utf8.decoder = utf8.decode(src[0:3]);
|
|
match (utf8.next(&d)) {
|
|
case let r: rune => void;
|
|
case => { abort(); };
|
|
};
|
|
match (utf8.prev(&d)) {
|
|
case let r: rune => { assert(!(r != 0x20ACu32: rune)); };
|
|
case => { abort(); };
|
|
};
|
|
};
|
|
|
|
@test fn prev_four_byte() void = {
|
|
let src: [4]u8;
|
|
src[0] = 0xF0u8; src[1] = 0x9Fu8; src[2] = 0xA6u8; src[3] = 0x80u8; // '🦀'
|
|
let d: utf8.decoder = utf8.decode(src[0:4]);
|
|
match (utf8.next(&d)) {
|
|
case let r: rune => void;
|
|
case => { abort(); };
|
|
};
|
|
match (utf8.prev(&d)) {
|
|
case let r: rune => { assert(!(r != 0x1F980u32: rune)); };
|
|
case => { abort(); };
|
|
};
|
|
};
|
|
|
|
// ---- decoder.prev: full forward + full reverse on mixed input ---------
|
|
// ref/hare/encoding/utf8/decode.ha:85-111. After consuming all runes
|
|
// forward, prev walks back through them in reverse order; prev at the
|
|
// start-of-input then returns done.
|
|
|
|
@test fn prev_mixed_roundtrip() void = {
|
|
let src: [16]u8;
|
|
src[0] = 0xE3u8; src[1] = 0x81u8; src[2] = 0x93u8; // こ
|
|
src[3] = 0xE3u8; src[4] = 0x82u8; src[5] = 0x93u8; // ん
|
|
src[6] = 0xE3u8; src[7] = 0x81u8; src[8] = 0xABu8; // に
|
|
src[9] = 0xE3u8; src[10] = 0x81u8; src[11] = 0xA1u8; // ち
|
|
src[12] = 0xE3u8; src[13] = 0x81u8; src[14] = 0xAFu8; // は
|
|
src[15] = 0u8;
|
|
let d: utf8.decoder = utf8.decode(src[0:16]);
|
|
let fwd: [6]u32;
|
|
let i: i32 = 0;
|
|
for (i < 6) {
|
|
match (utf8.next(&d)) {
|
|
case let r: rune => { fwd[i] = r: u32; };
|
|
case => { abort(); };
|
|
};
|
|
i += 1;
|
|
};
|
|
assert(!(utf8.position(&d) != 16));
|
|
i = 0;
|
|
for (i < 6) {
|
|
match (utf8.prev(&d)) {
|
|
case let r: rune => {
|
|
assert(!((r: u32) != fwd[5 - i]));
|
|
};
|
|
case => { abort(); };
|
|
};
|
|
i += 1;
|
|
};
|
|
match (utf8.prev(&d)) {
|
|
case utf8.done => void;
|
|
case => { abort(); };
|
|
};
|
|
};
|
|
|
|
// ---- decoder.prev: invalid inputs -------------------------------------
|
|
// ref/hare/encoding/utf8/decode.ha:113-150. Two-continuation walk
|
|
// reaches offs=0 without an initial byte → more; otherwise the forward
|
|
// re-decode catches the malformed run as invalid.
|
|
|
|
// ref/hare/encoding/utf8/decode.ha:117 — `[0xA0, 0xA1]` is all
|
|
// continuation bytes; walking back from offs=2 finds no initial byte
|
|
// within the 4-step bound nor before offs underflows, so the loop
|
|
// exits via the `>= 0` guard and returns `more`.
|
|
@test fn prev_continuation_only_more() void = {
|
|
let src: [2]u8;
|
|
src[0] = 0xA0u8; src[1] = 0xA1u8;
|
|
let d: utf8.decoder = utf8.decode(src[0:2]);
|
|
d.offs = 2;
|
|
match (utf8.prev(&d)) {
|
|
case utf8.more => void;
|
|
case let r: rune => { abort(); };
|
|
case utf8.done => { abort(); };
|
|
case let e: utf8.invalid => { abort(); };
|
|
};
|
|
};
|
|
|
|
// #70: after prev() returns `more` it leaves offs out of range (SIZE_MAX,
|
|
// via the unsigned underflow). A subsequent next() must see offs >= len
|
|
// and return `more` — NOT read one byte before the buffer and decode a
|
|
// garbage rune. The byte before the slice (buf[2] = 0x41 = 'A') is the
|
|
// planted sentinel the old i32 offs=-1 path returned as rune 65.
|
|
@test fn prev_more_then_next_no_oob() void = {
|
|
let buf: [4]u8;
|
|
buf[0] = 0x58u8; buf[1] = 0x59u8; buf[2] = 0x41u8; // 'A' sentinel
|
|
buf[3] = 0xA0u8; // lone continuation
|
|
let d: utf8.decoder = utf8.decode(buf[3:4]);
|
|
match (utf8.next(&d)) {
|
|
case let e: utf8.invalid => void;
|
|
case => { abort(); };
|
|
};
|
|
match (utf8.prev(&d)) {
|
|
case utf8.more => void;
|
|
case => { abort(); };
|
|
};
|
|
match (utf8.next(&d)) {
|
|
case utf8.more => void; // safe: not src[-1]'s 'A'
|
|
case let r: rune => { abort(); };
|
|
case utf8.done => { abort(); };
|
|
case let e: utf8.invalid => { abort(); };
|
|
};
|
|
};
|
|
|
|
@test fn prev_incomplete_invalid() void = {
|
|
let src: [2]u8;
|
|
src[0] = 0xE3u8; src[1] = 0x81u8;
|
|
let d: utf8.decoder = utf8.decode(src[0:2]);
|
|
d.offs = 2;
|
|
match (utf8.prev(&d)) {
|
|
case let e: utf8.invalid => void;
|
|
case let r: rune => { abort(); };
|
|
case utf8.done => { abort(); };
|
|
case utf8.more => { abort(); };
|
|
};
|
|
};
|
|
|
|
@test fn prev_surrogate_invalid() void = {
|
|
let src: [3]u8;
|
|
src[0] = 0xEDu8; src[1] = 0xA0u8; src[2] = 0x80u8;
|
|
let d: utf8.decoder = utf8.decode(src[0:3]);
|
|
d.offs = 3;
|
|
match (utf8.prev(&d)) {
|
|
case let e: utf8.invalid => void;
|
|
case let r: rune => { abort(); };
|
|
case utf8.done => { abort(); };
|
|
case utf8.more => { abort(); };
|
|
};
|
|
};
|
|
|
|
@test fn prev_overlong_invalid() void = {
|
|
let src: [4]u8;
|
|
src[0] = 0xF0u8; src[1] = 0x82u8; src[2] = 0x82u8; src[3] = 0xACu8;
|
|
let d: utf8.decoder = utf8.decode(src[0:4]);
|
|
d.offs = 4;
|
|
match (utf8.prev(&d)) {
|
|
case let e: utf8.invalid => void;
|
|
case let r: rune => { abort(); };
|
|
case utf8.done => { abort(); };
|
|
case utf8.more => { abort(); };
|
|
};
|
|
};
|
|
|
|
// ref/hare/encoding/utf8/decode.ha:146 — extra continuation byte after
|
|
// a complete 2-byte codepoint. prev from offs=3 finds 0xC2 (initial
|
|
// byte), forward-decodes back to offs=2, but n=3 ≠ 2 → invalid.
|
|
@test fn prev_extracont_invalid() void = {
|
|
let src: [3]u8;
|
|
src[0] = 0xC2u8; src[1] = 0xA3u8; src[2] = 0x95u8;
|
|
let d: utf8.decoder = utf8.decode(src[0:3]);
|
|
d.offs = 3;
|
|
match (utf8.prev(&d)) {
|
|
case let e: utf8.invalid => void;
|
|
case let r: rune => { abort(); };
|
|
case utf8.done => { abort(); };
|
|
case utf8.more => { abort(); };
|
|
};
|
|
};
|
|
|
|
// ref/hare/encoding/utf8/decode.ha:158-163 — max in-range codepoint
|
|
// reverse-decoded. Pins state-7 acceptance via prev: from offs=4 the
|
|
// walk skips 3 continuations, lands on 0xF4 (state-7 lead), and the
|
|
// forward re-decode reproduces U+10FFFF.
|
|
@test fn prev_max_in_range() void = {
|
|
let src: [4]u8;
|
|
src[0] = 0xF4u8; src[1] = 0x8Fu8; src[2] = 0xBFu8; src[3] = 0xBFu8;
|
|
let d: utf8.decoder = utf8.decode(src[0:4]);
|
|
d.offs = 4;
|
|
match (utf8.prev(&d)) {
|
|
case let r: rune => { assert(!(r != 0x10FFFFu32: rune)); };
|
|
case let e: utf8.invalid => { abort(); };
|
|
case utf8.done => { abort(); };
|
|
case utf8.more => { abort(); };
|
|
};
|
|
};
|
|
|
|
// ref/hare/encoding/utf8/decode.ha:166-169 — `[0xF5, 0x94, 0x80, 0x80]`:
|
|
// 0xF5 is not a legal initial byte (dfa state-0 cell is -1). prev from
|
|
// offs=4 walks 4 bytes without finding an initial byte, trips the
|
|
// `n - d.offs == 4` bound, and returns invalid. The only test that
|
|
// exercises this branch (prev_continuation_only_more underflows past
|
|
// 0 before reaching 4).
|
|
@test fn prev_min_out_of_range() void = {
|
|
let src: [4]u8;
|
|
src[0] = 0xF5u8; src[1] = 0x94u8; src[2] = 0x80u8; src[3] = 0x80u8;
|
|
let d: utf8.decoder = utf8.decode(src[0:4]);
|
|
d.offs = 4;
|
|
match (utf8.prev(&d)) {
|
|
case let e: utf8.invalid => void;
|
|
case let r: rune => { abort(); };
|
|
case utf8.done => { abort(); };
|
|
case utf8.more => { abort(); };
|
|
};
|
|
};
|
|
|
|
// ---- remaining / slice / position -------------------------------------
|
|
// ref/hare/encoding/utf8/decode.ha:172-198.
|
|
|
|
@test fn remaining_slice_position() void = {
|
|
let src: [16]u8;
|
|
src[0] = 0xE3u8; src[1] = 0x81u8; src[2] = 0x93u8; // こ
|
|
src[3] = 0xE3u8; src[4] = 0x82u8; src[5] = 0x93u8; // ん
|
|
src[6] = 0xE3u8; src[7] = 0x81u8; src[8] = 0xABu8; // に
|
|
src[9] = 0xE3u8; src[10] = 0x81u8; src[11] = 0xA1u8; // ち
|
|
src[12] = 0xE3u8; src[13] = 0x81u8; src[14] = 0xAFu8; // は
|
|
src[15] = 0u8;
|
|
// Two decode() calls instead of `let d2 = d1` per Hare style —
|
|
// let-copy on struct >8B is a Class A cstage+wwstage miscompile
|
|
// (task #32; rhs ident is dropped, slot zero-inits). Both calls
|
|
// take the same src[0:16] slice so d1.src.ptr == d2.src.ptr,
|
|
// matching Hare's `d2 = d1` for slice()'s same-source precondition.
|
|
let d1: utf8.decoder = utf8.decode(src[0:16]);
|
|
let d2: utf8.decoder = utf8.decode(src[0:16]);
|
|
|
|
assert(!(!bytes.equal(utf8.remaining(&d1), src[0:16])));
|
|
assert(!(utf8.slice(&d1, &d2).len != 0));
|
|
assert(!(utf8.slice(&d2, &d1).len != 0));
|
|
assert(!(utf8.position(&d1) != 0));
|
|
|
|
let i: i32 = 0;
|
|
for (i < 2) {
|
|
match (utf8.next(&d1)) { case let r: rune => void; case => { abort(); }; };
|
|
match (utf8.next(&d2)) { case let r: rune => void; case => { abort(); }; };
|
|
i += 1;
|
|
};
|
|
assert(!(utf8.position(&d1) != 6));
|
|
assert(!(!bytes.equal(utf8.remaining(&d1), src[6:16])));
|
|
assert(!(utf8.slice(&d1, &d2).len != 0));
|
|
|
|
i = 0;
|
|
for (i < 3) {
|
|
match (utf8.next(&d2)) { case let r: rune => void; case => { abort(); }; };
|
|
i += 1;
|
|
};
|
|
assert(!(utf8.position(&d2) != 15));
|
|
assert(!(!bytes.equal(utf8.remaining(&d2), src[15:16])));
|
|
assert(!(!bytes.equal(utf8.slice(&d1, &d2), src[6:15])));
|
|
|
|
i = 0;
|
|
for (i < 3) {
|
|
match (utf8.next(&d1)) { case let r: rune => void; case => { abort(); }; };
|
|
i += 1;
|
|
};
|
|
assert(!(utf8.slice(&d1, &d2).len != 0));
|
|
match (utf8.next(&d1)) { case let r: rune => void; case => { abort(); }; };
|
|
assert(!(utf8.remaining(&d1).len != 0));
|
|
};
|
|
|
|
// ---- strerror: constant rendering -------------------------------------
|
|
// ref/hare/encoding/utf8/types.ha:12. `invalid` is payload-free; the
|
|
// renderer collapses to a single fixture row.
|
|
|
|
@test fn strerror_cases() void = {
|
|
let e: utf8.invalid;
|
|
assert(!(!streq(utf8.strerror(e), "Invalid UTF-8")));
|
|
};
|
|
|
|
// ---- round-trip: encode → decode → equal rune --------------------------
|
|
|
|
@test fn roundtrip() void = {
|
|
let runes: [4]u32;
|
|
runes[0] = 0x41u32;
|
|
runes[1] = 0xE9u32;
|
|
runes[2] = 0x20ACu32;
|
|
runes[3] = 0x1F980u32;
|
|
let i: i32 = 0;
|
|
for (i < 4) {
|
|
let buf: [4]u8;
|
|
let n: i32 = utf8.encoderune(buf[0:4], runes[i]: rune);
|
|
let d: utf8.decoder = utf8.decode(buf[0:n]);
|
|
match (utf8.next(&d)) {
|
|
case let r: rune => {
|
|
assert(!((r: u32) != runes[i]));
|
|
};
|
|
case utf8.done => { abort(); };
|
|
case utf8.more => { abort(); };
|
|
case let e: utf8.invalid => { abort(); };
|
|
};
|
|
i += 1;
|
|
};
|
|
};
|