lib/bytes+test: port ltrim/rtrim/trim from Hare (u8... variadic)

lib/bytes/bytes.ww

@@ -142,6 +142,40 @@ export fn contains(s: []u8, needles: (u8 | []u8)...) bool = {
 	return false;
 };
 
+// ltrim — borrowed view of `in` with leading bytes in `trim` stripped.
+// `trim` must be non-empty. ref/hare/bytes/trim.ha:7.
+export fn ltrim(in: []u8, trim: u8...) []u8 = {
+	os.assert(trim.len > 0, "bytes.ltrim called with empty trim set");
+	let i: i32 = 0;
+	for (i < in.len && contains(trim, in[i])) { i += 1; };
+	let r: []u8;
+	r.ptr = in.ptr + (i: u64);
+	r.len = in.len - i;
+	r.cap = r.len;
+	return r;
+};
+
+// rtrim — borrowed view of `in` with trailing bytes in `trim` stripped.
+// `trim` must be non-empty. ref/hare/bytes/trim.ha:17. Hare's loop uses
+// `size` underflow at i==0 to terminate; ww indices are signed i32, so
+// the equivalent termination is spelled `i >= 0` explicitly.
+export fn rtrim(in: []u8, trim: u8...) []u8 = {
+	os.assert(trim.len > 0, "bytes.rtrim called with empty trim set");
+	let i: i32 = in.len - 1;
+	for (i >= 0 && contains(trim, in[i])) { i -= 1; };
+	let r: []u8;
+	r.ptr = in.ptr;
+	r.len = i + 1;
+	r.cap = r.len;
+	return r;
+};
+
+// trim — borrowed view of `in` with both ends in `trim` stripped.
+// ref/hare/bytes/trim.ha:27.
+export fn trim(in: []u8, trim: u8...) []u8 = {
+	return ltrim(rtrim(in, trim...), trim...);
+};
+
 // hasprefix — true iff `s` starts with `pre`.
 // ref/hare/bytes/contains.ha:21.
 export fn hasprefix(s: []u8, pre: []u8) bool = {

lib/bytes/bytestest.ww

@@ -467,6 +467,107 @@ fn expect_done(t: *bytes.tokenizer) void = {
 	if (!bytes.equal(r2, e_12[0:2])) { fail(); };
 };
 
+// ---- ltrim / rtrim / trim ---------------------------------------------
+// ref/hare/bytes/trim.ha:29 — Hare's @test fn trim pins
+// `trim([0,1,2,3,5,0], 0) == [1,2,3,5]`, `trim([0,0,0], 0) == []`,
+// `trim([], 0) == []`. ww spreads the matrix across ltrim/rtrim/trim.
+
+fn beq(got: []u8, want: []u8) bool = {
+	if (got.len != want.len) { return false; };
+	let i: i32 = 0;
+	for (i < got.len) {
+		if (got[i] != want[i]) { return false; };
+		i += 1;
+	};
+	return true;
+};
+
+@test fn ltrim_cases() void = {
+	let z: [1]u8;
+
+	// [0,0,1,2] / 0 -> [1,2]
+	signalled = 1780;
+	let a: [4]u8; a[0] = 0u8; a[1] = 0u8; a[2] = 1u8; a[3] = 2u8;
+	let ea: [2]u8; ea[0] = 1u8; ea[1] = 2u8;
+	if (!beq(bytes.ltrim(a[0:4], 0u8), ea[0:2])) { fail(); };
+
+	// [1,2,3] / 0 -> [1,2,3] (no leading match)
+	signalled = 1781;
+	let b: [3]u8; b[0] = 1u8; b[1] = 2u8; b[2] = 3u8;
+	if (!beq(bytes.ltrim(b[0:3], 0u8), b[0:3])) { fail(); };
+
+	// [0,0,0] / 0 -> [] (full match)
+	signalled = 1782;
+	let c: [3]u8;
+	if (!beq(bytes.ltrim(c[0:3], 0u8), z[0:0])) { fail(); };
+
+	// [] / 0 -> [] (empty input)
+	signalled = 1783;
+	if (!beq(bytes.ltrim(z[0:0], 0u8), z[0:0])) { fail(); };
+};
+
+@test fn rtrim_cases() void = {
+	let z: [1]u8;
+
+	// [1,2,0,0] / 0 -> [1,2]
+	signalled = 1790;
+	let a: [4]u8; a[0] = 1u8; a[1] = 2u8; a[2] = 0u8; a[3] = 0u8;
+	let ea: [2]u8; ea[0] = 1u8; ea[1] = 2u8;
+	if (!beq(bytes.rtrim(a[0:4], 0u8), ea[0:2])) { fail(); };
+
+	// [1,2,3] / 0 -> [1,2,3] (no trailing match)
+	signalled = 1791;
+	let b: [3]u8; b[0] = 1u8; b[1] = 2u8; b[2] = 3u8;
+	if (!beq(bytes.rtrim(b[0:3], 0u8), b[0:3])) { fail(); };
+
+	// [0,0,0] / 0 -> [] (full match)
+	signalled = 1792;
+	let c: [3]u8;
+	if (!beq(bytes.rtrim(c[0:3], 0u8), z[0:0])) { fail(); };
+
+	// [] / 0 -> [] (empty input)
+	signalled = 1793;
+	if (!beq(bytes.rtrim(z[0:0], 0u8), z[0:0])) { fail(); };
+};
+
+@test fn trim_cases() void = {
+	let z: [1]u8;
+
+	// [0,1,2,3,5,0] / 0 -> [1,2,3,5]
+	signalled = 1800;
+	let a: [6]u8;
+	a[0] = 0u8; a[1] = 1u8; a[2] = 2u8;
+	a[3] = 3u8; a[4] = 5u8; a[5] = 0u8;
+	let ea: [4]u8; ea[0] = 1u8; ea[1] = 2u8; ea[2] = 3u8; ea[3] = 5u8;
+	if (!beq(bytes.trim(a[0:6], 0u8), ea[0:4])) { fail(); };
+
+	// [0,5,0] / 5 -> [0,5,0] (only 5 in trim set; boundary mismatch)
+	signalled = 1801;
+	let b: [3]u8; b[0] = 0u8; b[1] = 5u8; b[2] = 0u8;
+	if (!beq(bytes.trim(b[0:3], 5u8), b[0:3])) { fail(); };
+
+	// [0,1,42,1,0] / {0,42} -> [1,42,1] (multi-byte trim set)
+	signalled = 1802;
+	let c: [5]u8;
+	c[0] = 0u8; c[1] = 1u8; c[2] = 42u8; c[3] = 1u8; c[4] = 0u8;
+	let ec: [3]u8; ec[0] = 1u8; ec[1] = 42u8; ec[2] = 1u8;
+	if (!beq(bytes.trim(c[0:5], 0u8, 42u8), ec[0:3])) { fail(); };
+
+	// [0,0,0] / 0 -> [] (full match)
+	signalled = 1803;
+	let d: [3]u8;
+	if (!beq(bytes.trim(d[0:3], 0u8), z[0:0])) { fail(); };
+
+	// [] / 0 -> [] (empty input, Hare ref/hare/bytes/trim.ha:34)
+	signalled = 1804;
+	if (!beq(bytes.trim(z[0:0], 0u8), z[0:0])) { fail(); };
+
+	// [1,2,3,5] / 0 -> [1,2,3,5] (Hare ref/hare/bytes/trim.ha:31)
+	signalled = 1805;
+	let e: [4]u8; e[0] = 1u8; e[1] = 2u8; e[2] = 3u8; e[3] = 5u8;
+	if (!beq(bytes.trim(e[0:4], 0u8), e[0:4])) { fail(); };
+};
+
 // ---- splitn / rsplitn / split -----------------------------------------
 // ref/hare/bytes/tokenize.ha:330 (@test fn split). Hare's table mixes
 // strings; ww spells the vectors as byte arrays explicitly.
@@ -662,5 +763,8 @@ export fn main() i32 = {
 	signalled = 13; splitn_cases();
 	signalled = 14; rsplitn_cases();
 	signalled = 15; split_cases();
+	signalled = 16; ltrim_cases();
+	signalled = 17; rtrim_cases();
+	signalled = 18; trim_cases();
 	return 0;
 };

selfhost/cmd/w6c/main.combined.ww

@@ -1023,6 +1023,40 @@ export fn contains(s: []u8, needles: (u8 | []u8)...) bool = {
 	return false;
 };
 
+// ltrim — borrowed view of `in` with leading bytes in `trim` stripped.
+// `trim` must be non-empty. ref/hare/bytes/trim.ha:7.
+export fn ltrim(in: []u8, trim: u8...) []u8 = {
+	os.assert(trim.len > 0, "bytes.ltrim called with empty trim set");
+	let i: i32 = 0;
+	for (i < in.len && contains(trim, in[i])) { i += 1; };
+	let r: []u8;
+	r.ptr = in.ptr + (i: u64);
+	r.len = in.len - i;
+	r.cap = r.len;
+	return r;
+};
+
+// rtrim — borrowed view of `in` with trailing bytes in `trim` stripped.
+// `trim` must be non-empty. ref/hare/bytes/trim.ha:17. Hare's loop uses
+// `size` underflow at i==0 to terminate; ww indices are signed i32, so
+// the equivalent termination is spelled `i >= 0` explicitly.
+export fn rtrim(in: []u8, trim: u8...) []u8 = {
+	os.assert(trim.len > 0, "bytes.rtrim called with empty trim set");
+	let i: i32 = in.len - 1;
+	for (i >= 0 && contains(trim, in[i])) { i -= 1; };
+	let r: []u8;
+	r.ptr = in.ptr;
+	r.len = i + 1;
+	r.cap = r.len;
+	return r;
+};
+
+// trim — borrowed view of `in` with both ends in `trim` stripped.
+// ref/hare/bytes/trim.ha:27.
+export fn trim(in: []u8, trim: u8...) []u8 = {
+	return ltrim(rtrim(in, trim...), trim...);
+};
+
 // hasprefix — true iff `s` starts with `pre`.
 // ref/hare/bytes/contains.ha:21.
 export fn hasprefix(s: []u8, pre: []u8) bool = {

selfhost/cmd/wwdump/main.combined.ww

@@ -1023,6 +1023,40 @@ export fn contains(s: []u8, needles: (u8 | []u8)...) bool = {
 	return false;
 };
 
+// ltrim — borrowed view of `in` with leading bytes in `trim` stripped.
+// `trim` must be non-empty. ref/hare/bytes/trim.ha:7.
+export fn ltrim(in: []u8, trim: u8...) []u8 = {
+	os.assert(trim.len > 0, "bytes.ltrim called with empty trim set");
+	let i: i32 = 0;
+	for (i < in.len && contains(trim, in[i])) { i += 1; };
+	let r: []u8;
+	r.ptr = in.ptr + (i: u64);
+	r.len = in.len - i;
+	r.cap = r.len;
+	return r;
+};
+
+// rtrim — borrowed view of `in` with trailing bytes in `trim` stripped.
+// `trim` must be non-empty. ref/hare/bytes/trim.ha:17. Hare's loop uses
+// `size` underflow at i==0 to terminate; ww indices are signed i32, so
+// the equivalent termination is spelled `i >= 0` explicitly.
+export fn rtrim(in: []u8, trim: u8...) []u8 = {
+	os.assert(trim.len > 0, "bytes.rtrim called with empty trim set");
+	let i: i32 = in.len - 1;
+	for (i >= 0 && contains(trim, in[i])) { i -= 1; };
+	let r: []u8;
+	r.ptr = in.ptr;
+	r.len = i + 1;
+	r.cap = r.len;
+	return r;
+};
+
+// trim — borrowed view of `in` with both ends in `trim` stripped.
+// ref/hare/bytes/trim.ha:27.
+export fn trim(in: []u8, trim: u8...) []u8 = {
+	return ltrim(rtrim(in, trim...), trim...);
+};
+
 // hasprefix — true iff `s` starts with `pre`.
 // ref/hare/bytes/contains.ha:21.
 export fn hasprefix(s: []u8, pre: []u8) bool = {

selfhost/test/smoke.combined.ww

@@ -914,6 +914,40 @@ export fn contains(s: []u8, needles: (u8 | []u8)...) bool = {
 	return false;
 };
 
+// ltrim — borrowed view of `in` with leading bytes in `trim` stripped.
+// `trim` must be non-empty. ref/hare/bytes/trim.ha:7.
+export fn ltrim(in: []u8, trim: u8...) []u8 = {
+	os.assert(trim.len > 0, "bytes.ltrim called with empty trim set");
+	let i: i32 = 0;
+	for (i < in.len && contains(trim, in[i])) { i += 1; };
+	let r: []u8;
+	r.ptr = in.ptr + (i: u64);
+	r.len = in.len - i;
+	r.cap = r.len;
+	return r;
+};
+
+// rtrim — borrowed view of `in` with trailing bytes in `trim` stripped.
+// `trim` must be non-empty. ref/hare/bytes/trim.ha:17. Hare's loop uses
+// `size` underflow at i==0 to terminate; ww indices are signed i32, so
+// the equivalent termination is spelled `i >= 0` explicitly.
+export fn rtrim(in: []u8, trim: u8...) []u8 = {
+	os.assert(trim.len > 0, "bytes.rtrim called with empty trim set");
+	let i: i32 = in.len - 1;
+	for (i >= 0 && contains(trim, in[i])) { i -= 1; };
+	let r: []u8;
+	r.ptr = in.ptr;
+	r.len = i + 1;
+	r.cap = r.len;
+	return r;
+};
+
+// trim — borrowed view of `in` with both ends in `trim` stripped.
+// ref/hare/bytes/trim.ha:27.
+export fn trim(in: []u8, trim: u8...) []u8 = {
+	return ltrim(rtrim(in, trim...), trim...);
+};
+
 // hasprefix — true iff `s` starts with `pre`.
 // ref/hare/bytes/contains.ha:21.
 export fn hasprefix(s: []u8, pre: []u8) bool = {