authorgravatar for r00ster91@proton.meWooster <r00ster91@proton.me> 2022-08-18 18:54:51+02:00
committergravatar for noreply@github.comGitHub <noreply@github.com> 2022-08-18 19:54:51+03:00
log4055e6055b5c0727f5e85a30760c9bccf525f4de
tree3295ef3babc22815422ced1f692bc98c9a43972d
parent2523b443a482d78fd1a31f5e6b0224fec03835e3
signaturebadge-question-mark Signed by PGP key 4AEE18F83AFDEB23

AstGen: disallow leading zeroes in int literals and int types

This makes `0123` and `u0123` etc. illegal. I'm now confident that this is a good change because I actually caught two C header translation mistakes in `haiku.zig` with this. Clearly, `0123` being octal in C (TIL) can cause confusion, and we make this easier to read by requiring `0o` as the prefix and now also disallowing leading zeroes in integers. For consistency and because it looks weird, we disallow it for integer types too (e.g. `u0123`). Fixes #11963 Fixes #12417

6 files changed, 74 insertions(+), 27 deletions(-)

lib/std/c/haiku.zig+2-2
......@@ -702,7 +702,7 @@ pub const T = struct {
702702 pub const CSETAF = 0x8002;
703703 pub const CSETAW = 0x8003;
704704 pub const CWAITEVENT = 0x8004;
705 pub const CSBRK = 08005;
705 pub const CSBRK = 0x8005;
706706 pub const CFLSH = 0x8006;
707707 pub const CXONC = 0x8007;
708708 pub const CQUERYCONNECTED = 0x8008;
......@@ -874,7 +874,7 @@ pub const S = struct {
874874 pub const IFDIR = 0o040000;
875875 pub const IFCHR = 0o020000;
876876 pub const IFIFO = 0o010000;
877 pub const INDEX_DIR = 04000000000;
877 pub const INDEX_DIR = 0o4000000000;
878878
879879 pub const IUMSK = 0o7777;
880880 pub const ISUID = 0o4000;
lib/std/zig/parser_test.zig+8-8
......@@ -4254,10 +4254,10 @@ test "zig fmt: integer literals with underscore separators" {
42544254 \\const
42554255 \\ x =
42564256 \\ 1_234_567
4257 \\ + (0b0_1-0o7_0+0xff_FF ) + 0_0;
4257 \\ + (0b0_1-0o7_0+0xff_FF ) + 1_0;
42584258 ,
42594259 \\const x =
4260 \\ 1_234_567 + (0b0_1 - 0o7_0 + 0xff_FF) + 0_0;
4260 \\ 1_234_567 + (0b0_1 - 0o7_0 + 0xff_FF) + 1_0;
42614261 \\
42624262 );
42634263}
......@@ -4266,7 +4266,7 @@ test "zig fmt: hex literals with underscore separators" {
42664266 try testTransform(
42674267 \\pub fn orMask(a: [ 1_000 ]u64, b: [ 1_000] u64) [1_000]u64 {
42684268 \\ var c: [1_000]u64 = [1]u64{ 0xFFFF_FFFF_FFFF_FFFF}**1_000;
4269 \\ for (c [ 0_0 .. ]) |_, i| {
4269 \\ for (c [ 1_0 .. ]) |_, i| {
42704270 \\ c[i] = (a[i] | b[i]) & 0xCCAA_CCAA_CCAA_CCAA;
42714271 \\ }
42724272 \\ return c;
......@@ -4276,7 +4276,7 @@ test "zig fmt: hex literals with underscore separators" {
42764276 ,
42774277 \\pub fn orMask(a: [1_000]u64, b: [1_000]u64) [1_000]u64 {
42784278 \\ var c: [1_000]u64 = [1]u64{0xFFFF_FFFF_FFFF_FFFF} ** 1_000;
4279 \\ for (c[0_0..]) |_, i| {
4279 \\ for (c[1_0..]) |_, i| {
42804280 \\ c[i] = (a[i] | b[i]) & 0xCCAA_CCAA_CCAA_CCAA;
42814281 \\ }
42824282 \\ return c;
......@@ -4288,14 +4288,14 @@ test "zig fmt: hex literals with underscore separators" {
42884288test "zig fmt: decimal float literals with underscore separators" {
42894289 try testTransform(
42904290 \\pub fn main() void {
4291 \\ const a:f64=(10.0e-0+(10.0e+0))+10_00.00_00e-2+00_00.00_10e+4;
4292 \\ const b:f64=010.0--0_10.0+0_1_0.0_0+1e2;
4291 \\ const a:f64=(10.0e-0+(10.0e+0))+10_00.00_00e-2+20_00.00_10e+4;
4292 \\ const b:f64=1_0.0--10_10.0+1_0_0.0_0+1e2;
42934293 \\ std.debug.warn("a: {}, b: {} -> a+b: {}\n", .{ a, b, a + b });
42944294 \\}
42954295 ,
42964296 \\pub fn main() void {
4297 \\ const a: f64 = (10.0e-0 + (10.0e+0)) + 10_00.00_00e-2 + 00_00.00_10e+4;
4298 \\ const b: f64 = 010.0 - -0_10.0 + 0_1_0.0_0 + 1e2;
4297 \\ const a: f64 = (10.0e-0 + (10.0e+0)) + 10_00.00_00e-2 + 20_00.00_10e+4;
4298 \\ const b: f64 = 1_0.0 - -10_10.0 + 1_0_0.0_0 + 1e2;
42994299 \\ std.debug.warn("a: {}, b: {} -> a+b: {}\n", .{ a, b, a + b });
43004300 \\}
43014301 \\
src/AstGen.zig+32-11
......@@ -4088,6 +4088,13 @@ fn testDecl(
40884088 true => .signed,
40894089 false => .unsigned,
40904090 };
4091 if (ident_name_raw.len >= 3 and ident_name_raw[1] == '0') {
4092 return astgen.failTok(
4093 test_name_token,
4094 "primitive integer type '{s}' has leading zero",
4095 .{ident_name_raw},
4096 );
4097 }
40914098 _ = parseBitCount(ident_name_raw[1..]) catch |err| switch (err) {
40924099 error.Overflow => return astgen.failTok(
40934100 test_name_token,
......@@ -6791,6 +6798,13 @@ fn identifier(
67916798 true => .signed,
67926799 false => .unsigned,
67936800 };
6801 if (ident_name_raw.len >= 3 and ident_name_raw[1] == '0') {
6802 return astgen.failNode(
6803 ident,
6804 "primitive integer type '{s}' has leading zero",
6805 .{ident_name_raw},
6806 );
6807 }
67946808 const bit_count = parseBitCount(ident_name_raw[1..]) catch |err| switch (err) {
67956809 error.Overflow => return astgen.failNode(
67966810 ident,
......@@ -7021,17 +7035,6 @@ fn integerLiteral(gz: *GenZir, rl: ResultLoc, node: Ast.Node.Index) InnerError!Z
70217035 const main_tokens = tree.nodes.items(.main_token);
70227036 const int_token = main_tokens[node];
70237037 const prefixed_bytes = tree.tokenSlice(int_token);
7024 if (std.fmt.parseInt(u64, prefixed_bytes, 0)) |small_int| {
7025 const result: Zir.Inst.Ref = switch (small_int) {
7026 0 => .zero,
7027 1 => .one,
7028 else => try gz.addInt(small_int),
7029 };
7030 return rvalue(gz, rl, result, node);
7031 } else |err| switch (err) {
7032 error.InvalidCharacter => unreachable, // Caught by the parser.
7033 error.Overflow => {},
7034 }
70357038
70367039 var base: u8 = 10;
70377040 var non_prefixed: []const u8 = prefixed_bytes;
......@@ -7046,6 +7049,24 @@ fn integerLiteral(gz: *GenZir, rl: ResultLoc, node: Ast.Node.Index) InnerError!Z
70467049 non_prefixed = prefixed_bytes[2..];
70477050 }
70487051
7052 if (base == 10 and prefixed_bytes.len >= 2 and prefixed_bytes[0] == '0') {
7053 return astgen.failNodeNotes(node, "integer literal '{s}' has leading zero", .{prefixed_bytes}, &.{
7054 try astgen.errNoteNode(node, "use '0o' prefix for octal literals", .{}),
7055 });
7056 }
7057
7058 if (std.fmt.parseUnsigned(u64, non_prefixed, base)) |small_int| {
7059 const result: Zir.Inst.Ref = switch (small_int) {
7060 0 => .zero,
7061 1 => .one,
7062 else => try gz.addInt(small_int),
7063 };
7064 return rvalue(gz, rl, result, node);
7065 } else |err| switch (err) {
7066 error.InvalidCharacter => unreachable, // Caught by the parser.
7067 error.Overflow => {},
7068 }
7069
70497070 const gpa = astgen.gpa;
70507071 var big_int = try std.math.big.int.Managed.init(gpa);
70517072 defer big_int.deinit();
test/behavior/math.zig+3-4
......@@ -239,10 +239,9 @@ test "quad hex float literal parsing in range" {
239239}
240240
241241test "underscore separator parsing" {
242 try expect(0_0_0_0 == 0);
243242 try expect(1_234_567 == 1234567);
244 try expect(001_234_567 == 1234567);
245 try expect(0_0_1_2_3_4_5_6_7 == 1234567);
243 try expect(1_234_567 == 1234567);
244 try expect(1_2_3_4_5_6_7 == 1234567);
246245
247246 try expect(0b0_0_0_0 == 0);
248247 try expect(0b1010_1010 == 0b10101010);
......@@ -260,7 +259,7 @@ test "underscore separator parsing" {
260259 try expect(0x1_0_1_0_1_0_1_0 == 0x10101010);
261260
262261 try expect(123_456.789_000e1_0 == 123456.789000e10);
263 try expect(0_1_2_3_4_5_6.7_8_9_0_0_0e0_0_1_0 == 123456.789000e10);
262 try expect(1_2_3_4_5_6.7_8_9_0_0_0e0_0_1_0 == 123456.789000e10);
264263
265264 try expect(0x1234_5678.9ABC_DEF0p-1_0 == 0x12345678.9ABCDEF0p-10);
266265 try expect(0x1_2_3_4_5_6_7_8.9_A_B_C_D_E_F_0p-0_0_0_1_0 == 0x12345678.9ABCDEF0p-10);
test/cases/compile_errors/invalid_underscore_placement_in_int_literal-1.zig+2-2
......@@ -1,5 +1,5 @@
11fn main() void {
2 var bad: u128 = 0010_;
2 var bad: u128 = 10_;
33 _ = bad;
44}
55
......@@ -8,4 +8,4 @@ fn main() void {
88// target=native
99//
1010// :2:21: error: expected expression, found 'invalid bytes'
11// :2:26: note: invalid byte: ';'
11// :2:24: note: invalid byte: ';'
test/cases/compile_errors/leading_zero_in_integer.zig created+27
......@@ -0,0 +1,27 @@
1export fn entry1() void {
2 const T = u000123;
3 _ = T;
4}
5export fn entry2() void {
6 _ = i0;
7 _ = u0;
8 var x: i01 = 1;
9 _ = x;
10}
11export fn entry3() void {
12 _ = 000123;
13}
14export fn entry4() void {
15 _ = 01;
16}
17
18// error
19// backend=llvm
20// target=native
21//
22// :2:15: error: primitive integer type 'u000123' has leading zero
23// :8:12: error: primitive integer type 'i01' has leading zero
24// :12:9: error: integer literal '000123' has leading zero
25// :12:9: note: use '0o' prefix for octal literals
26// :15:9: error: integer literal '01' has leading zero
27// :15:9: note: use '0o' prefix for octal literals