diff options
| author | Evan Haas <evan@lagerdata.com> | 2022-09-08 20:19:10 -0700 |
|---|---|---|
| committer | Veikka Tuominen <git@vexu.eu> | 2022-09-10 01:50:17 +0300 |
| commit | 8e631ee3e7b4e7b4466c0efafaffb4151447785f (patch) | |
| tree | 1873bad537882c07d240b8ceed55e0221f81705c /src | |
| parent | 9e070b653c89a9216f9dd9f78ed7c78c11460ac7 (diff) | |
| download | zig-8e631ee3e7b4e7b4466c0efafaffb4151447785f.tar.gz zig-8e631ee3e7b4e7b4466c0efafaffb4151447785f.zip | |
translate-c: Escape non-ASCII characters that appear in macros
Macro definitions are simply a slice of bytes, which may not be
UTF-8 encoded. If they are not UTF-8 encoded, escape non-printable
and non-ASCII characters as `\xNN`.
Fixes #12784
Diffstat (limited to 'src')
| -rw-r--r-- | src/translate_c.zig | 20 |
1 files changed, 18 insertions, 2 deletions
diff --git a/src/translate_c.zig b/src/translate_c.zig index 014f6b1934..f969bf1c8b 100644 --- a/src/translate_c.zig +++ b/src/translate_c.zig @@ -5957,20 +5957,36 @@ fn zigifyEscapeSequences(ctx: *Context, m: *MacroCtx) ![]const u8 { return bytes[0..i]; } +/// non-ASCII characters (c > 127) are also treated as non-printable by fmtSliceEscapeLower. +/// If a C string literal or char literal in a macro is not valid UTF-8, we need to escape +/// non-ASCII characters so that the Zig source we output will itself be UTF-8. +fn escapeUnprintables(ctx: *Context, m: *MacroCtx) ![]const u8 { + const zigified = try zigifyEscapeSequences(ctx, m); + if (std.unicode.utf8ValidateSlice(zigified)) return zigified; + + const formatter = std.fmt.fmtSliceEscapeLower(zigified); + const encoded_size = @intCast(usize, std.fmt.count("{s}", .{formatter})); + var output = try ctx.arena.alloc(u8, encoded_size); + return std.fmt.bufPrint(output, "{s}", .{formatter}) catch |err| switch (err) { + error.NoSpaceLeft => unreachable, + else => |e| return e, + }; +} + fn parseCPrimaryExprInner(c: *Context, m: *MacroCtx, scope: *Scope) ParseError!Node { const tok = m.next().?; const slice = m.slice(); switch (tok) { .CharLiteral => { if (slice[0] != '\'' or slice[1] == '\\' or slice.len == 3) { - return Tag.char_literal.create(c.arena, try zigifyEscapeSequences(c, m)); + return Tag.char_literal.create(c.arena, try escapeUnprintables(c, m)); } else { const str = try std.fmt.allocPrint(c.arena, "0x{s}", .{std.fmt.fmtSliceHexLower(slice[1 .. slice.len - 1])}); return Tag.integer_literal.create(c.arena, str); } }, .StringLiteral => { - return Tag.string_literal.create(c.arena, try zigifyEscapeSequences(c, m)); + return Tag.string_literal.create(c.arena, try escapeUnprintables(c, m)); }, .IntegerLiteral, .FloatLiteral => { return parseCNumLit(c, m); |
