Commit 3b622f4494b8fd899abc20e75e46726344f2d20c | zig

author	Andrew Kelley <andrew@ziglang.org> 2020-02-09 14:56:19-05:00
committer	Andrew Kelley <andrew@ziglang.org> 2020-02-09 14:56:19-05:00
log	3b622f4494b8fd899abc20e75e46726344f2d20c
tree	4110aaab9e8748020861c0ff7d4211314d079e0c
parent	c4d0f97b4c3f85796c4953d6e87b5c0186616bc1
signature	Commit is signed but in an unrecognized format.

fix off-by-one error in std.unicode.utf8ToUtf16LeWithNull

and fix larger-than-one-byte sentinels when being freed

Thank you to João Pedro for identifying both problems and
providing example code to solve them.

closes #4413

2 files changed, 7 insertions(+), 4 deletions(-)

lib
- std
  - mem.zig
  - unicode.zig

lib/std/mem.zig+1-1

...	@@ -233,7 +233,7 @@ pub const Allocator = struct {	...	@@ -233,7 +233,7 @@ pub const Allocator = struct {
233	pub fn free(self: *Allocator, memory: var) void {	233	pub fn free(self: *Allocator, memory: var) void {
234	const Slice = @typeInfo(@TypeOf(memory)).Pointer;	234	const Slice = @typeInfo(@TypeOf(memory)).Pointer;
235	const bytes = @sliceToBytes(memory);	235	const bytes = @sliceToBytes(memory);
236	const bytes_len = bytes.len + @boolToInt(Slice.sentinel != null);	236	const bytes_len = bytes.len + if (Slice.sentinel != null) @sizeOf(Slice.child) else 0;
237	if (bytes_len == 0) return;	237	if (bytes_len == 0) return;
238	const non_const_ptr = @intToPtr([*]u8, @ptrToInt(bytes.ptr));	238	const non_const_ptr = @intToPtr([*]u8, @ptrToInt(bytes.ptr));
239	@memset(non_const_ptr, undefined, bytes_len);	239	@memset(non_const_ptr, undefined, bytes_len);

lib/std/unicode.zig+6-3

 @@ -571,8 +571,9 @@ pub fn utf8ToUtf16LeWithNull(allocator: *mem.Allocator, utf8: []const u8) ![:0]u
         }
     }
+    const len = result.len;
     try result.append(0);
-    return result.toOwnedSlice()[0..:0];
+    return result.toOwnedSlice()[0..len :0];
 }
 /// Returns index of next character. If exact fit, returned index equals output slice length.
 @@ -619,12 +620,14 @@ test "utf8ToUtf16LeWithNull" {
         var bytes: [128]u8 = undefined;
         const allocator = &std.heap.FixedBufferAllocator.init(bytes[0..]).allocator;
         const utf16 = try utf8ToUtf16LeWithNull(allocator, "𐐷");
-        testing.expectEqualSlices(u8, "\x01\xd8\x37\xdc\x00\x00", @sliceToBytes(utf16[0..]));
+        testing.expectEqualSlices(u8, "\x01\xd8\x37\xdc", @sliceToBytes(utf16[0..]));
+        testing.expect(utf16[2] == 0);
     }
     {
         var bytes: [128]u8 = undefined;
         const allocator = &std.heap.FixedBufferAllocator.init(bytes[0..]).allocator;
         const utf16 = try utf8ToUtf16LeWithNull(allocator, "\u{10FFFF}");
-        testing.expectEqualSlices(u8, "\xff\xdb\xff\xdf\x00\x00", @sliceToBytes(utf16[0..]));
+        testing.expectEqualSlices(u8, "\xff\xdb\xff\xdf", @sliceToBytes(utf16[0..]));
+        testing.expect(utf16[2] == 0);
     }
 }