dom/script/ScriptDecoding.h

   1 /* -*- Mode: C++; tab-width: 8; indent-tabs-mode: nil; c-basic-offset: 2 -*- */
   2 /* vim: set ts=8 sts=2 et sw=2 tw=80: */
   3 /* This Source Code Form is subject to the terms of the Mozilla Public
   4  * License, v. 2.0. If a copy of the MPL was not distributed with this
   5  * file, You can obtain one at http://mozilla.org/MPL/2.0/. */
   6
   7 /* Script decoding templates for processing byte data as UTF-8 or UTF-16. */
   8
   9 #ifndef mozilla_dom_ScriptDecoding_h
  10 #define mozilla_dom_ScriptDecoding_h
  11
  12 #include "mozilla/Assertions.h"  // MOZ_ASSERT
  13 #include "mozilla/CheckedInt.h"  // mozilla::CheckedInt
  14 #include "mozilla/Encoding.h"    // mozilla::Decoder
  15 #include "mozilla/Span.h"        // mozilla::Span
  16 #include "mozilla/UniquePtr.h"   // mozilla::UniquePtr
  17
  18 #include <stddef.h>     // size_t
  19 #include <stdint.h>     // uint8_t, uint32_t
  20 #include <type_traits>  // std::is_same
  21
  22 namespace mozilla::dom {
  23
  24 template <typename Unit>
  25 struct ScriptDecoding {
  26   static_assert(std::is_same<Unit, char16_t>::value ||
  27                     std::is_same<Unit, Utf8Unit>::value,
  28                 "must be either UTF-8 or UTF-16");
  29 };
  30
  31 template <>
  32 struct ScriptDecoding<char16_t> {
  33   static CheckedInt<size_t> MaxBufferLength(const UniquePtr<Decoder>& aDecoder,
  34                                             uint32_t aByteLength) {
  35     return aDecoder->MaxUTF16BufferLength(aByteLength);
  36   }
  37
  38   static size_t DecodeInto(const UniquePtr<Decoder>& aDecoder,
  39                            const Span<const uint8_t>& aSrc,
  40                            Span<char16_t> aDest, bool aEndOfSource) {
  41     uint32_t result;
  42     size_t read;
  43     size_t written;
  44     std::tie(result, read, written, std::ignore) =
  45         aDecoder->DecodeToUTF16(aSrc, aDest, aEndOfSource);
  46     MOZ_ASSERT(result == kInputEmpty);
  47     MOZ_ASSERT(read == aSrc.Length());
  48     MOZ_ASSERT(written <= aDest.Length());
  49
  50     return written;
  51   }
  52 };
  53
  54 template <>
  55 struct ScriptDecoding<Utf8Unit> {
  56   static CheckedInt<size_t> MaxBufferLength(const UniquePtr<Decoder>& aDecoder,
  57                                             uint32_t aByteLength) {
  58     return aDecoder->MaxUTF8BufferLength(aByteLength);
  59   }
  60
  61   static size_t DecodeInto(const UniquePtr<Decoder>& aDecoder,
  62                            const Span<const uint8_t>& aSrc,
  63                            Span<Utf8Unit> aDest, bool aEndOfSource) {
  64     uint32_t result;
  65     size_t read;
  66     size_t written;
  67     // Until C++ char8_t happens, our decoder APIs deal in |uint8_t| while
  68     // |Utf8Unit| internally deals with |char|, so there's inevitable impedance
  69     // mismatch between |aDest| as |Utf8Unit| and |AsWritableBytes(aDest)| as
  70     // |Span<uint8_t>|.  :-(  The written memory will be interpreted through
  71     // |char Utf8Unit::mValue| which is *permissible* because any object's
  72     // memory can be interpreted as |char|.  Unfortunately, until
  73     // twos-complement is mandated, we have to play fast and loose and *hope*
  74     // interpreting memory storing |uint8_t| as |char| will pick up the desired
  75     // wrapped-around value.  ¯\_(ツ)_/¯
  76     std::tie(result, read, written, std::ignore) =
  77         aDecoder->DecodeToUTF8(aSrc, AsWritableBytes(aDest), aEndOfSource);
  78     MOZ_ASSERT(result == kInputEmpty);
  79     MOZ_ASSERT(read == aSrc.Length());
  80     MOZ_ASSERT(written <= aDest.Length());
  81
  82     return written;
  83   }
  84 };
  85
  86 }  // namespace mozilla::dom
  87
  88 #endif  // mozilla_dom_ScriptDecoding_h