libcxx/src/support/ibm/mbsnrtowcs.cpp

   1 //===----------------------------------------------------------------------===//
   2 //
   3 // Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
   4 // See https://llvm.org/LICENSE.txt for license information.
   5 // SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
   6 //
   7 //===----------------------------------------------------------------------===//
   8
   9 #include <cstddef> // size_t
  10 #include <cwchar>  // mbstate_t
  11 #include <limits.h> // MB_LEN_MAX
  12 #include <string.h> // wmemcpy
  13
  14 // Returns the number of wide characters found in the multi byte sequence `src`
  15 // (of `src_size_bytes`), that fit in the buffer `dst` (of `max_dest_chars`
  16 // elements size). The count returned excludes the null terminator.
  17 // When `dst` is NULL, no characters are copied to `dst`.
  18 // Returns (size_t) -1 when an invalid sequence is encountered.
  19 // Leaves *`src` pointing to the next character to convert or NULL
  20 // if a null character was converted from *`src`.
  21 _LIBCPP_EXPORTED_FROM_ABI
  22 size_t mbsnrtowcs(wchar_t *__restrict dst, const char **__restrict src,
  23                    size_t src_size_bytes, size_t max_dest_chars,
  24                    mbstate_t *__restrict ps) {
  25   const size_t terminated_sequence = static_cast<size_t>(0);
  26   const size_t invalid_sequence = static_cast<size_t>(-1);
  27   const size_t incomplete_sequence = static_cast<size_t>(-2);
  28
  29   size_t source_converted;
  30   size_t dest_converted;
  31   size_t result = 0;
  32
  33   // If `dst` is null then `max_dest_chars` should be ignored according to the
  34   // standard. Setting `max_dest_chars` to a large value has this effect.
  35   if (dst == nullptr)
  36     max_dest_chars = static_cast<size_t>(-1);
  37
  38   for (dest_converted = source_converted = 0;
  39        source_converted < src_size_bytes && (!dst || dest_converted < max_dest_chars);
  40        ++dest_converted, source_converted += result) {
  41     // Converts one multi byte character.
  42     // If result (char_size) is greater than 0, it's the size in bytes of that character.
  43     // If result (char_size) is zero, it indicates that the null character has been found.
  44     // Otherwise, it's an error and errno may be set.
  45     size_t source_remaining = src_size_bytes - source_converted;
  46     size_t dest_remaining = max_dest_chars - dest_converted;
  47
  48     if (dst == nullptr) {
  49       result = mbrtowc(NULL, *src + source_converted, source_remaining, ps);
  50     } else if (dest_remaining >= source_remaining) {
  51       // dst has enough space to translate in-place.
  52       result = mbrtowc(dst + dest_converted, *src + source_converted, source_remaining, ps);
  53     } else {
  54       /*
  55       * dst may not have enough space, so use a temporary buffer.
  56       *
  57       * We need to save a copy of the conversion state
  58       * here so we can restore it if the multibyte
  59       * character is too long for the buffer.
  60       */
  61       wchar_t buff[MB_LEN_MAX];
  62       mbstate_t mbstate_tmp;
  63
  64       if (ps != nullptr)
  65         mbstate_tmp = *ps;
  66       result = mbrtowc(buff, *src + source_converted, source_remaining, ps);
  67
  68       if (result > dest_remaining) {
  69         // Multi-byte sequence for character won't fit.
  70         if (ps != nullptr)
  71           *ps = mbstate_tmp;
  72         break;
  73       } else {
  74         // The buffer was used, so we need copy the translation to dst.
  75         wmemcpy(dst, buff, result);
  76       }
  77     }
  78
  79     // Don't do anything to change errno from here on.
  80     if (result == invalid_sequence || result == terminated_sequence || result == incomplete_sequence) {
  81       break;
  82     }
  83   }
  84
  85   if (dst) {
  86     if (result == terminated_sequence)
  87       *src = NULL;
  88     else
  89       *src += source_converted;
  90   }
  91   if (result == invalid_sequence)
  92     return invalid_sequence;
  93
  94   return dest_converted;
  95 }