String.cc

/*
 * Copyright (c) 2008-2014, Pedigree Developers
 *
 * Please see the CONTRIB file in the root of the source tree for a full
 * list of contributors.
 *
 * Permission to use, copy, modify, and distribute this software for any
 * purpose with or without fee is hereby granted, provided that the above
 * copyright notice and this permission notice appear in all copies.
 *
 * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES
 * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
 * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR
 * ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
 * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
 * ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
 * OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
 */
 
#include "pedigree/kernel/Log.h"
#include "pedigree/kernel/utilities/Cord.h"
#include "pedigree/kernel/utilities/String.h"
#include "pedigree/kernel/utilities/StringView.h"
#include "pedigree/kernel/utilities/assert.h"
#include "pedigree/kernel/utilities/utility.h"
 
#include <stdarg.h>
 
static constexpr size_t StringMinimumAllocationSize = 64;
 
String::String() : m_Data(nullptr), m_Length(0), m_Size(0), m_Hash(0) {}
 
String::String(const char* s) : String() {
  assign(s);
}
 
String::String(const char* s, size_t length) : String() {
  assign(s, length);
}
 
String::String(const char* s, size_t length, bool unsafe) : String() {
  assign(s, length, unsafe);
}
 
String::String(const String& x) : String() {
  assign(x);
}
 
String::String(const StringView& x) : String() {
  assign(x.str(), x.length(), true);
}
 
String::String(String&& x) noexcept : String() {
  move(static_cast<String&&>(x));
}
 
String::String(const Cord& x) : String() {
  assign(x);
}
 
String::~String() {
  clear();
}
 
void String::move(String&& other) noexcept {
  clear();
 
  // take ownership of the object
  m_Data = other.m_Data;
  m_Length = other.m_Length;
  m_Size = other.m_Size;
  m_Hash = other.m_Hash;
 
  // free other string but don't destroy the heap pointer if we had one
  // as it is now owned by this new instance
  other.m_Data = 0;
  other.clear();
}
 
String& String::operator=(String&& x) noexcept {
  move(static_cast<String&&>(x));
  return *this;
}
 
String& String::operator=(const String& x) {
  assign(x);
  return *this;
}
 
#if !STRING_DISABLE_EXPENSIVE_COPY_CONSTRUCTION
String& String::operator=(const char* s) {
  assign(s);
  return *this;
}
#endif
 
String& String::operator=(const Cord& x) {
  assign(x);
  return *this;
}
 
String& String::operator+=(const String& x) {
  assert(assignable());
 
  // Empty strings need not have storage, even for a terminating byte.
  if (!x.length())
    return *this;
 
  if (this == &x) {
    // The copy breaks self-append aliasing before the recursive append.
    // NOLINTNEXTLINE(performance-unnecessary-copy-initialization)
    String copy(x);
    return *this += copy;
  }
 
  size_t newLength = x.length() + m_Length;
 
  reserve(newLength + 1);
 
  char* dst = extract();
  const char* src = x.extract();
 
  // Copy!
  MemoryCopy(&dst[m_Length], src, x.length() + 1);
  m_Length += x.length();
 
  m_Hash = 0;  // hash is no longer valid
  return *this;
}
 
String& String::operator+=(const char* s) {
  assert(assignable());
 
  const char* current = extract();
  if (s && current) {
    const uintptr_t address = reinterpret_cast<uintptr_t>(s);
    const uintptr_t begin = reinterpret_cast<uintptr_t>(current);
    if ((address >= begin) && (address < (begin + m_Size))) {
      String copy(s);
      return *this += copy;
    }
  }
 
  size_t slen = StringLength(s);
  size_t newLength = slen + m_Length;
 
  reserve(slen + m_Length + 1);
  MemoryCopy(&m_Data[m_Length], s, slen + 1);
  m_Length += slen;
 
  m_Hash = 0;
  return *this;
}
 
bool String::operator==(const String& s) const {
 
  if (m_Length != s.m_Length) {
    return false;
  } else if (LIKELY(m_Hash && s.maybeHash())) {
    if (m_Hash != s.hash()) {
      // precomputed hash didn't match, don't bother
      return false;
    }
  }
 
  const char* buf = extract();
  const char* other_buf = s.extract();
 
  // Neither of these can be null because of the above conditions.
  return !StringMatchN(buf, other_buf, m_Length);
}
 
bool String::operator==(const StringView& s) const {
  // use StringView::operator==(const String &)
  return s == *this;
}
 
bool String::operator==(const char* s) const {
  const char* buf = extract();
 
  if ((!m_Length) && (s == 0)) {
    return true;
  } else if (s == 0) {
    // m_Length > 0 but other buffer is null.
    return false;
  } else if ((!m_Length) && *s) {
    // Quick check when we're zero-length.
    return false;
  } else if (StringLength(s) != m_Length) {
    return false;
  } else {
    return !StringMatchN(buf, s, m_Length);
  }
}
 
bool String::compare(const char* s, size_t len) const {
  if (m_Length != len) {
    // Mismatch in length
    return false;
  } else if (UNLIKELY(s == 0)) {
    // other buffer is null, don't match
    return false;
  } else {
    const char* buf = extract();
    return !StringMatchN(buf, s, m_Length);
  }
}
 
char String::operator[](size_t i) const {
  assert(i <= m_Length);
  const char* buf = extract();
  return buf[i];
}
 
uint32_t String::hash() const {
  if (!m_Hash) {
    return computeHash();
  }
 
  return m_Hash;
}
 
uint32_t String::hash() {
  if (!m_Hash) {
    computeHash();
  }
 
  return m_Hash;
}
 
uint32_t String::maybeHash() const {
  return m_Hash;
}
 
size_t String::nextCharacter(size_t c) const {
  const char* buf = extract();
  return ::nextCharacter(buf, c);
}
 
size_t String::prevCharacter(size_t c) const {
  const char* buf = extract();
  return ::prevCharacter(buf, c);
}
 
void String::assign(const String& x) {
  assert(assignable());
 
  if (this == &x)
    return;
 
  if (extract() && x.extract()) {
    assert(extract() != x.extract());
  }
 
  reserve(x.size(), false);
  MemoryCopy(m_Data, x.extract(), x.size());
  m_Length = x.length();
 
  // no need to recompute in this case
  m_Hash = x.m_Hash;
 
#if ADDITIONAL_CHECKS
  if (*this != x) {
    ERROR("mismatch: '" << *this << "' != '" << x << "'");
  }
  assert(*this == x);
#endif
}
 
void String::assign(const Cord& x) {
  assert(assignable());
 
  reserve(x.length() + 1);
 
  size_t offset = 0;
  char* buf = extract();
  for (auto& it : x.m_Segments) {
    StringCopyN(buf + offset, it.ptr, it.length);
    offset += it.length;
  }
  buf[offset] = 0;
 
  m_Length = offset;
 
  m_Hash = 0;
}
 
void String::assign(const char* s, size_t len, bool unsafe) {
  assert(assignable());
 
  const char* current = extract();
  if (s && current) {
    const uintptr_t address = reinterpret_cast<uintptr_t>(s);
    const uintptr_t begin = reinterpret_cast<uintptr_t>(current);
    if ((address >= begin) && (address < (begin + m_Size))) {
      String copy(s, len, unsafe);
      assign(copy);
      return;
    }
  }
 
  // Trying to assign self to self?
  assert((m_Data == nullptr) || (m_Data && (m_Data != s)));
 
  size_t copyLength = 0;
  size_t origLength = len;
  // len overrides all other optimizations
  if (len) {
    // Fix up length if the passed string is much smaller than the 'len'
    // parameter (otherwise we think we have a giant string).
    size_t trueLength = 0;
    if (unsafe) {
      trueLength = BoundedStringLength(s, len);
    } else {
      trueLength = StringLength(s);
    }
 
    if (trueLength < len) {
      len = trueLength;
    }
 
    m_Length = len;
    copyLength = len;
  } else if (!s || !*s) {
    m_Length = 0;
  } else {
    m_Length = StringLength(s);
    copyLength = m_Length;
  }
 
  if (!m_Length) {
    delete[] m_Data;
    m_Data = 0;
    m_Size = 0;
  } else {
    reserve(pedigree_std::max(origLength, copyLength + 1), false);
    MemoryCopy(m_Data, s, copyLength);
    m_Data[copyLength] = '\0';
  }
 
#if ADDITIONAL_CHECKS
  if (!len) {
    assert(*this == s);
  }
#endif
 
  m_Hash = 0;
}
 
void String::reserve(size_t size) {
  reserve(size, true);
}
 
void String::reserve(size_t size, bool zero) {
  assert(resizable());
 
  size = pedigree_std::max(size, StringMinimumAllocationSize);
 
  if (size > m_Size) {
    char* tmp = m_Data;
    m_Data = new char[size];
    if (tmp) {
      MemoryCopy(m_Data, tmp, m_Size > size ? size : m_Size);
      delete[] tmp;
    } else if (zero) {
      ByteSet(m_Data, 0, size);
    }
    m_Size = size;
  }
}
 
void String::downsize() {
  assert(resizable());
 
  size_t newSize = pedigree_std::max(m_Length + 1, StringMinimumAllocationSize);
 
  if (!m_Data || (newSize >= m_Size))
    return;
 
  char* oldData = m_Data;
 
  m_Data = new char[newSize];
  MemoryCopy(m_Data, oldData, m_Length + 1);
 
  delete[] oldData;
 
  m_Size = newSize;
}
 
void String::clear() noexcept {
  assert(assignable());
 
  if (m_Data) {
    delete[] m_Data;
  }
  m_Data = 0;
  m_Length = 0;
  m_Size = 0;
  m_Hash = 0;
}
 
void String::ltrim(size_t n) {
  assert(assignable());
 
  if (n > m_Length) {
    clear();
    return;
  }
 
  MemoryCopy(m_Data, &m_Data[n], m_Length - n);
  m_Length -= n;
  m_Data[m_Length] = 0;
 
  m_Hash = 0;
}
 
void String::rtrim(size_t n) {
  assert(assignable());
 
  if (n > m_Length) {
    clear();
    return;
  }
 
  m_Data[m_Length - n] = 0;
  m_Length -= n;
 
  m_Hash = 0;
}
 
String String::split(size_t offset) {
  String result;
  split(offset, result);
  return result;
}
 
void String::split(size_t offset, String& back) {
  assert(assignable());
 
  if (offset >= m_Length) {
    back.clear();
    return;
  }
 
  char* buf = extract();
 
  back.assign(&buf[offset], m_Length - offset, true);
  m_Length = offset;
  buf[m_Length] = 0;
 
  m_Hash = 0;
}
 
void String::strip() {
  assert(assignable());
 
  lstrip();
  rstrip();
}
 
void String::lstrip() {
  assert(assignable());
 
  char* buf = extract();
  if (!buf) {
    // nothing to strip
    return;
  }
 
  if (!iswhitespace(buf[0]))
    return;
 
  // finish up the byte tail
  size_t n = 0;
  while (n < m_Length && iswhitespace(buf[n]))
    n++;
 
  // Move the data to cover up the whitespace and avoid reallocating m_Data
  m_Length -= n;
  MemoryCopy(buf, (buf + n), m_Length);
  buf[m_Length] = 0;
 
  m_Hash = 0;
}
 
void String::rstrip() {
  assert(assignable());
 
  char* buf = extract();
  if (!buf) {
    // nothing to strip
    return;
  }
 
  if (!iswhitespace(buf[m_Length - 1]))
    return;
 
  size_t n = m_Length;
  while (n > 0 && iswhitespace(buf[n - 1]))
    n--;
 
  // m_Size is still valid - it's the size of the buffer. m_Length is now
  // updated to contain the proper length of the string, but the buffer is
  // not reallocated.
  m_Length = n;
  buf[m_Length] = 0;
 
  m_Hash = 0;
}
 
Vector<String> String::tokenise(char token) {
  Vector<String> list;
  tokenise(token, list);
  return list;
}
 
size_t String::Utf32ToUtf8(uint32_t utf32, char* utf8) {
  // clear out the string before conversion
  ByteSet(utf8, 0, 4);
 
  size_t nbuf = 0;
  if (utf32 <= 0x7F) {
    utf8[0] = utf32 & 0x7F;
    nbuf = 1;
  } else if (utf32 <= 0x7FF) {
    utf8[0] = 0xC0 | ((utf32 >> 6) & 0x1F);
    utf8[1] = 0x80 | (utf32 & 0x3F);
    nbuf = 2;
  } else if (utf32 <= 0xFFFF) {
    utf8[0] = 0xE0 | ((utf32 >> 12) & 0x0F);
    utf8[1] = 0x80 | ((utf32 >> 6) & 0x3F);
    utf8[2] = 0x80 | (utf32 & 0x3F);
    nbuf = 3;
  } else if (utf32 <= 0x10FFFF) {
    utf8[0] = 0xF0 | ((utf32 >> 18) & 0x07);
    utf8[1] = 0x80 | ((utf32 >> 12) & 0x3F);
    utf8[2] = 0x80 | ((utf32 >> 6) & 0x3F);
    utf8[3] = 0x80 | (utf32 & 0x3F);
    nbuf = 4;
  }
 
  return nbuf;
}
 
void String::tokenise(char token, Vector<StringView>& output) const {
  const char* orig_buffer = extract();
  const char* buffer = orig_buffer;
 
  output.clear();
  // reserve for the worst-case, where we tokenise every character of this string
  // output.reserve(m_Length, false);
 
  const char* pos = buffer ? StringFind(buffer, token) : nullptr;
  while (pos && (*buffer)) {
    if (pos == buffer) {
      ++buffer;
      continue;
    }
 
    if (pos > buffer) {
      output.createBack(buffer, pos - buffer);
    }
 
    buffer = pos + 1;
 
    pos = StringFind(buffer, token);
  }
 
  if (buffer && !pos) {
    // might be able to just copy this string rather than copy & move
    if (buffer == orig_buffer) {
      output.createBack(view());
    } else {
      size_t length = m_Length - (buffer - orig_buffer);
      if (length) {
        output.createBack(buffer, length);
      }
    }
  }
}
 
void String::tokenise(char token, Vector<String>& output) const {
  Vector<StringView> views;
  tokenise(token, views);
 
  output.clear();
  output.reserve(views.count(), false);
  for (auto& it : views) {
    output.createBack(it);
  }
}
 
void String::lchomp() {
  assert(assignable());
 
  if (!m_Length)
    return;
 
  char* buf = extract();
 
  StringCopy(buf, &buf[1]);
  --m_Length;
 
  m_Hash = 0;
}
 
void String::chomp() {
  assert(assignable());
 
  if (!m_Length)
    return;
 
  char* buf = extract();
 
  m_Length--;
  buf[m_Length] = '\0';
 
  m_Hash = 0;
}
 
void String::Format(const char* fmt, ...) {
  assert(assignable());
 
  reserve(256);
  va_list vl;
  va_start(vl, fmt);
  m_Length = VStringFormat(m_Data, fmt, vl);
  va_end(vl);
 
  m_Hash = 0;
}
 
bool String::endswith(const char c) const {
  if (!m_Length) {
    return false;
  }
 
  const char* buf = extract();
  return buf[m_Length - 1] == c;
}
 
bool String::endswith(const String& s) const {
  // Not a suffix check.
  if (m_Length == s.length())
    return *this == s;
 
  const char* otherbuf = s.extract();
  return endswith(otherbuf, s.length());
}
 
bool String::endswith(const char* s, size_t len) const {
  if (!len) {
    len = StringLength(s);
  }
 
  // Suffix exceeds our length.
  if (m_Length < len)
    return false;
 
  const char* mybuf = extract();
  mybuf += m_Length - len;
 
  return !MemoryCompare(mybuf, s, len);
}
 
bool String::startswith(const char c) const {
  if (!m_Length) {
    return false;
  }
 
  const char* buf = extract();
  return buf[0] == c;
}
 
bool String::startswith(const String& s) const {
  // Not a prefix check.
  if (m_Length == s.length())
    return *this == s;
 
  const char* otherbuf = s.extract();
  return startswith(otherbuf, s.length());
}
 
bool String::startswith(const char* s, size_t len) const {
  if (!len) {
    len = StringLength(s);
  }
 
  // Prefix exceeds our length.
  if (m_Length < len)
    return false;
 
  const char* mybuf = extract();
 
  // Do the check.
  return !MemoryCompare(mybuf, s, len);
}
 
bool String::iswhitespace(const char c) const {
  return (c <= ' ' || c == '\x7f');
}
 
char* String::extract() const {
  return m_Data;
}
 
ssize_t String::find(const char c) const {
  if (!m_Length)
    return -1;
 
  ssize_t signedLength = m_Length;
 
  char* buf = extract();
  for (ssize_t i = 0; i < signedLength; ++i) {
    if (buf[i] == c) {
      return i;
    }
  }
 
  return -1;
}
 
ssize_t String::rfind(const char c) const {
  if (!m_Length)
    return -1;
 
  char* buf = extract();
  for (ssize_t i = m_Length - 1, n = 0; i >= 0; --i, ++n) {
    if (buf[i] == c) {
      return n;
    }
  }
 
  return -1;
}
 
void String::computeHash() {
  if (m_Length) {
    m_Hash = spookyHash(extract(), m_Length);
  } else {
    m_Hash = 0;
  }
}
 
uint32_t String::computeHash() const {
  if (m_Length) {
    return spookyHash(extract(), m_Length);
  } else {
    return 0;
  }
}
 
String String::copy() const {
  String result;
  result.assign(*this);
  return result;
}
 
StringView String::view() const {
  // hash already calculated, enable hashing
  const char* buf = extract();
  return StringView(buf, m_Length, m_Hash, true);
}
 
bool String::resizable() const {
  return true;
}
 
bool String::assignable() const {
  return true;
}
 
void String::setLength(size_t n) {
  m_Length = n;
}
 
void String::setSize(size_t n) {
  m_Size = n;
}