1 files changed, 301 insertions, 249 deletions

diff --git a/contrib/libc++/src/experimental/filesystem/path.cpp b/contrib/libc++/src/experimental/filesystem/path.cpp
index 38c449832f6e..96b81f7b0a70 100644
--- a/contrib/libc++/src/experimental/filesystem/path.cpp
+++ b/contrib/libc++/src/experimental/filesystem/path.cpp
@@ -6,242 +6,279 @@
 // Source Licenses. See LICENSE.TXT for details.
 //
 //===----------------------------------------------------------------------===//
+#undef NDEBUG
 #include "experimental/filesystem"
-#include "experimental/string_view"
+#include "string_view"
 #include "utility"
-
-_LIBCPP_BEGIN_NAMESPACE_EXPERIMENTAL_FILESYSTEM
-
-_LIBCPP_CONSTEXPR path::value_type path::preferred_separator;
+#include "cassert"
 
 namespace { namespace parser
 {
+using namespace std;
+using namespace std::experimental::filesystem;
+
+using string_view_t = path::__string_view;
+using string_view_pair = pair<string_view_t, string_view_t>;
+using PosPtr = path::value_type const*;
+
+struct PathParser {
+  enum ParserState : unsigned char {
+    // Zero is a special sentinel value used by default constructed iterators.
+    PS_BeforeBegin = 1,
+    PS_InRootName,
+    PS_InRootDir,
+    PS_InFilenames,
+    PS_InTrailingSep,
+    PS_AtEnd
+  };
 
-using string_type = string_view;
-using value_type = path::value_type;
-
-using string_view_pair = pair<string_view, string_view>;
-
-// status reporting
-constexpr size_t npos = static_cast<size_t>(-1);
-
-inline bool good(size_t pos) { return pos != npos; }
-
-// lexical elements
-constexpr value_type preferred_separator = path::preferred_separator;
-constexpr value_type const * preferred_separator_str = "/";
-constexpr value_type const * dot = ".";
-
-// forward //
-bool is_separator(string_type const &, size_t);
-bool is_root_name(const string_type&, size_t);
-bool is_root_directory(string_type const &, size_t);
-bool is_trailing_separator(string_type const &, size_t);
-
-size_t start_of(string_type const &, size_t);
-size_t end_of(string_type const &, size_t);
-
-size_t root_name_start(const string_type& s);
-size_t root_name_end(const string_type&);
-
-size_t root_directory_start(string_type const &);
-size_t root_directory_end(string_type const &);
+  const string_view_t Path;
+  string_view_t RawEntry;
+  ParserState State;
 
-string_view_pair separate_filename(string_type const &);
-string_view extract_raw(string_type const &, size_t);
-string_view extract_preferred(string_type const &, size_t);
+private:
+  PathParser(string_view_t P, ParserState State) noexcept
+      : Path(P), State(State) {}
 
-inline bool is_separator(const string_type& s, size_t pos) {
-    return (pos < s.size() && s[pos] == preferred_separator);
-}
+public:
+  PathParser(string_view_t P, string_view_t E, unsigned char S)
+      : Path(P), RawEntry(E), State(static_cast<ParserState>(S)) {
+    // S cannot be '0' or PS_BeforeBegin.
+  }
 
-inline bool is_root_name(const string_type& s, size_t pos) {
-  return good(pos) && pos == 0 ? root_name_start(s) == pos : false;
-}
+  static PathParser CreateBegin(string_view_t P) noexcept {
+    PathParser PP(P, PS_BeforeBegin);
+    PP.increment();
+    return PP;
+  }
 
-inline bool is_root_directory(const string_type& s, size_t pos) {
-    return good(pos) ? root_directory_start(s) == pos : false;
-}
+  static PathParser CreateEnd(string_view_t P) noexcept {
+    PathParser PP(P, PS_AtEnd);
+    return PP;
+  }
 
-inline bool is_trailing_separator(const string_type& s, size_t pos) {
-    return (pos < s.size() && is_separator(s, pos) &&
-            end_of(s, pos) == s.size()-1 &&
-            !is_root_directory(s, pos) && !is_root_name(s, pos));
-}
+  PosPtr peek() const noexcept {
+    auto End = &Path.back() + 1;
+    auto TkEnd = getNextTokenStartPos();
+    return TkEnd == End ? nullptr : TkEnd;
+  }
 
-size_t start_of(const string_type& s, size_t pos) {
-    if (pos >= s.size()) return npos;
-    bool in_sep = (s[pos] == preferred_separator);
-    while (pos - 1 < s.size() &&
-        (s[pos-1] == preferred_separator) == in_sep)
-    { --pos; }
-    if (pos == 2 && !in_sep && s[0] == preferred_separator &&
-        s[1] == preferred_separator)
-    { return 0; }
-    return pos;
-}
+  void increment() noexcept {
+    const PosPtr End = &Path.back() + 1;
+    const PosPtr Start = getNextTokenStartPos();
+    if (Start == End)
+      return makeState(PS_AtEnd);
+
+    switch (State) {
+    case PS_BeforeBegin: {
+      PosPtr TkEnd = consumeSeparator(Start, End);
+      // If we consumed exactly two separators we have a root name.
+      if (TkEnd && TkEnd == Start + 2) {
+        // FIXME Do we need to consume a name or is '//' a root name on its own?
+        // what about '//.', '//..', '//...'?
+        auto NameEnd = consumeName(TkEnd, End);
+        if (NameEnd)
+          TkEnd = NameEnd;
+        return makeState(PS_InRootName, Start, TkEnd);
+      }
+      else if (TkEnd)
+        return makeState(PS_InRootDir, Start, TkEnd);
+      else
+        return makeState(PS_InFilenames, Start, consumeName(Start, End));
+    }
 
-size_t end_of(const string_type& s, size_t pos) {
-    if (pos >= s.size()) return npos;
-    // special case for root name
-    if (pos == 0 && is_root_name(s, pos)) return root_name_end(s);
-    bool in_sep = (s[pos] == preferred_separator);
-    while (pos + 1 < s.size() && (s[pos+1] == preferred_separator) == in_sep)
-    { ++pos; }
-    return pos;
-}
+    case PS_InRootName:
+      return makeState(PS_InRootDir, Start, consumeSeparator(Start, End));
+    case PS_InRootDir:
+      return makeState(PS_InFilenames, Start, consumeName(Start, End));
+
+    case PS_InFilenames: {
+      PosPtr SepEnd = consumeSeparator(Start, End);
+      if (SepEnd != End) {
+        PosPtr TkEnd = consumeName(SepEnd, End);
+        if (TkEnd)
+          return makeState(PS_InFilenames, SepEnd, TkEnd);
+      }
+      return makeState(PS_InTrailingSep, Start, SepEnd);
+    }
 
-inline size_t root_name_start(const string_type& s) {
-    return good(root_name_end(s)) ? 0 : npos;
-}
+    case PS_InTrailingSep:
+      return makeState(PS_AtEnd);
 
-size_t root_name_end(const string_type& s) {
-    if (s.size() < 2 || s[0] != preferred_separator
-        || s[1] != preferred_separator) {
-        return npos;
+    case PS_AtEnd:
+      _LIBCPP_UNREACHABLE();
     }
-    if (s.size() == 2) {
-        return 1;
+  }
+
+  void decrement() noexcept {
+    const PosPtr REnd = &Path.front() - 1;
+    const PosPtr RStart = getCurrentTokenStartPos() - 1;
+    assert(RStart != REnd);
+
+    switch (State) {
+    case PS_AtEnd: {
+      // Try to consume a trailing separator or root directory first.
+      if (PosPtr SepEnd = consumeSeparator(RStart, REnd)) {
+        if (SepEnd == REnd)
+          return makeState((RStart == REnd + 2) ? PS_InRootName : PS_InRootDir,
+                           Path.data(), RStart + 1);
+        // Check if we're seeing the root directory separator
+        auto PP = CreateBegin(Path);
+        bool InRootDir = PP.State == PS_InRootName &&
+            &PP.RawEntry.back() == SepEnd;
+        return makeState(InRootDir ? PS_InRootDir : PS_InTrailingSep,
+                         SepEnd + 1, RStart + 1);
+      } else {
+        PosPtr TkStart = consumeName(RStart, REnd);
+        if (TkStart == REnd + 2 && consumeSeparator(TkStart, REnd) == REnd)
+          return makeState(PS_InRootName, Path.data(), RStart + 1);
+        else
+          return makeState(PS_InFilenames, TkStart + 1, RStart + 1);
+      }
     }
-    size_t index = 2; // current position
-    if (s[index] == preferred_separator) {
-        return npos;
+    case PS_InTrailingSep:
+      return makeState(PS_InFilenames, consumeName(RStart, REnd) + 1, RStart + 1);
+    case PS_InFilenames: {
+      PosPtr SepEnd = consumeSeparator(RStart, REnd);
+      if (SepEnd == REnd)
+        return makeState((RStart == REnd + 2) ? PS_InRootName : PS_InRootDir,
+                         Path.data(), RStart + 1);
+      PosPtr TkEnd = consumeName(SepEnd, REnd);
+      if (TkEnd == REnd + 2 && consumeSeparator(TkEnd, REnd) == REnd)
+        return makeState(PS_InRootDir, SepEnd + 1, RStart + 1);
+      return makeState(PS_InFilenames, TkEnd + 1, SepEnd + 1);
     }
-    while (index + 1 < s.size() && s[index+1] != preferred_separator) {
-        ++index;
+    case PS_InRootDir:
+      return makeState(PS_InRootName, Path.data(), RStart + 1);
+    case PS_InRootName:
+    case PS_BeforeBegin:
+      _LIBCPP_UNREACHABLE();
     }
-    return index;
-}
-
-size_t root_directory_start(const string_type& s) {
-    size_t e = root_name_end(s);
-    if (!good(e))
-    return is_separator(s, 0) ? 0 : npos;
-    return is_separator(s, e + 1) ? e + 1 : npos;
-}
-
-size_t root_directory_end(const string_type& s) {
-    size_t st = root_directory_start(s);
-    if (!good(st)) return npos;
-    size_t index = st;
-    while (index + 1 < s.size() && s[index + 1] == preferred_separator)
-      { ++index; }
-    return index;
-}
-
-string_view_pair separate_filename(string_type const & s) {
-    if (s == "." || s == ".." || s.empty()) return string_view_pair{s, ""};
-    auto pos = s.find_last_of('.');
-    if (pos == string_type::npos) return string_view_pair{s, string_view{}};
-    return string_view_pair{s.substr(0, pos), s.substr(pos)};
-}
-
-inline string_view extract_raw(const string_type& s, size_t pos) {
-    size_t end_i = end_of(s, pos);
-    if (!good(end_i)) return string_view{};
-    return string_view(s).substr(pos, end_i - pos + 1);
-}
-
-string_view extract_preferred(const string_type& s, size_t pos) {
-    string_view raw = extract_raw(s, pos);
-    if (raw.empty())
-        return raw;
-    if (is_trailing_separator(s, pos))
-        return string_view{dot};
-    if (is_separator(s, pos) && !is_root_name(s, pos))
-        return string_view(preferred_separator_str);
-    return raw;
-}
-
-}} // namespace parser
-
-
-////////////////////////////////////////////////////////////////////////////////
-//                            path_view_iterator
-////////////////////////////////////////////////////////////////////////////////
-namespace {
-
-struct path_view_iterator {
-  const string_view __s_;
-  size_t __pos_;
+  }
 
-  explicit path_view_iterator(string_view const& __s) : __s_(__s), __pos_(__s_.empty() ? parser::npos : 0) {}
-  explicit path_view_iterator(string_view const& __s, size_t __p) : __s_(__s), __pos_(__p) {}
+  /// \brief Return a view with the "preferred representation" of the current
+  ///   element. For example trailing separators are represented as a '.'
+  string_view_t operator*() const noexcept {
+    switch (State) {
+    case PS_BeforeBegin:
+    case PS_AtEnd:
+      return "";
+    case PS_InRootDir:
+      return "/";
+    case PS_InTrailingSep:
+      return ".";
+    case PS_InRootName:
+    case PS_InFilenames:
+      return RawEntry;
+    }
+    _LIBCPP_UNREACHABLE();
+  }
 
-  string_view operator*() const {
-    return parser::extract_preferred(__s_, __pos_);
+  explicit operator bool() const noexcept {
+    return State != PS_BeforeBegin && State != PS_AtEnd;
   }
 
-  path_view_iterator& operator++() {
+  PathParser& operator++() noexcept {
     increment();
     return *this;
   }
 
-  path_view_iterator& operator--() {
+  PathParser& operator--() noexcept {
     decrement();
     return *this;
   }
 
-  void increment() {
-    if (__pos_ == parser::npos) return;
-    while (! set_position(parser::end_of(__s_, __pos_)+1))
-        ;
-    return;
+private:
+  void makeState(ParserState NewState, PosPtr Start, PosPtr End) noexcept {
+    State = NewState;
+    RawEntry = string_view_t(Start, End - Start);
+  }
+  void makeState(ParserState NewState) noexcept {
+    State = NewState;
+    RawEntry = {};
   }
 
-  void decrement() {
-    if (__pos_ == 0) {
-      set_position(0);
-    }
-    else if (__pos_ == parser::npos) {
-      auto const str_size = __s_.size();
-      set_position(parser::start_of(
-          __s_, str_size != 0 ? str_size - 1 : str_size));
-    } else {
-      while (!set_position(parser::start_of(__s_, __pos_-1)))
-        ;
+  /// \brief Return a pointer to the first character after the currently
+  ///   lexed element.
+  PosPtr getNextTokenStartPos() const noexcept {
+    switch (State) {
+    case PS_BeforeBegin:
+      return &Path.front();
+    case PS_InRootName:
+    case PS_InRootDir:
+    case PS_InFilenames:
+      return &RawEntry.back() + 1;
+    case PS_InTrailingSep:
+    case PS_AtEnd:
+      return &Path.back() + 1;
     }
+    _LIBCPP_UNREACHABLE();
   }
 
-  bool set_position(size_t pos) {
-    if (pos >= __s_.size()) {
-      __pos_ = parser::npos;
-    } else {
-      __pos_ = pos;
+  /// \brief Return a pointer to the first character in the currently lexed
+  ///   element.
+  PosPtr getCurrentTokenStartPos() const noexcept {
+    switch (State) {
+    case PS_BeforeBegin:
+    case PS_InRootName:
+      return &Path.front();
+    case PS_InRootDir:
+    case PS_InFilenames:
+    case PS_InTrailingSep:
+      return &RawEntry.front();
+    case PS_AtEnd:
+      return &Path.back() + 1;
     }
-    return valid_iterator_position();
+    _LIBCPP_UNREACHABLE();
   }
 
-  bool valid_iterator_position() const {
-    if (__pos_ == parser::npos) return true; // end position is valid
-    return (!parser::is_separator      (__s_, __pos_) ||
-          parser::is_root_directory    (__s_, __pos_) ||
-          parser::is_trailing_separator(__s_, __pos_) ||
-          parser::is_root_name         (__s_, __pos_));
+  PosPtr consumeSeparator(PosPtr P, PosPtr End) const noexcept {
+    if (P == End || *P != '/')
+      return nullptr;
+    const int Inc = P < End ? 1 : -1;
+    P += Inc;
+    while (P != End && *P == '/')
+      P += Inc;
+    return P;
   }
 
-  bool is_end() const { return __pos_ == parser::npos; }
-
-  inline bool operator==(path_view_iterator const& __p) {
-      return __pos_ == __p.__pos_;
+  PosPtr consumeName(PosPtr P, PosPtr End) const noexcept {
+    if (P == End || *P == '/')
+      return nullptr;
+    const int Inc = P < End ? 1 : -1;
+    P += Inc;
+    while (P != End && *P != '/')
+      P += Inc;
+    return P;
   }
 };
 
-path_view_iterator pbegin(path const& p) {
-    return path_view_iterator(p.native());
+string_view_pair separate_filename(string_view_t const & s) {
+    if (s == "." || s == ".." || s.empty()) return string_view_pair{s, ""};
+    auto pos = s.find_last_of('.');
+    if (pos == string_view_t::npos) return string_view_pair{s, string_view{}};
+    return string_view_pair{s.substr(0, pos), s.substr(pos)};
 }
 
-path_view_iterator pend(path const& p) {
-    path_view_iterator __p(p.native());
-    __p.__pos_ = parser::npos;
-    return __p;
+string_view_t createView(PosPtr S, PosPtr E) noexcept {
+  return {S, static_cast<size_t>(E - S) + 1};
 }
 
-} // end namespace
+}} // namespace parser
+
+_LIBCPP_BEGIN_NAMESPACE_EXPERIMENTAL_FILESYSTEM
+
+using parser::string_view_t;
+using parser::string_view_pair;
+using parser::PathParser;
+using parser::createView;
+
 ///////////////////////////////////////////////////////////////////////////////
 //                            path definitions
 ///////////////////////////////////////////////////////////////////////////////
 
+constexpr path::value_type path::preferred_separator;
+
 path & path::replace_extension(path const & replacement)
 {
     path p = extension();
@@ -260,131 +297,146 @@ path & path::replace_extension(path const & replacement)
 ///////////////////////////////////////////////////////////////////////////////
 // path.decompose
 
-string_view path::__root_name() const
+string_view_t path::__root_name() const
 {
-    return parser::is_root_name(__pn_, 0)
-      ? parser::extract_preferred(__pn_, 0)
-      : string_view{};
+    auto PP = PathParser::CreateBegin(__pn_);
+    if (PP.State == PathParser::PS_InRootName)
+      return *PP;
+    return {};
 }
 
-string_view path::__root_directory() const
+string_view_t path::__root_directory() const
 {
-    auto start_i = parser::root_directory_start(__pn_);
-    if(!parser::good(start_i)) {
-        return {};
-    }
-    return parser::extract_preferred(__pn_, start_i);
+    auto PP = PathParser::CreateBegin(__pn_);
+    if (PP.State == PathParser::PS_InRootName)
+      ++PP;
+    if (PP.State == PathParser::PS_InRootDir)
+      return *PP;
+    return {};
 }
 
-string_view path::__relative_path() const
+string_view_t path::__root_path_raw() const
 {
-    if (empty()) {
-        return {__pn_};
-    }
-    auto end_i = parser::root_directory_end(__pn_);
-    if (not parser::good(end_i)) {
-        end_i = parser::root_name_end(__pn_);
+    auto PP = PathParser::CreateBegin(__pn_);
+    if (PP.State == PathParser::PS_InRootName) {
+      auto NextCh = PP.peek();
+      if (NextCh && *NextCh == '/') {
+        ++PP;
+        assert(PP.State == PathParser::PS_InRootDir);
+        return createView(__pn_.data(), &PP.RawEntry.back());
+      }
+      return PP.RawEntry;
     }
-    if (not parser::good(end_i)) {
-        return {__pn_};
-    }
-    return string_view(__pn_).substr(end_i+1);
+    if (PP.State == PathParser::PS_InRootDir)
+      return *PP;
+    return {};
 }
 
-string_view path::__parent_path() const
+string_view_t path::__relative_path() const
 {
-    if (empty() || pbegin(*this) == --pend(*this)) {
-        return {};
-    }
-    auto end_it = --(--pend(*this));
-    auto end_i = parser::end_of(__pn_, end_it.__pos_);
-    return string_view(__pn_).substr(0, end_i+1);
+    auto PP = PathParser::CreateBegin(__pn_);
+    while (PP.State <= PathParser::PS_InRootDir)
+      ++PP;
+    if (PP.State == PathParser::PS_AtEnd)
+      return {};
+    return createView(PP.RawEntry.data(), &__pn_.back());
 }
 
-string_view path::__filename() const
+string_view_t path::__parent_path() const
 {
-    return empty() ? string_view{} : *--pend(*this);
+    if (empty())
+      return {};
+    auto PP = PathParser::CreateEnd(__pn_);
+    --PP;
+    if (PP.RawEntry.data() == __pn_.data())
+      return {};
+    --PP;
+    return createView(__pn_.data(), &PP.RawEntry.back());
 }
 
-string_view path::__stem() const
+string_view_t path::__filename() const
+{
+    if (empty()) return {};
+    return *(--PathParser::CreateEnd(__pn_));
+}
+
+string_view_t path::__stem() const
 {
     return parser::separate_filename(__filename()).first;
 }
 
-string_view path::__extension() const
+string_view_t path::__extension() const
 {
     return parser::separate_filename(__filename()).second;
 }
 
 ////////////////////////////////////////////////////////////////////////////
 // path.comparisons
-int path::__compare(const value_type* __s) const {
-    path_view_iterator thisIter(this->native());
-    path_view_iterator sIter(__s);
-    while (!thisIter.is_end() && !sIter.is_end()) {
-        int res = (*thisIter).compare(*sIter);
+int path::__compare(string_view_t __s) const {
+    auto PP = PathParser::CreateBegin(__pn_);
+    auto PP2 = PathParser::CreateBegin(__s);
+    while (PP && PP2) {
+        int res = (*PP).compare(*PP2);
         if (res != 0) return res;
-        ++thisIter; ++sIter;
+        ++PP; ++PP2;
     }
-    if (thisIter.is_end() && sIter.is_end())
+    if (PP.State == PP2.State && PP.State == PathParser::PS_AtEnd)
         return 0;
-    if (thisIter.is_end())
+    if (PP.State == PathParser::PS_AtEnd)
         return -1;
     return 1;
 }
 
 ////////////////////////////////////////////////////////////////////////////
 // path.nonmembers
-size_t hash_value(const path& __p) _NOEXCEPT {
-  path_view_iterator thisIter(__p.native());
-  struct HashPairT {
-    size_t first;
-    size_t second;
-  };
-  HashPairT hp = {0, 0};
+size_t hash_value(const path& __p) noexcept {
+  auto PP = PathParser::CreateBegin(__p.native());
+  size_t hash_value = 0;
   std::hash<string_view> hasher;
-  std::__scalar_hash<decltype(hp)> pair_hasher;
-  while (!thisIter.is_end()) {
-    hp.second = hasher(*thisIter);
-    hp.first = pair_hasher(hp);
-    ++thisIter;
+  while (PP) {
+    hash_value = __hash_combine(hash_value, hasher(*PP));
+    ++PP;
   }
-  return hp.first;
+  return hash_value;
 }
 
 ////////////////////////////////////////////////////////////////////////////
 // path.itr
 path::iterator path::begin() const
 {
-    path_view_iterator pit = pbegin(*this);
+    auto PP = PathParser::CreateBegin(__pn_);
     iterator it;
     it.__path_ptr_ = this;
-    it.__pos_ = pit.__pos_;
-    it.__elem_.__assign_view(*pit);
+    it.__state_ = PP.State;
+    it.__entry_ = PP.RawEntry;
+    it.__stashed_elem_.__assign_view(*PP);
     return it;
 }
 
 path::iterator path::end() const
 {
     iterator it{};
+    it.__state_ = PathParser::PS_AtEnd;
     it.__path_ptr_ = this;
-    it.__pos_ = parser::npos;
     return it;
 }
 
 path::iterator& path::iterator::__increment() {
-  path_view_iterator it(__path_ptr_->native(), __pos_);
-  it.increment();
-  __pos_ = it.__pos_;
-  __elem_.__assign_view(*it);
+  static_assert(__at_end == PathParser::PS_AtEnd, "");
+  PathParser PP(__path_ptr_->native(), __entry_, __state_);
+  ++PP;
+  __state_ = PP.State;
+  __entry_ = PP.RawEntry;
+  __stashed_elem_.__assign_view(*PP);
   return *this;
 }
 
 path::iterator& path::iterator::__decrement() {
-  path_view_iterator it(__path_ptr_->native(), __pos_);
-  it.decrement();
-  __pos_ = it.__pos_;
-  __elem_.__assign_view(*it);
+  PathParser PP(__path_ptr_->native(), __entry_, __state_);
+  --PP;
+  __state_ = PP.State;
+  __entry_ = PP.RawEntry;
+  __stashed_elem_.__assign_view(*PP);
   return *this;
 }