MFC 261991:

Upgrade our copy of llvm/clang to 3.4 release.  This version supports
all of the features in the current working draft of the upcoming C++
standard, provisionally named C++1y.

The code generator's performance is greatly increased, and the loop
auto-vectorizer is now enabled at -Os and -O2 in addition to -O3.  The
PowerPC backend has made several major improvements to code generation
quality and compile time, and the X86, SPARC, ARM32, Aarch64 and SystemZ
backends have all seen major feature work.

Release notes for llvm and clang can be found here:
<http://llvm.org/releases/3.4/docs/ReleaseNotes.html>
<http://llvm.org/releases/3.4/tools/clang/docs/ReleaseNotes.html>

MFC 262121 (by emaste):

Update lldb for clang/llvm 3.4 import

This commit largely restores the lldb source to the upstream r196259
snapshot with the addition of threaded inferior support and a few bug
fixes.

Specific upstream lldb revisions restored include:
   SVN      git
  181387  779e6ac
  181703  7bef4e2
  182099  b31044e
  182650  f2dcf35
  182683  0d91b80
  183862  15c1774
  183929  99447a6
  184177  0b2934b
  184948  4dc3761
  184954  007e7bc
  186990  eebd175

Sponsored by:	DARPA, AFRL

MFC 262186 (by emaste):

Fix mismerge in r262121

A break statement was lost in the merge.  The error had no functional
impact, but restore it to reduce the diff against upstream.

MFC 262303:

Pull in r197521 from upstream clang trunk (by rdivacky):

  Use the integrated assembler by default on FreeBSD/ppc and ppc64.

Requested by:	jhibbits

MFC 262611:

Pull in r196874 from upstream llvm trunk:

  Fix a crash that occurs when PWD is invalid.

  MCJIT needs to be able to run in hostile environments, even when PWD
  is invalid. There's no need to crash MCJIT in this case.

  The obvious fix is to simply leave MCContext's CompilationDir empty
  when PWD can't be determined. This way, MCJIT clients,
  and other clients that link with LLVM don't need a valid working directory.

  If we do want to guarantee valid CompilationDir, that should be done
  only for clients of getCompilationDir(). This is as simple as checking
  for an empty string.

  The only current use of getCompilationDir is EmitGenDwarfInfo, which
  won't conceivably run with an invalid working dir. However, in the
  purely hypothetically and untestable case that this happens, the
  AT_comp_dir will be omitted from the compilation_unit DIE.

This should help fix assertions occurring with ports-mgmt/tinderbox,
when it is using jails, and sometimes invalidates clang's current
working directory.

Reported by:	decke

MFC 262809:

Pull in r203007 from upstream clang trunk:

  Don't produce an alias between destructors with different calling conventions.

  Fixes pr19007.

(Please note that is an LLVM PR identifier, not a FreeBSD one.)

This should fix Firefox and/or libxul crashes (due to problems with
regparm/stdcall calling conventions) on i386.

Reported by:	multiple users on freebsd-current
PR:		bin/187103

MFC 263048:

Repair recognition of "CC" as an alias for the C++ compiler, since it
was silently broken by upstream for a Windows-specific use-case.

Apparently some versions of CMake still rely on this archaic feature...

Reported by:	rakuco

MFC 263049:

Garbage collect the old way of adding the libstdc++ include directories
in clang's InitHeaderSearch.cpp.  This has been superseded by David
Chisnall's commit in r255321.

Moreover, if libc++ is used, the libstdc++ include directories should
not be in the search path at all.  These directories are now only used
if you pass -stdlib=libstdc++.

1 files changed, 1136 insertions, 0 deletions

diff --git a/contrib/llvm/tools/clang/lib/Index/CommentToXML.cpp b/contrib/llvm/tools/clang/lib/Index/CommentToXML.cpp
new file mode 100644
index 0000000..0a9619d
--- /dev/null
+++ b/contrib/llvm/tools/clang/lib/Index/CommentToXML.cpp
@@ -0,0 +1,1136 @@
+//===--- CommentToXML.cpp - Convert comments to XML representation --------===//
+//
+//                     The LLVM Compiler Infrastructure
+//
+// This file is distributed under the University of Illinois Open Source
+// License. See LICENSE.TXT for details.
+//
+//===----------------------------------------------------------------------===//
+
+#include "clang/Index/CommentToXML.h"
+#include "SimpleFormatContext.h"
+#include "clang/AST/Attr.h"
+#include "clang/AST/ASTContext.h"
+#include "clang/AST/Comment.h"
+#include "clang/AST/CommentVisitor.h"
+#include "clang/Format/Format.h"
+#include "clang/Index/USRGeneration.h"
+#include "clang/Lex/Lexer.h"
+#include "llvm/ADT/StringExtras.h"
+#include "llvm/ADT/TinyPtrVector.h"
+#include "llvm/Support/raw_ostream.h"
+
+using namespace clang;
+using namespace clang::comments;
+using namespace clang::index;
+
+namespace {
+
+/// This comparison will sort parameters with valid index by index, then vararg
+/// parameters, and invalid (unresolved) parameters last.
+class ParamCommandCommentCompareIndex {
+public:
+  bool operator()(const ParamCommandComment *LHS,
+                  const ParamCommandComment *RHS) const {
+    unsigned LHSIndex = UINT_MAX;
+    unsigned RHSIndex = UINT_MAX;
+
+    if (LHS->isParamIndexValid()) {
+      if (LHS->isVarArgParam())
+        LHSIndex = UINT_MAX - 1;
+      else
+        LHSIndex = LHS->getParamIndex();
+    }
+    if (RHS->isParamIndexValid()) {
+      if (RHS->isVarArgParam())
+        RHSIndex = UINT_MAX - 1;
+      else
+        RHSIndex = RHS->getParamIndex();
+    }
+    return LHSIndex < RHSIndex;
+  }
+};
+
+/// This comparison will sort template parameters in the following order:
+/// \li real template parameters (depth = 1) in index order;
+/// \li all other names (depth > 1);
+/// \li unresolved names.
+class TParamCommandCommentComparePosition {
+public:
+  bool operator()(const TParamCommandComment *LHS,
+                  const TParamCommandComment *RHS) const {
+    // Sort unresolved names last.
+    if (!LHS->isPositionValid())
+      return false;
+    if (!RHS->isPositionValid())
+      return true;
+
+    if (LHS->getDepth() > 1)
+      return false;
+    if (RHS->getDepth() > 1)
+      return true;
+
+    // Sort template parameters in index order.
+    if (LHS->getDepth() == 1 && RHS->getDepth() == 1)
+      return LHS->getIndex(0) < RHS->getIndex(0);
+
+    // Leave all other names in source order.
+    return true;
+  }
+};
+
+/// Separate parts of a FullComment.
+struct FullCommentParts {
+  /// Take a full comment apart and initialize members accordingly.
+  FullCommentParts(const FullComment *C,
+                   const CommandTraits &Traits);
+
+  const BlockContentComment *Brief;
+  const BlockContentComment *Headerfile;
+  const ParagraphComment *FirstParagraph;
+  SmallVector<const BlockCommandComment *, 4> Returns;
+  SmallVector<const ParamCommandComment *, 8> Params;
+  SmallVector<const TParamCommandComment *, 4> TParams;
+  llvm::TinyPtrVector<const BlockCommandComment *> Exceptions;
+  SmallVector<const BlockContentComment *, 8> MiscBlocks;
+};
+
+FullCommentParts::FullCommentParts(const FullComment *C,
+                                   const CommandTraits &Traits) :
+    Brief(NULL), Headerfile(NULL), FirstParagraph(NULL) {
+  for (Comment::child_iterator I = C->child_begin(), E = C->child_end();
+       I != E; ++I) {
+    const Comment *Child = *I;
+    if (!Child)
+      continue;
+    switch (Child->getCommentKind()) {
+    case Comment::NoCommentKind:
+      continue;
+
+    case Comment::ParagraphCommentKind: {
+      const ParagraphComment *PC = cast<ParagraphComment>(Child);
+      if (PC->isWhitespace())
+        break;
+      if (!FirstParagraph)
+        FirstParagraph = PC;
+
+      MiscBlocks.push_back(PC);
+      break;
+    }
+
+    case Comment::BlockCommandCommentKind: {
+      const BlockCommandComment *BCC = cast<BlockCommandComment>(Child);
+      const CommandInfo *Info = Traits.getCommandInfo(BCC->getCommandID());
+      if (!Brief && Info->IsBriefCommand) {
+        Brief = BCC;
+        break;
+      }
+      if (!Headerfile && Info->IsHeaderfileCommand) {
+        Headerfile = BCC;
+        break;
+      }
+      if (Info->IsReturnsCommand) {
+        Returns.push_back(BCC);
+        break;
+      }
+      if (Info->IsThrowsCommand) {
+        Exceptions.push_back(BCC);
+        break;
+      }
+      MiscBlocks.push_back(BCC);
+      break;
+    }
+
+    case Comment::ParamCommandCommentKind: {
+      const ParamCommandComment *PCC = cast<ParamCommandComment>(Child);
+      if (!PCC->hasParamName())
+        break;
+
+      if (!PCC->isDirectionExplicit() && !PCC->hasNonWhitespaceParagraph())
+        break;
+
+      Params.push_back(PCC);
+      break;
+    }
+
+    case Comment::TParamCommandCommentKind: {
+      const TParamCommandComment *TPCC = cast<TParamCommandComment>(Child);
+      if (!TPCC->hasParamName())
+        break;
+
+      if (!TPCC->hasNonWhitespaceParagraph())
+        break;
+
+      TParams.push_back(TPCC);
+      break;
+    }
+
+    case Comment::VerbatimBlockCommentKind:
+      MiscBlocks.push_back(cast<BlockCommandComment>(Child));
+      break;
+
+    case Comment::VerbatimLineCommentKind: {
+      const VerbatimLineComment *VLC = cast<VerbatimLineComment>(Child);
+      const CommandInfo *Info = Traits.getCommandInfo(VLC->getCommandID());
+      if (!Info->IsDeclarationCommand)
+        MiscBlocks.push_back(VLC);
+      break;
+    }
+
+    case Comment::TextCommentKind:
+    case Comment::InlineCommandCommentKind:
+    case Comment::HTMLStartTagCommentKind:
+    case Comment::HTMLEndTagCommentKind:
+    case Comment::VerbatimBlockLineCommentKind:
+    case Comment::FullCommentKind:
+      llvm_unreachable("AST node of this kind can't be a child of "
+                       "a FullComment");
+    }
+  }
+
+  // Sort params in order they are declared in the function prototype.
+  // Unresolved parameters are put at the end of the list in the same order
+  // they were seen in the comment.
+  std::stable_sort(Params.begin(), Params.end(),
+                   ParamCommandCommentCompareIndex());
+
+  std::stable_sort(TParams.begin(), TParams.end(),
+                   TParamCommandCommentComparePosition());
+}
+
+void printHTMLStartTagComment(const HTMLStartTagComment *C,
+                              llvm::raw_svector_ostream &Result) {
+  Result << "<" << C->getTagName();
+
+  if (C->getNumAttrs() != 0) {
+    for (unsigned i = 0, e = C->getNumAttrs(); i != e; i++) {
+      Result << " ";
+      const HTMLStartTagComment::Attribute &Attr = C->getAttr(i);
+      Result << Attr.Name;
+      if (!Attr.Value.empty())
+        Result << "=\"" << Attr.Value << "\"";
+    }
+  }
+
+  if (!C->isSelfClosing())
+    Result << ">";
+  else
+    Result << "/>";
+}
+
+class CommentASTToHTMLConverter :
+    public ConstCommentVisitor<CommentASTToHTMLConverter> {
+public:
+  /// \param Str accumulator for HTML.
+  CommentASTToHTMLConverter(const FullComment *FC,
+                            SmallVectorImpl<char> &Str,
+                            const CommandTraits &Traits) :
+      FC(FC), Result(Str), Traits(Traits)
+  { }
+
+  // Inline content.
+  void visitTextComment(const TextComment *C);
+  void visitInlineCommandComment(const InlineCommandComment *C);
+  void visitHTMLStartTagComment(const HTMLStartTagComment *C);
+  void visitHTMLEndTagComment(const HTMLEndTagComment *C);
+
+  // Block content.
+  void visitParagraphComment(const ParagraphComment *C);
+  void visitBlockCommandComment(const BlockCommandComment *C);
+  void visitParamCommandComment(const ParamCommandComment *C);
+  void visitTParamCommandComment(const TParamCommandComment *C);
+  void visitVerbatimBlockComment(const VerbatimBlockComment *C);
+  void visitVerbatimBlockLineComment(const VerbatimBlockLineComment *C);
+  void visitVerbatimLineComment(const VerbatimLineComment *C);
+
+  void visitFullComment(const FullComment *C);
+
+  // Helpers.
+
+  /// Convert a paragraph that is not a block by itself (an argument to some
+  /// command).
+  void visitNonStandaloneParagraphComment(const ParagraphComment *C);
+
+  void appendToResultWithHTMLEscaping(StringRef S);
+
+private:
+  const FullComment *FC;
+  /// Output stream for HTML.
+  llvm::raw_svector_ostream Result;
+
+  const CommandTraits &Traits;
+};
+} // end unnamed namespace
+
+void CommentASTToHTMLConverter::visitTextComment(const TextComment *C) {
+  appendToResultWithHTMLEscaping(C->getText());
+}
+
+void CommentASTToHTMLConverter::visitInlineCommandComment(
+                                  const InlineCommandComment *C) {
+  // Nothing to render if no arguments supplied.
+  if (C->getNumArgs() == 0)
+    return;
+
+  // Nothing to render if argument is empty.
+  StringRef Arg0 = C->getArgText(0);
+  if (Arg0.empty())
+    return;
+
+  switch (C->getRenderKind()) {
+  case InlineCommandComment::RenderNormal:
+    for (unsigned i = 0, e = C->getNumArgs(); i != e; ++i) {
+      appendToResultWithHTMLEscaping(C->getArgText(i));
+      Result << " ";
+    }
+    return;
+
+  case InlineCommandComment::RenderBold:
+    assert(C->getNumArgs() == 1);
+    Result << "<b>";
+    appendToResultWithHTMLEscaping(Arg0);
+    Result << "</b>";
+    return;
+  case InlineCommandComment::RenderMonospaced:
+    assert(C->getNumArgs() == 1);
+    Result << "<tt>";
+    appendToResultWithHTMLEscaping(Arg0);
+    Result<< "</tt>";
+    return;
+  case InlineCommandComment::RenderEmphasized:
+    assert(C->getNumArgs() == 1);
+    Result << "<em>";
+    appendToResultWithHTMLEscaping(Arg0);
+    Result << "</em>";
+    return;
+  }
+}
+
+void CommentASTToHTMLConverter::visitHTMLStartTagComment(
+                                  const HTMLStartTagComment *C) {
+  printHTMLStartTagComment(C, Result);
+}
+
+void CommentASTToHTMLConverter::visitHTMLEndTagComment(
+                                  const HTMLEndTagComment *C) {
+  Result << "</" << C->getTagName() << ">";
+}
+
+void CommentASTToHTMLConverter::visitParagraphComment(
+                                  const ParagraphComment *C) {
+  if (C->isWhitespace())
+    return;
+
+  Result << "<p>";
+  for (Comment::child_iterator I = C->child_begin(), E = C->child_end();
+       I != E; ++I) {
+    visit(*I);
+  }
+  Result << "</p>";
+}
+
+void CommentASTToHTMLConverter::visitBlockCommandComment(
+                                  const BlockCommandComment *C) {
+  const CommandInfo *Info = Traits.getCommandInfo(C->getCommandID());
+  if (Info->IsBriefCommand) {
+    Result << "<p class=\"para-brief\">";
+    visitNonStandaloneParagraphComment(C->getParagraph());
+    Result << "</p>";
+    return;
+  }
+  if (Info->IsReturnsCommand) {
+    Result << "<p class=\"para-returns\">"
+              "<span class=\"word-returns\">Returns</span> ";
+    visitNonStandaloneParagraphComment(C->getParagraph());
+    Result << "</p>";
+    return;
+  }
+  // We don't know anything about this command.  Just render the paragraph.
+  visit(C->getParagraph());
+}
+
+void CommentASTToHTMLConverter::visitParamCommandComment(
+                                  const ParamCommandComment *C) {
+  if (C->isParamIndexValid()) {
+    if (C->isVarArgParam()) {
+      Result << "<dt class=\"param-name-index-vararg\">";
+      appendToResultWithHTMLEscaping(C->getParamNameAsWritten());
+    } else {
+      Result << "<dt class=\"param-name-index-"
+             << C->getParamIndex()
+             << "\">";
+      appendToResultWithHTMLEscaping(C->getParamName(FC));
+    }
+  } else {
+    Result << "<dt class=\"param-name-index-invalid\">";
+    appendToResultWithHTMLEscaping(C->getParamNameAsWritten());
+  }
+  Result << "</dt>";
+
+  if (C->isParamIndexValid()) {
+    if (C->isVarArgParam())
+      Result << "<dd class=\"param-descr-index-vararg\">";
+    else
+      Result << "<dd class=\"param-descr-index-"
+             << C->getParamIndex()
+             << "\">";
+  } else
+    Result << "<dd class=\"param-descr-index-invalid\">";
+
+  visitNonStandaloneParagraphComment(C->getParagraph());
+  Result << "</dd>";
+}
+
+void CommentASTToHTMLConverter::visitTParamCommandComment(
+                                  const TParamCommandComment *C) {
+  if (C->isPositionValid()) {
+    if (C->getDepth() == 1)
+      Result << "<dt class=\"tparam-name-index-"
+             << C->getIndex(0)
+             << "\">";
+    else
+      Result << "<dt class=\"tparam-name-index-other\">";
+    appendToResultWithHTMLEscaping(C->getParamName(FC));
+  } else {
+    Result << "<dt class=\"tparam-name-index-invalid\">";
+    appendToResultWithHTMLEscaping(C->getParamNameAsWritten());
+  }
+
+  Result << "</dt>";
+
+  if (C->isPositionValid()) {
+    if (C->getDepth() == 1)
+      Result << "<dd class=\"tparam-descr-index-"
+             << C->getIndex(0)
+             << "\">";
+    else
+      Result << "<dd class=\"tparam-descr-index-other\">";
+  } else
+    Result << "<dd class=\"tparam-descr-index-invalid\">";
+
+  visitNonStandaloneParagraphComment(C->getParagraph());
+  Result << "</dd>";
+}
+
+void CommentASTToHTMLConverter::visitVerbatimBlockComment(
+                                  const VerbatimBlockComment *C) {
+  unsigned NumLines = C->getNumLines();
+  if (NumLines == 0)
+    return;
+
+  Result << "<pre>";
+  for (unsigned i = 0; i != NumLines; ++i) {
+    appendToResultWithHTMLEscaping(C->getText(i));
+    if (i + 1 != NumLines)
+      Result << '\n';
+  }
+  Result << "</pre>";
+}
+
+void CommentASTToHTMLConverter::visitVerbatimBlockLineComment(
+                                  const VerbatimBlockLineComment *C) {
+  llvm_unreachable("should not see this AST node");
+}
+
+void CommentASTToHTMLConverter::visitVerbatimLineComment(
+                                  const VerbatimLineComment *C) {
+  Result << "<pre>";
+  appendToResultWithHTMLEscaping(C->getText());
+  Result << "</pre>";
+}
+
+void CommentASTToHTMLConverter::visitFullComment(const FullComment *C) {
+  FullCommentParts Parts(C, Traits);
+
+  bool FirstParagraphIsBrief = false;
+  if (Parts.Headerfile)
+    visit(Parts.Headerfile);
+  if (Parts.Brief)
+    visit(Parts.Brief);
+  else if (Parts.FirstParagraph) {
+    Result << "<p class=\"para-brief\">";
+    visitNonStandaloneParagraphComment(Parts.FirstParagraph);
+    Result << "</p>";
+    FirstParagraphIsBrief = true;
+  }
+
+  for (unsigned i = 0, e = Parts.MiscBlocks.size(); i != e; ++i) {
+    const Comment *C = Parts.MiscBlocks[i];
+    if (FirstParagraphIsBrief && C == Parts.FirstParagraph)
+      continue;
+    visit(C);
+  }
+
+  if (Parts.TParams.size() != 0) {
+    Result << "<dl>";
+    for (unsigned i = 0, e = Parts.TParams.size(); i != e; ++i)
+      visit(Parts.TParams[i]);
+    Result << "</dl>";
+  }
+
+  if (Parts.Params.size() != 0) {
+    Result << "<dl>";
+    for (unsigned i = 0, e = Parts.Params.size(); i != e; ++i)
+      visit(Parts.Params[i]);
+    Result << "</dl>";
+  }
+
+  if (Parts.Returns.size() != 0) {
+    Result << "<div class=\"result-discussion\">";
+    for (unsigned i = 0, e = Parts.Returns.size(); i != e; ++i)
+      visit(Parts.Returns[i]);
+    Result << "</div>";
+  }
+
+  Result.flush();
+}
+
+void CommentASTToHTMLConverter::visitNonStandaloneParagraphComment(
+                                  const ParagraphComment *C) {
+  if (!C)
+    return;
+
+  for (Comment::child_iterator I = C->child_begin(), E = C->child_end();
+       I != E; ++I) {
+    visit(*I);
+  }
+}
+
+void CommentASTToHTMLConverter::appendToResultWithHTMLEscaping(StringRef S) {
+  for (StringRef::iterator I = S.begin(), E = S.end(); I != E; ++I) {
+    const char C = *I;
+    switch (C) {
+    case '&':
+      Result << "&amp;";
+      break;
+    case '<':
+      Result << "&lt;";
+      break;
+    case '>':
+      Result << "&gt;";
+      break;
+    case '"':
+      Result << "&quot;";
+      break;
+    case '\'':
+      Result << "&#39;";
+      break;
+    case '/':
+      Result << "&#47;";
+      break;
+    default:
+      Result << C;
+      break;
+    }
+  }
+}
+
+namespace {
+class CommentASTToXMLConverter :
+    public ConstCommentVisitor<CommentASTToXMLConverter> {
+public:
+  /// \param Str accumulator for XML.
+  CommentASTToXMLConverter(const FullComment *FC,
+                           SmallVectorImpl<char> &Str,
+                           const CommandTraits &Traits,
+                           const SourceManager &SM,
+                           SimpleFormatContext &SFC,
+                           unsigned FUID) :
+      FC(FC), Result(Str), Traits(Traits), SM(SM),
+      FormatRewriterContext(SFC),
+      FormatInMemoryUniqueId(FUID) { }
+
+  // Inline content.
+  void visitTextComment(const TextComment *C);
+  void visitInlineCommandComment(const InlineCommandComment *C);
+  void visitHTMLStartTagComment(const HTMLStartTagComment *C);
+  void visitHTMLEndTagComment(const HTMLEndTagComment *C);
+
+  // Block content.
+  void visitParagraphComment(const ParagraphComment *C);
+
+  void appendParagraphCommentWithKind(const ParagraphComment *C,
+                                      StringRef Kind);
+
+  void visitBlockCommandComment(const BlockCommandComment *C);
+  void visitParamCommandComment(const ParamCommandComment *C);
+  void visitTParamCommandComment(const TParamCommandComment *C);
+  void visitVerbatimBlockComment(const VerbatimBlockComment *C);
+  void visitVerbatimBlockLineComment(const VerbatimBlockLineComment *C);
+  void visitVerbatimLineComment(const VerbatimLineComment *C);
+
+  void visitFullComment(const FullComment *C);
+
+  // Helpers.
+  void appendToResultWithXMLEscaping(StringRef S);
+
+  void formatTextOfDeclaration(const DeclInfo *DI,
+                               SmallString<128> &Declaration);
+
+private:
+  const FullComment *FC;
+
+  /// Output stream for XML.
+  llvm::raw_svector_ostream Result;
+
+  const CommandTraits &Traits;
+  const SourceManager &SM;
+  SimpleFormatContext &FormatRewriterContext;
+  unsigned FormatInMemoryUniqueId;
+};
+
+void getSourceTextOfDeclaration(const DeclInfo *ThisDecl,
+                                SmallVectorImpl<char> &Str) {
+  ASTContext &Context = ThisDecl->CurrentDecl->getASTContext();
+  const LangOptions &LangOpts = Context.getLangOpts();
+  llvm::raw_svector_ostream OS(Str);
+  PrintingPolicy PPolicy(LangOpts);
+  PPolicy.PolishForDeclaration = true;
+  PPolicy.TerseOutput = true;
+  ThisDecl->CurrentDecl->print(OS, PPolicy,
+                               /*Indentation*/0, /*PrintInstantiation*/false);
+}
+
+void CommentASTToXMLConverter::formatTextOfDeclaration(
+    const DeclInfo *DI, SmallString<128> &Declaration) {
+  // FIXME. formatting API expects null terminated input string.
+  // There might be more efficient way of doing this.
+  std::string StringDecl = Declaration.str();
+
+  // Formatter specific code.
+  // Form a unique in memory buffer name.
+  SmallString<128> filename;
+  filename += "xmldecl";
+  filename += llvm::utostr(FormatInMemoryUniqueId);
+  filename += ".xd";
+  FileID ID = FormatRewriterContext.createInMemoryFile(filename, StringDecl);
+  SourceLocation Start = FormatRewriterContext.Sources.getLocForStartOfFile(ID)
+      .getLocWithOffset(0);
+  unsigned Length = Declaration.size();
+
+  std::vector<CharSourceRange> Ranges(
+      1, CharSourceRange::getCharRange(Start, Start.getLocWithOffset(Length)));
+  ASTContext &Context = DI->CurrentDecl->getASTContext();
+  const LangOptions &LangOpts = Context.getLangOpts();
+  Lexer Lex(ID, FormatRewriterContext.Sources.getBuffer(ID),
+            FormatRewriterContext.Sources, LangOpts);
+  tooling::Replacements Replace = reformat(
+      format::getLLVMStyle(), Lex, FormatRewriterContext.Sources, Ranges);
+  applyAllReplacements(Replace, FormatRewriterContext.Rewrite);
+  Declaration = FormatRewriterContext.getRewrittenText(ID);
+}
+
+} // end unnamed namespace
+
+void CommentASTToXMLConverter::visitTextComment(const TextComment *C) {
+  appendToResultWithXMLEscaping(C->getText());
+}
+
+void CommentASTToXMLConverter::visitInlineCommandComment(
+    const InlineCommandComment *C) {
+  // Nothing to render if no arguments supplied.
+  if (C->getNumArgs() == 0)
+    return;
+
+  // Nothing to render if argument is empty.
+  StringRef Arg0 = C->getArgText(0);
+  if (Arg0.empty())
+    return;
+
+  switch (C->getRenderKind()) {
+  case InlineCommandComment::RenderNormal:
+    for (unsigned i = 0, e = C->getNumArgs(); i != e; ++i) {
+      appendToResultWithXMLEscaping(C->getArgText(i));
+      Result << " ";
+    }
+    return;
+  case InlineCommandComment::RenderBold:
+    assert(C->getNumArgs() == 1);
+    Result << "<bold>";
+    appendToResultWithXMLEscaping(Arg0);
+    Result << "</bold>";
+    return;
+  case InlineCommandComment::RenderMonospaced:
+    assert(C->getNumArgs() == 1);
+    Result << "<monospaced>";
+    appendToResultWithXMLEscaping(Arg0);
+    Result << "</monospaced>";
+    return;
+  case InlineCommandComment::RenderEmphasized:
+    assert(C->getNumArgs() == 1);
+    Result << "<emphasized>";
+    appendToResultWithXMLEscaping(Arg0);
+    Result << "</emphasized>";
+    return;
+  }
+}
+
+void CommentASTToXMLConverter::visitHTMLStartTagComment(
+    const HTMLStartTagComment *C) {
+  Result << "<rawHTML><![CDATA[";
+  printHTMLStartTagComment(C, Result);
+  Result << "]]></rawHTML>";
+}
+
+void
+CommentASTToXMLConverter::visitHTMLEndTagComment(const HTMLEndTagComment *C) {
+  Result << "<rawHTML>&lt;/" << C->getTagName() << "&gt;</rawHTML>";
+}
+
+void
+CommentASTToXMLConverter::visitParagraphComment(const ParagraphComment *C) {
+  appendParagraphCommentWithKind(C, StringRef());
+}
+
+void CommentASTToXMLConverter::appendParagraphCommentWithKind(
+                                  const ParagraphComment *C,
+                                  StringRef ParagraphKind) {
+  if (C->isWhitespace())
+    return;
+
+  if (ParagraphKind.empty())
+    Result << "<Para>";
+  else
+    Result << "<Para kind=\"" << ParagraphKind << "\">";
+
+  for (Comment::child_iterator I = C->child_begin(), E = C->child_end();
+       I != E; ++I) {
+    visit(*I);
+  }
+  Result << "</Para>";
+}
+
+void CommentASTToXMLConverter::visitBlockCommandComment(
+    const BlockCommandComment *C) {
+  StringRef ParagraphKind;
+
+  switch (C->getCommandID()) {
+  case CommandTraits::KCI_attention:
+  case CommandTraits::KCI_author:
+  case CommandTraits::KCI_authors:
+  case CommandTraits::KCI_bug:
+  case CommandTraits::KCI_copyright:
+  case CommandTraits::KCI_date:
+  case CommandTraits::KCI_invariant:
+  case CommandTraits::KCI_note:
+  case CommandTraits::KCI_post:
+  case CommandTraits::KCI_pre:
+  case CommandTraits::KCI_remark:
+  case CommandTraits::KCI_remarks:
+  case CommandTraits::KCI_sa:
+  case CommandTraits::KCI_see:
+  case CommandTraits::KCI_since:
+  case CommandTraits::KCI_todo:
+  case CommandTraits::KCI_version:
+  case CommandTraits::KCI_warning:
+    ParagraphKind = C->getCommandName(Traits);
+  default:
+    break;
+  }
+
+  appendParagraphCommentWithKind(C->getParagraph(), ParagraphKind);
+}
+
+void CommentASTToXMLConverter::visitParamCommandComment(
+    const ParamCommandComment *C) {
+  Result << "<Parameter><Name>";
+  appendToResultWithXMLEscaping(C->isParamIndexValid()
+                                    ? C->getParamName(FC)
+                                    : C->getParamNameAsWritten());
+  Result << "</Name>";
+
+  if (C->isParamIndexValid()) {
+    if (C->isVarArgParam())
+      Result << "<IsVarArg />";
+    else
+      Result << "<Index>" << C->getParamIndex() << "</Index>";
+  }
+
+  Result << "<Direction isExplicit=\"" << C->isDirectionExplicit() << "\">";
+  switch (C->getDirection()) {
+  case ParamCommandComment::In:
+    Result << "in";
+    break;
+  case ParamCommandComment::Out:
+    Result << "out";
+    break;
+  case ParamCommandComment::InOut:
+    Result << "in,out";
+    break;
+  }
+  Result << "</Direction><Discussion>";
+  visit(C->getParagraph());
+  Result << "</Discussion></Parameter>";
+}
+
+void CommentASTToXMLConverter::visitTParamCommandComment(
+                                  const TParamCommandComment *C) {
+  Result << "<Parameter><Name>";
+  appendToResultWithXMLEscaping(C->isPositionValid() ? C->getParamName(FC)
+                                : C->getParamNameAsWritten());
+  Result << "</Name>";
+
+  if (C->isPositionValid() && C->getDepth() == 1) {
+    Result << "<Index>" << C->getIndex(0) << "</Index>";
+  }
+
+  Result << "<Discussion>";
+  visit(C->getParagraph());
+  Result << "</Discussion></Parameter>";
+}
+
+void CommentASTToXMLConverter::visitVerbatimBlockComment(
+                                  const VerbatimBlockComment *C) {
+  unsigned NumLines = C->getNumLines();
+  if (NumLines == 0)
+    return;
+
+  switch (C->getCommandID()) {
+  case CommandTraits::KCI_code:
+    Result << "<Verbatim xml:space=\"preserve\" kind=\"code\">";
+    break;
+  default:
+    Result << "<Verbatim xml:space=\"preserve\" kind=\"verbatim\">";
+    break;
+  }
+  for (unsigned i = 0; i != NumLines; ++i) {
+    appendToResultWithXMLEscaping(C->getText(i));
+    if (i + 1 != NumLines)
+      Result << '\n';
+  }
+  Result << "</Verbatim>";
+}
+
+void CommentASTToXMLConverter::visitVerbatimBlockLineComment(
+                                  const VerbatimBlockLineComment *C) {
+  llvm_unreachable("should not see this AST node");
+}
+
+void CommentASTToXMLConverter::visitVerbatimLineComment(
+                                  const VerbatimLineComment *C) {
+  Result << "<Verbatim xml:space=\"preserve\" kind=\"verbatim\">";
+  appendToResultWithXMLEscaping(C->getText());
+  Result << "</Verbatim>";
+}
+
+void CommentASTToXMLConverter::visitFullComment(const FullComment *C) {
+  FullCommentParts Parts(C, Traits);
+
+  const DeclInfo *DI = C->getDeclInfo();
+  StringRef RootEndTag;
+  if (DI) {
+    switch (DI->getKind()) {
+    case DeclInfo::OtherKind:
+      RootEndTag = "</Other>";
+      Result << "<Other";
+      break;
+    case DeclInfo::FunctionKind:
+      RootEndTag = "</Function>";
+      Result << "<Function";
+      switch (DI->TemplateKind) {
+      case DeclInfo::NotTemplate:
+        break;
+      case DeclInfo::Template:
+        Result << " templateKind=\"template\"";
+        break;
+      case DeclInfo::TemplateSpecialization:
+        Result << " templateKind=\"specialization\"";
+        break;
+      case DeclInfo::TemplatePartialSpecialization:
+        llvm_unreachable("partial specializations of functions "
+                         "are not allowed in C++");
+      }
+      if (DI->IsInstanceMethod)
+        Result << " isInstanceMethod=\"1\"";
+      if (DI->IsClassMethod)
+        Result << " isClassMethod=\"1\"";
+      break;
+    case DeclInfo::ClassKind:
+      RootEndTag = "</Class>";
+      Result << "<Class";
+      switch (DI->TemplateKind) {
+      case DeclInfo::NotTemplate:
+        break;
+      case DeclInfo::Template:
+        Result << " templateKind=\"template\"";
+        break;
+      case DeclInfo::TemplateSpecialization:
+        Result << " templateKind=\"specialization\"";
+        break;
+      case DeclInfo::TemplatePartialSpecialization:
+        Result << " templateKind=\"partialSpecialization\"";
+        break;
+      }
+      break;
+    case DeclInfo::VariableKind:
+      RootEndTag = "</Variable>";
+      Result << "<Variable";
+      break;
+    case DeclInfo::NamespaceKind:
+      RootEndTag = "</Namespace>";
+      Result << "<Namespace";
+      break;
+    case DeclInfo::TypedefKind:
+      RootEndTag = "</Typedef>";
+      Result << "<Typedef";
+      break;
+    case DeclInfo::EnumKind:
+      RootEndTag = "</Enum>";
+      Result << "<Enum";
+      break;
+    }
+
+    {
+      // Print line and column number.
+      SourceLocation Loc = DI->CurrentDecl->getLocation();
+      std::pair<FileID, unsigned> LocInfo = SM.getDecomposedLoc(Loc);
+      FileID FID = LocInfo.first;
+      unsigned FileOffset = LocInfo.second;
+
+      if (!FID.isInvalid()) {
+        if (const FileEntry *FE = SM.getFileEntryForID(FID)) {
+          Result << " file=\"";
+          appendToResultWithXMLEscaping(FE->getName());
+          Result << "\"";
+        }
+        Result << " line=\"" << SM.getLineNumber(FID, FileOffset)
+               << "\" column=\"" << SM.getColumnNumber(FID, FileOffset)
+               << "\"";
+      }
+    }
+
+    // Finish the root tag.
+    Result << ">";
+
+    bool FoundName = false;
+    if (const NamedDecl *ND = dyn_cast<NamedDecl>(DI->CommentDecl)) {
+      if (DeclarationName DeclName = ND->getDeclName()) {
+        Result << "<Name>";
+        std::string Name = DeclName.getAsString();
+        appendToResultWithXMLEscaping(Name);
+        FoundName = true;
+        Result << "</Name>";
+      }
+    }
+    if (!FoundName)
+      Result << "<Name>&lt;anonymous&gt;</Name>";
+
+    {
+      // Print USR.
+      SmallString<128> USR;
+      generateUSRForDecl(DI->CommentDecl, USR);
+      if (!USR.empty()) {
+        Result << "<USR>";
+        appendToResultWithXMLEscaping(USR);
+        Result << "</USR>";
+      }
+    }
+  } else {
+    // No DeclInfo -- just emit some root tag and name tag.
+    RootEndTag = "</Other>";
+    Result << "<Other><Name>unknown</Name>";
+  }
+
+  if (Parts.Headerfile) {
+    Result << "<Headerfile>";
+    visit(Parts.Headerfile);
+    Result << "</Headerfile>";
+  }
+
+  {
+    // Pretty-print the declaration.
+    Result << "<Declaration>";
+    SmallString<128> Declaration;
+    getSourceTextOfDeclaration(DI, Declaration);
+    formatTextOfDeclaration(DI, Declaration);
+    appendToResultWithXMLEscaping(Declaration);
+    Result << "</Declaration>";
+  }
+
+  bool FirstParagraphIsBrief = false;
+  if (Parts.Brief) {
+    Result << "<Abstract>";
+    visit(Parts.Brief);
+    Result << "</Abstract>";
+  } else if (Parts.FirstParagraph) {
+    Result << "<Abstract>";
+    visit(Parts.FirstParagraph);
+    Result << "</Abstract>";
+    FirstParagraphIsBrief = true;
+  }
+
+  if (Parts.TParams.size() != 0) {
+    Result << "<TemplateParameters>";
+    for (unsigned i = 0, e = Parts.TParams.size(); i != e; ++i)
+      visit(Parts.TParams[i]);
+    Result << "</TemplateParameters>";
+  }
+
+  if (Parts.Params.size() != 0) {
+    Result << "<Parameters>";
+    for (unsigned i = 0, e = Parts.Params.size(); i != e; ++i)
+      visit(Parts.Params[i]);
+    Result << "</Parameters>";
+  }
+
+  if (Parts.Exceptions.size() != 0) {
+    Result << "<Exceptions>";
+    for (unsigned i = 0, e = Parts.Exceptions.size(); i != e; ++i)
+      visit(Parts.Exceptions[i]);
+    Result << "</Exceptions>";
+  }
+
+  if (Parts.Returns.size() != 0) {
+    Result << "<ResultDiscussion>";
+    for (unsigned i = 0, e = Parts.Returns.size(); i != e; ++i)
+      visit(Parts.Returns[i]);
+    Result << "</ResultDiscussion>";
+  }
+
+  if (DI->CommentDecl->hasAttrs()) {
+    const AttrVec &Attrs = DI->CommentDecl->getAttrs();
+    for (unsigned i = 0, e = Attrs.size(); i != e; i++) {
+      const AvailabilityAttr *AA = dyn_cast<AvailabilityAttr>(Attrs[i]);
+      if (!AA) {
+        if (const DeprecatedAttr *DA = dyn_cast<DeprecatedAttr>(Attrs[i])) {
+          if (DA->getMessage().empty())
+            Result << "<Deprecated/>";
+          else {
+            Result << "<Deprecated>";
+            appendToResultWithXMLEscaping(DA->getMessage());
+            Result << "</Deprecated>";
+          }
+        }
+        else if (const UnavailableAttr *UA = dyn_cast<UnavailableAttr>(Attrs[i])) {
+          if (UA->getMessage().empty())
+            Result << "<Unavailable/>";
+          else {
+            Result << "<Unavailable>";
+            appendToResultWithXMLEscaping(UA->getMessage());
+            Result << "</Unavailable>";
+          }
+        }
+        continue;
+      }
+
+      // 'availability' attribute.
+      Result << "<Availability";
+      StringRef Distribution;
+      if (AA->getPlatform()) {
+        Distribution = AvailabilityAttr::getPrettyPlatformName(
+                                        AA->getPlatform()->getName());
+        if (Distribution.empty())
+          Distribution = AA->getPlatform()->getName();
+      }
+      Result << " distribution=\"" << Distribution << "\">";
+      VersionTuple IntroducedInVersion = AA->getIntroduced();
+      if (!IntroducedInVersion.empty()) {
+        Result << "<IntroducedInVersion>"
+               << IntroducedInVersion.getAsString()
+               << "</IntroducedInVersion>";
+      }
+      VersionTuple DeprecatedInVersion = AA->getDeprecated();
+      if (!DeprecatedInVersion.empty()) {
+        Result << "<DeprecatedInVersion>"
+               << DeprecatedInVersion.getAsString()
+               << "</DeprecatedInVersion>";
+      }
+      VersionTuple RemovedAfterVersion = AA->getObsoleted();
+      if (!RemovedAfterVersion.empty()) {
+        Result << "<RemovedAfterVersion>"
+               << RemovedAfterVersion.getAsString()
+               << "</RemovedAfterVersion>";
+      }
+      StringRef DeprecationSummary = AA->getMessage();
+      if (!DeprecationSummary.empty()) {
+        Result << "<DeprecationSummary>";
+        appendToResultWithXMLEscaping(DeprecationSummary);
+        Result << "</DeprecationSummary>";
+      }
+      if (AA->getUnavailable())
+        Result << "<Unavailable/>";
+      Result << "</Availability>";
+    }
+  }
+
+  {
+    bool StartTagEmitted = false;
+    for (unsigned i = 0, e = Parts.MiscBlocks.size(); i != e; ++i) {
+      const Comment *C = Parts.MiscBlocks[i];
+      if (FirstParagraphIsBrief && C == Parts.FirstParagraph)
+        continue;
+      if (!StartTagEmitted) {
+        Result << "<Discussion>";
+        StartTagEmitted = true;
+      }
+      visit(C);
+    }
+    if (StartTagEmitted)
+      Result << "</Discussion>";
+  }
+
+  Result << RootEndTag;
+
+  Result.flush();
+}
+
+void CommentASTToXMLConverter::appendToResultWithXMLEscaping(StringRef S) {
+  for (StringRef::iterator I = S.begin(), E = S.end(); I != E; ++I) {
+    const char C = *I;
+    switch (C) {
+    case '&':
+      Result << "&amp;";
+      break;
+    case '<':
+      Result << "&lt;";
+      break;
+    case '>':
+      Result << "&gt;";
+      break;
+    case '"':
+      Result << "&quot;";
+      break;
+    case '\'':
+      Result << "&apos;";
+      break;
+    default:
+      Result << C;
+      break;
+    }
+  }
+}
+
+void CommentToXMLConverter::convertCommentToHTML(const FullComment *FC,
+                                                 SmallVectorImpl<char> &HTML,
+                                                 const ASTContext &Context) {
+  CommentASTToHTMLConverter Converter(FC, HTML,
+                                      Context.getCommentCommandTraits());
+  Converter.visit(FC);
+}
+
+void CommentToXMLConverter::convertHTMLTagNodeToText(
+    const comments::HTMLTagComment *HTC, SmallVectorImpl<char> &Text,
+    const ASTContext &Context) {
+  CommentASTToHTMLConverter Converter(0, Text,
+                                      Context.getCommentCommandTraits());
+  Converter.visit(HTC);
+}
+
+void CommentToXMLConverter::convertCommentToXML(const FullComment *FC,
+                                                SmallVectorImpl<char> &XML,
+                                                const ASTContext &Context) {
+  if (!FormatContext) {
+    FormatContext = new SimpleFormatContext(Context.getLangOpts());
+  } else if ((FormatInMemoryUniqueId % 1000) == 0) {
+    // Delete after some number of iterations, so the buffers don't grow
+    // too large.
+    delete FormatContext;
+    FormatContext = new SimpleFormatContext(Context.getLangOpts());
+  }
+
+  CommentASTToXMLConverter Converter(FC, XML, Context.getCommentCommandTraits(),
+                                     Context.getSourceManager(), *FormatContext,
+                                     FormatInMemoryUniqueId++);
+  Converter.visit(FC);
+}
+