MC: Simple example parser for MC assembly markup.

Nothing fancy, just a simple demonstration parser.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@167181 91177308-0d34-0410-b5e6-96231b3b80d8

4 files changed, 269 insertions, 0 deletions

diff --git a/tools/llvm-mcmarkup/CMakeLists.txt b/tools/llvm-mcmarkup/CMakeLists.txt
new file mode 100644
index 0000000000..0a51e99f19
--- /dev/null
+++ b/tools/llvm-mcmarkup/CMakeLists.txt
@@ -0,0 +1,5 @@
+set(LLVM_LINK_COMPONENTS support)
+
+add_llvm_tool(llvm-mcmarkup
+  llvm-mcmarkup.cpp
+  )
diff --git a/tools/llvm-mcmarkup/LLVMBuild.txt b/tools/llvm-mcmarkup/LLVMBuild.txt
new file mode 100644
index 0000000000..6423493a54
--- /dev/null
+++ b/tools/llvm-mcmarkup/LLVMBuild.txt
@@ -0,0 +1,22 @@
+;===- ./tools/llvm-mcmarkup/LLVMBuild.txt ----------------------*- Conf -*--===;
+;
+;                     The LLVM Compiler Infrastructure
+;
+; This file is distributed under the University of Illinois Open Source
+; License. See LICENSE.TXT for details.
+;
+;===------------------------------------------------------------------------===;
+;
+; This is an LLVMBuild description file for the components in this subdirectory.
+;
+; For more information on the LLVMBuild system, please see:
+;
+;   http://llvm.org/docs/LLVMBuild.html
+;
+;===------------------------------------------------------------------------===;
+
+[component_0]
+type = Tool
+name = llvm-mcmarkup
+parent = Tools
+required_libraries = Support
diff --git a/tools/llvm-mcmarkup/Makefile b/tools/llvm-mcmarkup/Makefile
new file mode 100644
index 0000000000..5633a9c301
--- /dev/null
+++ b/tools/llvm-mcmarkup/Makefile
@@ -0,0 +1,17 @@
+##===- tools/llvm-mcmarkup/Makefile ------------------------*- Makefile -*-===##
+#
+#                     The LLVM Compiler Infrastructure
+#
+# This file is distributed under the University of Illinois Open Source
+# License. See LICENSE.TXT for details.
+#
+##===----------------------------------------------------------------------===##
+
+LEVEL := ../..
+TOOLNAME := llvm-mcmarkup
+LINK_COMPONENTS := support
+
+# This tool has no plugins, optimize startup time.
+TOOL_NO_EXPORTS = 1
+
+include $(LEVEL)/Makefile.common
diff --git a/tools/llvm-mcmarkup/llvm-mcmarkup.cpp b/tools/llvm-mcmarkup/llvm-mcmarkup.cpp
new file mode 100644
index 0000000000..888761f10f
--- /dev/null
+++ b/tools/llvm-mcmarkup/llvm-mcmarkup.cpp
@@ -0,0 +1,225 @@
+//===-- llvm-mcmarkup.cpp - Parse the MC assembly markup tags -------------===//
+//
+//                     The LLVM Compiler Infrastructure
+//
+// This file is distributed under the University of Illinois Open Source
+// License. See LICENSE.TXT for details.
+//
+//===----------------------------------------------------------------------===//
+//
+// Example simple parser implementation for the MC assembly markup language.
+//
+//===----------------------------------------------------------------------===//
+
+#include "llvm/ADT/OwningPtr.h"
+#include "llvm/Support/CommandLine.h"
+#include "llvm/Support/Format.h"
+#include "llvm/Support/ManagedStatic.h"
+#include "llvm/Support/MemoryBuffer.h"
+#include "llvm/Support/PrettyStackTrace.h"
+#include "llvm/Support/Signals.h"
+#include "llvm/Support/SourceMgr.h"
+#include "llvm/Support/raw_ostream.h"
+#include "llvm/Support/system_error.h"
+using namespace llvm;
+
+static cl::list<std::string>
+       InputFilenames(cl::Positional, cl::desc("<input files>"),
+                      cl::ZeroOrMore);
+static cl::opt<bool>
+DumpTags("dump-tags", cl::desc("List all tags encountered in input"));
+
+static StringRef ToolName;
+
+/// Trivial lexer for the markup parser. Input is always handled a character
+/// at a time. The lexer just encapsulates EOF and lookahead handling.
+class MarkupLexer {
+  StringRef::const_iterator Start;
+  StringRef::const_iterator CurPtr;
+  StringRef::const_iterator End;
+public:
+  MarkupLexer(StringRef Source)
+    : Start(Source.begin()), CurPtr(Source.begin()), End(Source.end()) {}
+  // When processing non-markup, input is consumed a character at a time.
+  bool isEOF() { return CurPtr == End; }
+  int getNextChar() {
+    if (CurPtr == End) return EOF;
+    return *CurPtr++;
+  }
+  int peekNextChar() {
+    if (CurPtr == End) return EOF;
+    return *CurPtr;
+  }
+  StringRef::const_iterator getPosition() const { return CurPtr; }
+};
+
+/// A markup tag is a name and a (usually empty) list of modifiers.
+class MarkupTag {
+  StringRef Name;
+  StringRef Modifiers;
+  SMLoc StartLoc;
+public:
+  MarkupTag(StringRef n, StringRef m, SMLoc Loc)
+    : Name(n), Modifiers(m), StartLoc(Loc) {}
+  StringRef getName() const { return Name; }
+  StringRef getModifiers() const { return Modifiers; }
+  SMLoc getLoc() const { return StartLoc; }
+};
+
+/// A simple parser implementation for creating MarkupTags from input text.
+class MarkupParser {
+  MarkupLexer &Lex;
+  SourceMgr &SM;
+public:
+  MarkupParser(MarkupLexer &lex, SourceMgr &SrcMgr) : Lex(lex), SM(SrcMgr) {}
+  /// Create a MarkupTag from the current position in the MarkupLexer.
+  /// The parseTag() method should be called when the lexer has processed
+  /// the opening '<' character. Input will be consumed up to and including
+  /// the ':' which terminates the tag open.
+  MarkupTag parseTag();
+  /// Issue a diagnostic and terminate program execution.
+  void FatalError(SMLoc Loc, StringRef Msg);
+};
+
+void MarkupParser::FatalError(SMLoc Loc, StringRef Msg) {
+  SM.PrintMessage(Loc, SourceMgr::DK_Error, Msg);
+  exit(1);
+}
+
+// Example handler for when a tag is recognized.
+static void processStartTag(MarkupTag &Tag) {
+  // If we're just printing the tags, do that, otherwise do some simple
+  // colorization.
+  if (DumpTags) {
+    outs() << Tag.getName();
+    if (Tag.getModifiers().size())
+      outs() << " " << Tag.getModifiers();
+    outs() << "\n";
+    return;
+  }
+
+  if (!outs().has_colors())
+    return;
+  // Color registers as red and immediates as cyan. Those don't have nested
+  // tags, so don't bother keeping a stack of colors to reset to.
+  if (Tag.getName() == "reg")
+    outs().changeColor(raw_ostream::RED);
+  else if (Tag.getName() == "imm")
+    outs().changeColor(raw_ostream::CYAN);
+}
+
+// Example handler for when the end of a tag is recognized.
+static void processEndTag(MarkupTag &Tag) {
+  // If we're printing the tags, there's nothing more to do here. Otherwise,
+  // set the color back the normal.
+  if (DumpTags)
+    return;
+  if (!outs().has_colors())
+    return;
+  // Just reset to basic white.
+  outs().changeColor(raw_ostream::WHITE, false);
+}
+
+MarkupTag MarkupParser::parseTag() {
+  // First off, extract the tag into it's own StringRef so we can look at it
+  // outside of the context of consuming input.
+  StringRef::const_iterator Start = Lex.getPosition();
+  SMLoc Loc = SMLoc::getFromPointer(Start - 1);
+  while(Lex.getNextChar() != ':') {
+    // EOF is an error.
+    if (Lex.isEOF())
+      FatalError(SMLoc::getFromPointer(Start), "unterminated markup tag");
+  }
+  StringRef RawTag(Start, Lex.getPosition() - Start - 1);
+  std::pair<StringRef, StringRef> SplitTag = RawTag.split(' ');
+  return MarkupTag(SplitTag.first, SplitTag.second, Loc);
+}
+
+static void parseMCMarkup(StringRef Filename) {
+  OwningPtr<MemoryBuffer> BufferPtr;
+  if (error_code ec = MemoryBuffer::getFileOrSTDIN(Filename, BufferPtr)) {
+    errs() << ToolName << ": " << ec.message() << '\n';
+    return;
+  }
+  MemoryBuffer *Buffer = BufferPtr.take();
+
+  SourceMgr SrcMgr;
+
+  // Tell SrcMgr about this buffer, which is what the parser will pick up.
+  SrcMgr.AddNewSourceBuffer(Buffer, SMLoc());
+
+  StringRef InputSource = Buffer->getBuffer();
+  MarkupLexer Lex(InputSource);
+  MarkupParser Parser(Lex, SrcMgr);
+
+  SmallVector<MarkupTag, 4> TagStack;
+
+  for (int CurChar = Lex.getNextChar();
+       CurChar != EOF;
+       CurChar = Lex.getNextChar()) {
+    switch (CurChar) {
+    case '<': {
+      // A "<<" is output as a literal '<' and does not start a markup tag.
+      if (Lex.peekNextChar() == '<') {
+        (void)Lex.getNextChar();
+        break;
+      }
+      // Parse the markup entry.
+      TagStack.push_back(Parser.parseTag());
+
+      // Do any special handling for the start of a tag.
+      processStartTag(TagStack.back());
+      continue;
+    }
+    case '>': {
+      SMLoc Loc = SMLoc::getFromPointer(Lex.getPosition() - 1);
+      // A ">>" is output as a literal '>' and does not end a markup tag.
+      if (Lex.peekNextChar() == '>') {
+        (void)Lex.getNextChar();
+        break;
+      }
+      // Close out the innermost tag.
+      if (TagStack.empty())
+        Parser.FatalError(Loc, "'>' without matching '<'");
+
+      // Do any special handling for the end of a tag.
+      processEndTag(TagStack.back());
+
+      TagStack.pop_back();
+      continue;
+    }
+    default:
+      break;
+    }
+    // For anything else, just echo the character back out.
+    if (!DumpTags && CurChar != EOF)
+      outs() << (char)CurChar;
+  }
+
+  // If there are any unterminated markup tags, issue diagnostics for them.
+  while (!TagStack.empty()) {
+    MarkupTag &Tag = TagStack.back();
+    SrcMgr.PrintMessage(Tag.getLoc(), SourceMgr::DK_Error,
+                        "unterminated markup tag");
+    TagStack.pop_back();
+  }
+}
+
+int main(int argc, char **argv) {
+  // Print a stack trace if we signal out.
+  sys::PrintStackTraceOnErrorSignal();
+  PrettyStackTraceProgram X(argc, argv);
+
+  llvm_shutdown_obj Y;  // Call llvm_shutdown() on exit.
+  cl::ParseCommandLineOptions(argc, argv, "llvm MC markup parser\n");
+
+  ToolName = argv[0];
+
+  // If no input files specified, read from stdin.
+  if (InputFilenames.size() == 0)
+    InputFilenames.push_back("-");
+
+  std::for_each(InputFilenames.begin(), InputFilenames.end(),
+                parseMCMarkup);
+  return 0;
+}