From 49a70d832d25ca29232cb918b57420a0b2fb1b8d Mon Sep 17 00:00:00 2001
From: Mark Olesen <Mark.Olesen@Germany>
Date: Sun, 7 Feb 2010 14:31:40 +0100
Subject: [PATCH] ENH: add C++-based wmkdepend parser (uses Coco/R grammar).
- This avoids dependency on lex/flex and provides better encapsulation
for buffer switching. As a result, the maximum number of open files
only corresponds to the include depth.
---
wmake/src/Makefile | 16 +-
wmake/src/wmkdepend.cpp | 164 +++++
wmake/src/wmkdependParser.atg | 497 ++++++++++++++++
wmake/src/wmkdependParser.cpp | 575 ++++++++++++++++++
wmake/src/wmkdependParser.h | 310 ++++++++++
wmake/src/wmkdependScanner.cpp | 1022 ++++++++++++++++++++++++++++++++
wmake/src/wmkdependScanner.h | 477 +++++++++++++++
7 files changed, 3058 insertions(+), 3 deletions(-)
create mode 100644 wmake/src/wmkdepend.cpp
create mode 100644 wmake/src/wmkdependParser.atg
create mode 100644 wmake/src/wmkdependParser.cpp
create mode 100644 wmake/src/wmkdependParser.h
create mode 100644 wmake/src/wmkdependScanner.cpp
create mode 100644 wmake/src/wmkdependScanner.h
diff --git a/wmake/src/Makefile b/wmake/src/Makefile
index a74c250f881..cac2a639a52 100644
--- a/wmake/src/Makefile
+++ b/wmake/src/Makefile
@@ -2,7 +2,7 @@
# ========= |
# \\ / F ield | OpenFOAM: The Open Source CFD Toolbox
# \\ / O peration |
-# \\ / A nd | Copyright (C) 1991-2009 OpenCFD Ltd.
+# \\ / A nd | Copyright (C) 1991-2010 OpenCFD Ltd.
# \\/ M anipulation |
#------------------------------------------------------------------------------
# License
@@ -60,16 +60,17 @@ include $(RULES)/$(WM_LINK_LANGUAGE)
# targets
#------------------------------------------------------------------------------
-all: $(BIN)/dirToString $(BIN)/wmkdep
+all: $(BIN)/dirToString $(BIN)/wmkdep $(BIN)/wmkdepend
clean:
- rm -f $(BIN)/dirToString $(BIN)/wmkdep 2>/dev/null
+ rm -f $(BIN)/dirToString $(BIN)/wmkdep $(BIN)/wmkdepend 2>/dev/null
$(BIN)/dirToString: dirToString.c
@mkdir -p $(BIN)
$(cc) $(cFLAGS) dirToString.c -o $(BIN)/dirToString
+
$(BIN)/wmkdep: wmkdep.l
@mkdir -p $(BIN)
flex wmkdep.l
@@ -77,4 +78,13 @@ $(BIN)/wmkdep: wmkdep.l
@rm -f lex.yy.c 2>/dev/null
+# for bootstrapping - use generated files directly (instead of from .atg file)
+$(BIN)/wmkdepend: wmkdepend.cpp \
+ wmkdependParser.cpp wmkdependScanner.cpp \
+ wmkdependParser.h wmkdependScanner.h
+ @mkdir -p $(BIN)
+ $(CC) $(c++FLAGS) \
+ wmkdepend.cpp wmkdependParser.cpp wmkdependScanner.cpp \
+ -o $(BIN)/wmkdepend
+
#------------------------------------------------------------------------------
diff --git a/wmake/src/wmkdepend.cpp b/wmake/src/wmkdepend.cpp
new file mode 100644
index 00000000000..159fb3849b9
--- /dev/null
+++ b/wmake/src/wmkdepend.cpp
@@ -0,0 +1,164 @@
+/*---------------------------------------------------------------------------*\
+ ========= |
+ \\ / F ield | OpenFOAM: The Open Source CFD Toolbox
+ \\ / O peration |
+ \\ / A nd | Copyright (C) 2010-2010 OpenCFD Ltd.
+ \\/ M anipulation |
+------------------------------------------------------------------------------
+License
+ This file is part of OpenFOAM.
+
+ OpenFOAM is free software; you can redistribute it and/or modify it
+ under the terms of the GNU General Public License as published by the
+ Free Software Foundation; either version 2 of the License, or (at your
+ option) any later version.
+
+ OpenFOAM is distributed in the hope that it will be useful, but WITHOUT
+ ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License
+ for more details.
+
+ You should have received a copy of the GNU General Public License
+ along with OpenFOAM; if not, write to the Free Software Foundation,
+ Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA
+
+Application
+ wmkdepend
+
+Description
+ A fast dependency list generator that emulates the behaviour and
+ output of cpp -M. However, the output contains no duplications and
+ is ~40% faster than cpp.
+
+ The algorithm uses flex to scan for includes and searches the files
+ found. Each file is entered into a hash table so that files are scanned
+ only once. This is why this program is faster than cpp.
+
+Usage
+ wmkdep [ -Idirectory ... -Idirectory ] filename
+
+\*---------------------------------------------------------------------------*/
+
+#include <cstdio>
+#include <stdlib.h>
+#include <cstring>
+
+#include "wmkdependParser.h"
+
+// Note: since we use the Coco/R default error messages, we must use
+// wide streams for stderr.
+
+void printUsage(const char* message = NULL)
+{
+ if (message)
+ {
+ fwprintf(stderr, L"\nError: %s\n\n", message);
+ }
+
+ fwprintf
+ (
+ stderr,
+ L"Usage: wmkdepend [ -Idirectory ... -Idirectory ] filename\n"
+ );
+}
+
+
+int main(int argc, char* argv[])
+{
+ if (argc == 1)
+ {
+ printUsage("Error: input file not supplied");
+ ::exit(1);
+ }
+
+ for (int i=1; i < argc; i++)
+ {
+ if (strncmp(argv[i], "-I", 2) == 0 && strlen(argv[i]) > 2)
+ {
+ std::string dirName(argv[i] + 2);
+
+ // add trailing slash if required
+ if (dirName.rfind('/') != dirName.size()-1)
+ {
+ dirName += '/';
+ }
+
+ wmake::Parser::includeDirs.push_back(dirName);
+ }
+ }
+
+ std::string sourceFile(argv[argc-1]);
+
+ fwprintf
+ (
+ stderr,
+ L"Making dependency list for source file %s\n",
+ sourceFile.c_str()
+ );
+
+ std::string::size_type basePos = sourceFile.rfind('/');
+ if (basePos == std::string::npos)
+ {
+ basePos = 0;
+ }
+ else
+ {
+ basePos++;
+ }
+
+ std::string::size_type dotPos = sourceFile.rfind('.');
+ if
+ (
+ dotPos == std::string::npos
+ || dotPos == sourceFile.size()-1
+ || dotPos <= basePos
+ )
+ {
+ fwprintf
+ (
+ stderr,
+ L"Cannot find extension in source file name %s\n",
+ sourceFile.c_str()
+ );
+ ::exit(1);
+ }
+
+ std::string depFile = sourceFile.substr(0, dotPos);
+ depFile += ".dep";
+
+ const std::string sourceExt = sourceFile.substr(dotPos);
+ if (sourceExt == ".java")
+ {
+ // import directories to ignore
+ wmake::Parser::ignoreDir("java.*");
+ wmake::Parser::ignoreDir("org.*");
+ wmake::Parser::ignoreDir("com.*");
+ wmake::Parser::ignoreDir("sunw.*");
+ wmake::Parser::ignoreDir("sun.*");
+ wmake::Parser::ignoreDir("launcher.*");
+
+ std::cout
+ << "$(CLASSES_DIR)/"
+ << sourceFile.substr(basePos, dotPos - basePos) << ".class: "
+ << depFile << "\n";
+ }
+ else
+ {
+ std::cout
+ << "$(OBJECTS_DIR)/"
+ << sourceFile.substr(basePos, dotPos - basePos) << ".o: "
+ << depFile << "\n";
+ }
+
+
+ wmake::Parser::sourceFile = sourceFile;
+ wmake::Parser::depFile = depFile;
+
+ wmake::Parser::includeFile(sourceFile);
+
+ return 0;
+}
+
+
+
+/*****************************************************************************/
diff --git a/wmake/src/wmkdependParser.atg b/wmake/src/wmkdependParser.atg
new file mode 100644
index 00000000000..2da6ebd3133
--- /dev/null
+++ b/wmake/src/wmkdependParser.atg
@@ -0,0 +1,497 @@
+/*---------------------------------------------------------------------------*\
+ Attributed Grammar for Coco/R (-*- C++ -*- version)
+ compile with:
+ coco-cpp wmkdependParser.atg
+\*---------------------------------------------------------------------------*/
+[copy]
+/*---------------------------------*- C++ -*---------------------------------*\
+ ========= |
+ \\ / F ield | OpenFOAM: The Open Source CFD Toolbox
+ \\ / O peration |
+ \\ / A nd | Copyright (C) 2010-2010 OpenCFD Ltd.
+ \\/ M anipulation |
+-------------------------------------------------------------------------------
+License
+ This file is part of OpenFOAM.
+
+ OpenFOAM is free software; you can redistribute it and/or modify it
+ under the terms of the GNU General Public License as published by the
+ Free Software Foundation; either version 2 of the License, or (at your
+ option) any later version.
+
+ OpenFOAM is distributed in the hope that it will be useful, but WITHOUT
+ ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License
+ for more details.
+
+ You should have received a copy of the GNU General Public License
+ along with OpenFOAM; if not, write to the Free Software Foundation,
+ Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA
+
+@file wmkdependParser.atg
+
+Description
+ An attributed Coco/R grammar to parse C/C++, Fortran and Java files
+ for include and import statements.
+
+SourceFiles
+ generated
+
+\*---------------------------------------------------------------------------*/
+[/copy]
+#include <iostream>
+#include <string>
+#include <list>
+
+//! @brief A simple HashTable implementation
+/**
+ * @note This hash table is only vaguely STL-like. In accordance with
+ * its present purpose, this hash table only supports a constIterator
+ * and no deletions. For simplicity, the constIterator increment is
+ * simply via a next() method. Instead of comparing to an end value,
+ * the constIterator valid() method is used.
+ * For example,
+ * @code
+ * for
+ * (
+ * HashTable<foo>::constIterator iter = myHash.begin();
+ * iter.valid();
+ * iter.next()
+ * )
+ * {
+ * std::cerr<< "key: " << iter.key() << "\n";
+ * }
+ * @endcode
+ *
+ */
+class StringHashSet
+{
+ //! An entry within the HashTable
+ struct hashedEntry
+ {
+ const std::string key_; //<! The lookup key
+ hashedEntry *next_; //<! Pointer to next hashedEntry in sub-list
+
+ hashedEntry(const std::string& key, hashedEntry *next=0)
+ :
+ key_(key), next_(next)
+ {}
+ };
+
+ const int size_; //<! fixed HashTable size
+ hashedEntry** table_;
+
+public:
+
+ //! Construct with a default size
+ StringHashSet(int size = 500)
+ :
+ size_(size),
+ table_(new hashedEntry*[size_])
+ {
+ memset(table_, 0, size_ * sizeof(hashedEntry*));
+ }
+
+ //! Destructor
+ ~StringHashSet()
+ {
+ for (int hashIdx = 0; hashIdx < size_; ++hashIdx)
+ {
+ hashedEntry* ep = table_[hashIdx];
+ while (ep)
+ {
+ hashedEntry* del = ep;
+ ep = ep->next_;
+ delete del;
+ }
+ }
+ delete[] table_;
+ table_ = 0;
+ }
+
+ //! Return hash index for lookup name in hash table
+ bool hashKeyIndex(const std::string& name) const
+ {
+ int hashIdx = 0;
+
+ // calculate hash index
+ for
+ (
+ std::string::const_iterator iter = name.begin();
+ iter != name.end();
+ ++iter
+ )
+ {
+ hashIdx = hashIdx << 1 ^ *iter;
+ }
+
+ if (hashIdx < 0)
+ {
+ hashIdx = -hashIdx;
+ }
+
+ return hashIdx % size_;
+ }
+
+
+ //! Return true if name is found in hash table
+ bool found(const std::string& name) const
+ {
+ const int hashIdx = hashKeyIndex(name);
+
+ for (hashedEntry* ep = table_[hashIdx]; ep; ep = ep->next_)
+ {
+ if (name == ep->key_)
+ {
+ // found
+ return true;
+ }
+ }
+
+ // entry not found
+ return false;
+ }
+
+
+ //! Return true if name is found in hash table, insert if not found
+ bool foundOrInsert(const std::string& name)
+ {
+ const int hashIdx = hashKeyIndex(name);
+
+ for (hashedEntry* ep = table_[hashIdx]; ep; ep = ep->next_)
+ {
+ if (name == ep->key_)
+ {
+ // found - return true
+ return true;
+ }
+ }
+
+ // not found - insert it
+ table_[hashIdx] = new hashedEntry(name, table_[hashIdx]);
+
+ // entry not found (but was added) - return false
+ return false;
+ }
+
+};
+
+
+/*---------------------------------------------------------------------------*/
+
+COMPILER wmkdepend
+ // grammar pragmas:
+ $namespace=wmake
+ $prefix=wmkdepend
+ $define=FORCE_UTF8
+
+/*---------------------------------------------------------------------------*/
+private:
+
+ //! Hash of files already visited
+ static StringHashSet visitedFiles_;
+
+ //! Hash of (java) directories already visited
+ static StringHashSet visitedDirs_;
+
+ //! Replace all '.' with '/'
+ static void dotToSlash(std::string& name);
+
+ //! Import (java) directories
+ static void importDir(const std::string& dirName);
+
+ //! Import (java) file
+ static void importFile(const std::string& name);
+
+public:
+ //! Include directories to search
+ static std::list<std::string> includeDirs;
+
+ //! The name of the top-level source file
+ static std::string sourceFile;
+
+ //! The name of the top-level dep file
+ static std::string depFile;
+
+ //! Add directory to list of visited dirs, thus effectively ignoring it
+ static void ignoreDir(const std::string& name);
+
+ //! Include file
+ static void includeFile(const std::string& name);
+
+/*---------------------------------------------------------------------------*/
+[code]
+#include <sys/types.h>
+#include <dirent.h>
+
+StringHashSet Parser::visitedFiles_;
+StringHashSet Parser::visitedDirs_;
+
+std::list<std::string> Parser::includeDirs;
+std::string Parser::sourceFile;
+std::string Parser::depFile;
+
+
+void Parser::dotToSlash(std::string& name)
+{
+ std::string::size_type start = 0;
+
+ while ((start = name.find('.', start)) != std::string::npos)
+ {
+ name.replace(start, 1, 1, '/');
+ start++;
+ }
+}
+
+
+void Parser::ignoreDir(const std::string& name)
+{
+ visitedDirs_.foundOrInsert(name);
+}
+
+
+void Parser::includeFile(const std::string& name)
+{
+ if (visitedFiles_.foundOrInsert(name))
+ {
+ return;
+ }
+
+ // use stdio and buffering within Coco/R -- (faster)
+ FILE *fh = fopen(name.c_str(), "r");
+ if (fh)
+ {
+ std::cout << depFile << ": " << name << "\n";
+ }
+ else
+ {
+ for
+ (
+ std::list<std::string>::const_iterator iter = includeDirs.begin();
+ iter != includeDirs.end();
+ ++iter
+ )
+ {
+ const std::string pathName = *iter + name;
+
+ fh = fopen(pathName.c_str(), "r");
+ if (fh)
+ {
+ std::cout << depFile << ": " << pathName << "\n";
+ break;
+ }
+ }
+ }
+
+ if (fh)
+ {
+ Scanner scanner(fh);
+ Parser parser(&scanner);
+
+ parser.Parse();
+ fclose(fh);
+ }
+ else
+ {
+ fwprintf
+ (
+ stderr,
+ L"could not open file %s for source file %s\n",
+ name.c_str(), sourceFile.c_str()
+ );
+ }
+}
+
+
+void Parser::importFile(const std::string& name)
+{
+ // check if a globbed form was already visited
+ std::string::size_type dotPos = name.find('.');
+ if (dotPos != std::string::npos)
+ {
+ std::string dirGlob = name.substr(0, dotPos);
+ dirGlob += ".*";
+
+ if (visitedDirs_.found(dirGlob))
+ {
+ return;
+ }
+ }
+
+ std::string javaFileName = name;
+
+ dotToSlash(javaFileName);
+ javaFileName += ".java";
+
+ includeFile(javaFileName);
+}
+
+
+void Parser::importDir(const std::string& name)
+{
+ if (visitedDirs_.foundOrInsert(name))
+ {
+ return;
+ }
+
+ std::string dirName = name;
+ dotToSlash(dirName);
+
+ DIR *source = opendir(dirName.c_str());
+
+ if (source)
+ {
+ struct dirent *list;
+
+ // Read and parse all the entries in the directory
+ while ((list = readdir(source)) != NULL)
+ {
+ const char* ext = strstr(list->d_name, ".java");
+
+ // avoid matching on something like '.java~'
+ if (ext && strlen(ext) == 5)
+ {
+ std::string pathName = dirName + list->d_name;
+ includeFile(pathName);
+ }
+ }
+
+ closedir(source);
+ }
+ else
+ {
+ fwprintf
+ (
+ stderr,
+ L"could not open directory %s\n",
+ dirName.c_str()
+ );
+ return;
+ }
+}
+
+[/code]
+
+
+/*---------------------------------------------------------------------------*/
+
+CHARACTERS
+ letter = 'A'..'Z' + 'a'..'z' + '_'.
+ digit = "0123456789".
+ cr = '\r'.
+ lf = '\n'.
+ tab = '\t'.
+ stringCh = ANY - '"' - '\\' - cr - lf.
+ printable = '\u0020' .. '\u007e'.
+ java_letter = letter + '$'.
+
+// * * * * * * * * * * * * * * * * TOKENS * * * * * * * * * * * * * * * * * //
+
+TOKENS
+
+// string
+string =
+ '"' { stringCh | '\\' printable } '"'.
+
+// single-quoted string (eg, Fortran)
+sqstring =
+ '\'' { stringCh | '\\' printable } '\''.
+
+// for java import
+package_name =
+ java_letter { java_letter | digit }
+ { '.' java_letter { java_letter | digit } } .
+
+// for java import
+package_dir =
+ java_letter { java_letter | digit }
+ { '.' java_letter { java_letter | digit } } ".*" .
+
+
+// * * * * * * * * * * * PRAGMAS / COMMENTS / IGNORE * * * * * * * * * * * //
+
+COMMENTS FROM "/*" TO "*/" NESTED
+COMMENTS FROM "//" TO lf
+
+IGNORE tab
+
+// * * * * * * * * * * * * * * * PRODUCTIONS * * * * * * * * * * * * * * * //
+
+PRODUCTIONS
+
+wmkdepend
+=
+{
+ // C/C++-style includes
+ '#'
+ [
+ "include"
+ [
+ string (.
+ if (isUTF8())
+ {
+ includeFile(t->toStringUTF8(1, t->length()-2));
+ }
+ else
+ {
+ includeFile(t->toString(1, t->length()-2));
+ }
+ .)
+ ]
+ ]
+ [ ANY { ANY } ] '\n' // skip trailing junk
+
+ // Fortran-style includes
+ | "include"
+ [
+ sqstring (.
+ if (isUTF8())
+ {
+ includeFile(t->toStringUTF8(1, t->length()-2));
+ }
+ else
+ {
+ includeFile(t->toString(1, t->length()-2));
+ }
+ .)
+ ]
+ [ ANY { ANY } ] '\n' // skip trailing junk
+
+ // Java imports
+ | "import"
+ (
+ package_dir (.
+ if (isUTF8())
+ {
+ importDir(t->toStringUTF8());
+ }
+ else
+ {
+ importDir(t->toString());
+ }
+ .)
+ | package_name (.
+ if (isUTF8())
+ {
+ importFile(t->toStringUTF8());
+ }
+ else
+ {
+ importFile(t->toString());
+ }
+ .)
+ )
+ ';'
+ [ ANY { ANY } ] '\n' // skip trailing junk
+
+ | [ ANY { ANY } ] '\n' // skip any other lines
+
+}
+.
+
+
+/*---------------------------------------------------------------------------*/
+
+END wmkdepend.
+
+// ************************************************************************* //
diff --git a/wmake/src/wmkdependParser.cpp b/wmake/src/wmkdependParser.cpp
new file mode 100644
index 00000000000..b6d5e4e71a1
--- /dev/null
+++ b/wmake/src/wmkdependParser.cpp
@@ -0,0 +1,575 @@
+/*---------------------------------*- C++ -*---------------------------------*\
+ ========= |
+ \\ / F ield | OpenFOAM: The Open Source CFD Toolbox
+ \\ / O peration |
+ \\ / A nd | Copyright (C) 2010-2010 OpenCFD Ltd.
+ \\/ M anipulation |
+-------------------------------------------------------------------------------
+License
+ This file is part of OpenFOAM.
+
+ OpenFOAM is free software; you can redistribute it and/or modify it
+ under the terms of the GNU General Public License as published by the
+ Free Software Foundation; either version 2 of the License, or (at your
+ option) any later version.
+
+ OpenFOAM is distributed in the hope that it will be useful, but WITHOUT
+ ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License
+ for more details.
+
+ You should have received a copy of the GNU General Public License
+ along with OpenFOAM; if not, write to the Free Software Foundation,
+ Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA
+
+@file wmkdependParser.atg
+
+Description
+ An attributed Coco/R grammar to parse C/C++, Fortran and Java files
+ for include and import statements.
+
+SourceFiles
+ generated
+
+\*---------------------------------------------------------------------------*/
+// This file was generated with Coco/R C++ (7 Feb 2010)
+// http://www.ssw.uni-linz.ac.at/coco/
+// with these defines:
+// - FORCE_UTF8
+
+
+#include <cstdio>
+#include <cstdlib>
+#include <cstring>
+#include <cwchar>
+#include <sstream>
+
+#include "wmkdependParser.h"
+
+namespace wmake {
+
+
+#include <sys/types.h>
+#include <dirent.h>
+
+StringHashSet Parser::visitedFiles_;
+StringHashSet Parser::visitedDirs_;
+
+std::list<std::string> Parser::includeDirs;
+std::string Parser::sourceFile;
+std::string Parser::depFile;
+
+
+void Parser::dotToSlash(std::string& name)
+{
+ std::string::size_type start = 0;
+
+ while ((start = name.find('.', start)) != std::string::npos)
+ {
+ name.replace(start, 1, 1, '/');
+ start++;
+ }
+}
+
+
+void Parser::ignoreDir(const std::string& name)
+{
+ visitedDirs_.foundOrInsert(name);
+}
+
+
+void Parser::includeFile(const std::string& name)
+{
+ if (visitedFiles_.foundOrInsert(name))
+ {
+ return;
+ }
+
+ // use stdio and buffering within Coco/R -- (faster)
+ FILE *fh = fopen(name.c_str(), "r");
+ if (fh)
+ {
+ std::cout << depFile << ": " << name << "\n";
+ }
+ else
+ {
+ for
+ (
+ std::list<std::string>::const_iterator iter = includeDirs.begin();
+ iter != includeDirs.end();
+ ++iter
+ )
+ {
+ const std::string pathName = *iter + name;
+
+ fh = fopen(pathName.c_str(), "r");
+ if (fh)
+ {
+ std::cout << depFile << ": " << pathName << "\n";
+ break;
+ }
+ }
+ }
+
+ if (fh)
+ {
+ Scanner scanner(fh);
+ Parser parser(&scanner);
+
+ parser.Parse();
+ fclose(fh);
+ }
+ else
+ {
+ fwprintf
+ (
+ stderr,
+ L"could not open file %s for source file %s\n",
+ name.c_str(), sourceFile.c_str()
+ );
+ }
+}
+
+
+void Parser::importFile(const std::string& name)
+{
+ // check if a globbed form was already visited
+ std::string::size_type dotPos = name.find('.');
+ if (dotPos != std::string::npos)
+ {
+ std::string dirGlob = name.substr(0, dotPos);
+ dirGlob += ".*";
+
+ if (visitedDirs_.found(dirGlob))
+ {
+ return;
+ }
+ }
+
+ std::string javaFileName = name;
+
+ dotToSlash(javaFileName);
+ javaFileName += ".java";
+
+ includeFile(javaFileName);
+}
+
+
+void Parser::importDir(const std::string& name)
+{
+ if (visitedDirs_.foundOrInsert(name))
+ {
+ return;
+ }
+
+ std::string dirName = name;
+ dotToSlash(dirName);
+
+ DIR *source = opendir(dirName.c_str());
+
+ if (source)
+ {
+ struct dirent *list;
+
+ // Read and parse all the entries in the directory
+ while ((list = readdir(source)) != NULL)
+ {
+ const char* ext = strstr(list->d_name, ".java");
+
+ // avoid matching on something like '.java~'
+ if (ext && strlen(ext) == 5)
+ {
+ std::string pathName = dirName + list->d_name;
+ includeFile(pathName);
+ }
+ }
+
+ closedir(source);
+ }
+ else
+ {
+ fwprintf
+ (
+ stderr,
+ L"could not open directory %s\n",
+ dirName.c_str()
+ );
+ return;
+ }
+}
+
+
+
+
+//! @cond fileScope
+//
+
+//
+// Create by copying str - only used locally
+inline static wchar_t* coco_string_create(const wchar_t* str)
+{
+ const int len = wcslen(str);
+ wchar_t* dst = new wchar_t[len + 1];
+ wcsncpy(dst, str, len);
+ dst[len] = 0;
+ return dst;
+}
+
+
+// Free storage and nullify the argument
+inline static void coco_string_delete(wchar_t* &str)
+{
+ delete[] str;
+ str = NULL;
+}
+//
+//! @endcond fileScope
+
+
+// ----------------------------------------------------------------------------
+// Parser Implementation
+// ----------------------------------------------------------------------------
+
+void Parser::SynErr(int n)
+{
+ if (errDist >= minErrDist) errors->SynErr(la->line, la->col, n);
+ errDist = 0;
+}
+
+
+void Parser::SemErr(const std::wstring& msg)
+{
+ if (errDist >= minErrDist) errors->Error(t->line, t->col, msg);
+ errDist = 0;
+}
+
+
+bool Parser::isUTF8() const
+{
+ return scanner && scanner->buffer && scanner->buffer->isUTF8();
+}
+
+
+void Parser::Get()
+{
+ for (;;)
+ {
+ t = la;
+ la = scanner->Scan();
+ if (la->kind <= maxT)
+ {
+ ++errDist;
+ break;
+ }
+ if (dummyToken != t)
+ {
+ dummyToken->kind = t->kind;
+ dummyToken->pos = t->pos;
+ dummyToken->col = t->col;
+ dummyToken->line = t->line;
+ dummyToken->next = NULL;
+ coco_string_delete(dummyToken->val);
+ dummyToken->val = coco_string_create(t->val);
+ t = dummyToken;
+ }
+ la = t;
+ }
+}
+
+
+void Parser::Expect(int n)
+{
+ if (la->kind == n)
+ {
+ Get();
+ }
+ else
+ {
+ SynErr(n);
+ }
+}
+
+
+void Parser::ExpectWeak(int n, int follow)
+{
+ if (la->kind == n)
+ {
+ Get();
+ }
+ else
+ {
+ SynErr(n);
+ while (!StartOf(follow))
+ {
+ Get();
+ }
+ }
+}
+
+
+bool Parser::WeakSeparator(int n, int syFol, int repFol)
+{
+ if (la->kind == n)
+ {
+ Get();
+ return true;
+ }
+ else if (StartOf(repFol))
+ {
+ return false;
+ }
+ else
+ {
+ SynErr(n);
+ while (!(StartOf(syFol) || StartOf(repFol) || StartOf(0)))
+ {
+ Get();
+ }
+ return StartOf(syFol);
+ }
+}
+
+
+void Parser::wmkdepend()
+{
+ while (StartOf(1)) {
+ if (la->kind == 5) {
+ Get();
+ if (la->kind == 6) {
+ Get();
+ if (la->kind == 1) {
+ Get();
+ if (isUTF8())
+ {
+ includeFile(t->toStringUTF8(1, t->length()-2));
+ }
+ else
+ {
+ includeFile(t->toString(1, t->length()-2));
+ }
+
+ }
+ }
+ if (StartOf(2)) {
+ Get();
+ while (StartOf(3)) {
+ Get();
+ }
+ }
+ Expect(7);
+ } else if (la->kind == 6) {
+ Get();
+ if (la->kind == 2) {
+ Get();
+ if (isUTF8())
+ {
+ includeFile(t->toStringUTF8(1, t->length()-2));
+ }
+ else
+ {
+ includeFile(t->toString(1, t->length()-2));
+ }
+
+ }
+ if (StartOf(4)) {
+ Get();
+ while (StartOf(3)) {
+ Get();
+ }
+ }
+ Expect(7);
+ } else if (la->kind == 8) {
+ Get();
+ if (la->kind == 4) {
+ Get();
+ if (isUTF8())
+ {
+ importDir(t->toStringUTF8());
+ }
+ else
+ {
+ importDir(t->toString());
+ }
+
+ } else if (la->kind == 3) {
+ Get();
+ if (isUTF8())
+ {
+ importFile(t->toStringUTF8());
+ }
+ else
+ {
+ importFile(t->toString());
+ }
+
+ } else SynErr(11);
+ Expect(9);
+ if (StartOf(3)) {
+ Get();
+ while (StartOf(3)) {
+ Get();
+ }
+ }
+ Expect(7);
+ } else {
+ if (StartOf(5)) {
+ Get();
+ while (StartOf(3)) {
+ Get();
+ }
+ }
+ Expect(7);
+ }
+ }
+}
+
+
+
+void Parser::Parse()
+{
+ t = NULL;
+ // might call Parse() twice
+ if (dummyToken) {
+ coco_string_delete(dummyToken->val);
+ delete dummyToken;
+ }
+ dummyToken = new Token(coco_string_create(L"Dummy Token"));
+ la = dummyToken;
+ Get();
+ wmkdepend();
+ Expect(0); // expect end-of-file automatically added
+}
+
+
+Parser::Parser(Scanner* scan, Errors* err)
+:
+ dummyToken(NULL),
+ deleteErrorsDestruct_(!err),
+ errDist(minErrDist),
+ scanner(scan),
+ errors(err),
+ t(NULL),
+ la(NULL)
+{
+ if (!errors) // add in default error handling
+ {
+ errors = new Errors();
+ }
+ // user-defined initializations:
+}
+
+
+bool Parser::StartOf(int s)
+{
+ const bool T = true;
+ const bool x = false;
+
+ static const bool set[6][12] =
+ {
+ {T,x,x,x, x,x,x,x, x,x,x,x},
+ {x,T,T,T, T,T,T,T, T,T,T,x},
+ {x,x,T,T, T,T,x,x, T,T,T,x},
+ {x,T,T,T, T,T,T,x, T,T,T,x},
+ {x,T,x,T, T,T,T,x, T,T,T,x},
+ {x,T,T,T, T,x,x,x, x,T,T,x}
+ };
+
+ return set[s][la->kind];
+}
+
+
+Parser::~Parser()
+{
+ if (deleteErrorsDestruct_) { delete errors; } // delete default error handling
+ if (dummyToken) {
+ coco_string_delete(dummyToken->val);
+ delete dummyToken;
+ }
+ // user-defined destruction:
+}
+
+
+// ----------------------------------------------------------------------------
+// Errors Implementation
+// ----------------------------------------------------------------------------
+
+Errors::Errors()
+:
+ count(0)
+{}
+
+
+Errors::~Errors()
+{}
+
+
+void Errors::clear()
+{
+ count = 0;
+}
+
+
+std::wstring Errors::strerror(int n)
+{
+ switch (n) {
+ case 0: return L"EOF expected"; break;
+ case 1: return L"string expected"; break;
+ case 2: return L"sqstring expected"; break;
+ case 3: return L"package_name expected"; break;
+ case 4: return L"package_dir expected"; break;
+ case 5: return L"\"#\" expected"; break;
+ case 6: return L"\"include\" expected"; break;
+ case 7: return L"\"\\n\" expected"; break;
+ case 8: return L"\"import\" expected"; break;
+ case 9: return L"\";\" expected"; break;
+ case 10: return L"??? expected"; break;
+ case 11: return L"invalid wmkdepend"; break;
+ default:
+ {
+ // std::wostringstream buf; (this typedef might be missing)
+ std::basic_ostringstream<wchar_t> buf;
+ buf << "error " << n;
+ return buf.str();
+ }
+ break;
+ }
+}
+
+
+void Errors::Warning(const std::wstring& msg)
+{
+ fwprintf(stderr, L"%ls\n", msg.c_str());
+}
+
+
+void Errors::Warning(int line, int col, const std::wstring& msg)
+{
+ fwprintf(stderr, L"-- line %d col %d: %ls\n", line, col, msg.c_str());
+}
+
+
+void Errors::Error(int line, int col, const std::wstring& msg)
+{
+ fwprintf(stderr, L"-- line %d col %d: %ls\n", line, col, msg.c_str());
+ count++;
+}
+
+
+void Errors::SynErr(int line, int col, int n)
+{
+ this->Error(line, col, this->strerror(n));
+}
+
+
+void Errors::Exception(const std::wstring& msg)
+{
+ fwprintf(stderr, L"%ls", msg.c_str());
+ ::exit(1);
+}
+
+
+// * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * //
+
+} // End namespace
+
+// ************************************************************************* //
diff --git a/wmake/src/wmkdependParser.h b/wmake/src/wmkdependParser.h
new file mode 100644
index 00000000000..8c7703adbd6
--- /dev/null
+++ b/wmake/src/wmkdependParser.h
@@ -0,0 +1,310 @@
+/*---------------------------------*- C++ -*---------------------------------*\
+ ========= |
+ \\ / F ield | OpenFOAM: The Open Source CFD Toolbox
+ \\ / O peration |
+ \\ / A nd | Copyright (C) 2010-2010 OpenCFD Ltd.
+ \\/ M anipulation |
+-------------------------------------------------------------------------------
+License
+ This file is part of OpenFOAM.
+
+ OpenFOAM is free software; you can redistribute it and/or modify it
+ under the terms of the GNU General Public License as published by the
+ Free Software Foundation; either version 2 of the License, or (at your
+ option) any later version.
+
+ OpenFOAM is distributed in the hope that it will be useful, but WITHOUT
+ ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License
+ for more details.
+
+ You should have received a copy of the GNU General Public License
+ along with OpenFOAM; if not, write to the Free Software Foundation,
+ Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA
+
+@file wmkdependParser.atg
+
+Description
+ An attributed Coco/R grammar to parse C/C++, Fortran and Java files
+ for include and import statements.
+
+SourceFiles
+ generated
+
+\*---------------------------------------------------------------------------*/
+// This file was generated with Coco/R C++ (7 Feb 2010)
+// http://www.ssw.uni-linz.ac.at/coco/
+// with these defines:
+// - FORCE_UTF8
+
+
+#ifndef COCO_wmkdependPARSER_H__
+#define COCO_wmkdependPARSER_H__
+
+#include <iostream>
+#include <string>
+#include <list>
+
+//! @brief A simple HashTable implementation
+/**
+ * @note This hash table is only vaguely STL-like. In accordance with
+ * its present purpose, this hash table only supports a constIterator
+ * and no deletions. For simplicity, the constIterator increment is
+ * simply via a next() method. Instead of comparing to an end value,
+ * the constIterator valid() method is used.
+ * For example,
+ * @code
+ * for
+ * (
+ * HashTable<foo>::constIterator iter = myHash.begin();
+ * iter.valid();
+ * iter.next()
+ * )
+ * {
+ * std::cerr<< "key: " << iter.key() << "\n";
+ * }
+ * @endcode
+ *
+ */
+class StringHashSet
+{
+ //! An entry within the HashTable
+ struct hashedEntry
+ {
+ const std::string key_; //<! The lookup key
+ hashedEntry *next_; //<! Pointer to next hashedEntry in sub-list
+
+ hashedEntry(const std::string& key, hashedEntry *next=0)
+ :
+ key_(key), next_(next)
+ {}
+ };
+
+ const int size_; //<! fixed HashTable size
+ hashedEntry** table_;
+
+public:
+
+ //! Construct with a default size
+ StringHashSet(int size = 500)
+ :
+ size_(size),
+ table_(new hashedEntry*[size_])
+ {
+ memset(table_, 0, size_ * sizeof(hashedEntry*));
+ }
+
+ //! Destructor
+ ~StringHashSet()
+ {
+ for (int hashIdx = 0; hashIdx < size_; ++hashIdx)
+ {
+ hashedEntry* ep = table_[hashIdx];
+ while (ep)
+ {
+ hashedEntry* del = ep;
+ ep = ep->next_;
+ delete del;
+ }
+ }
+ delete[] table_;
+ table_ = 0;
+ }
+
+ //! Return hash index for lookup name in hash table
+ bool hashKeyIndex(const std::string& name) const
+ {
+ int hashIdx = 0;
+
+ // calculate hash index
+ for
+ (
+ std::string::const_iterator iter = name.begin();
+ iter != name.end();
+ ++iter
+ )
+ {
+ hashIdx = hashIdx << 1 ^ *iter;
+ }
+
+ if (hashIdx < 0)
+ {
+ hashIdx = -hashIdx;
+ }
+
+ return hashIdx % size_;
+ }
+
+
+ //! Return true if name is found in hash table
+ bool found(const std::string& name) const
+ {
+ const int hashIdx = hashKeyIndex(name);
+
+ for (hashedEntry* ep = table_[hashIdx]; ep; ep = ep->next_)
+ {
+ if (name == ep->key_)
+ {
+ // found
+ return true;
+ }
+ }
+
+ // entry not found
+ return false;
+ }
+
+
+ //! Return true if name is found in hash table, insert if not found
+ bool foundOrInsert(const std::string& name)
+ {
+ const int hashIdx = hashKeyIndex(name);
+
+ for (hashedEntry* ep = table_[hashIdx]; ep; ep = ep->next_)
+ {
+ if (name == ep->key_)
+ {
+ // found - return true
+ return true;
+ }
+ }
+
+ // not found - insert it
+ table_[hashIdx] = new hashedEntry(name, table_[hashIdx]);
+
+ // entry not found (but was added) - return false
+ return false;
+ }
+
+};
+
+
+/*---------------------------------------------------------------------------*/
+
+
+
+#include "wmkdependScanner.h"
+
+namespace wmake {
+
+
+/*---------------------------------------------------------------------------*\
+ Class Errors Declaration
+\*---------------------------------------------------------------------------*/
+//! Parser error handing
+class Errors
+{
+public:
+ int count; //!< The number of errors detected
+
+ //! Return a string describing the given error code.
+ static std::wstring strerror(int n);
+
+ Errors(); //!< Construct null - start with no errors
+ virtual ~Errors(); //!< Destructor
+ virtual void clear(); //!< Clear the error count
+
+ //! Handle a general warning 'msg'
+ virtual void Warning(const std::wstring& msg);
+ //! Handle a general warning 'msg'
+ virtual void Warning(int line, int col, const std::wstring& msg);
+ //! Handle general error 'msg' (eg, a semantic error)
+ virtual void Error(int line, int col, const std::wstring& msg);
+ //! Handle syntax error 'n', uses strerror for the message, calls Error()
+ virtual void SynErr(int line, int col, int n);
+ //! Handle a general exception 'msg'
+ virtual void Exception(const std::wstring& msg);
+
+}; // Errors
+
+
+
+/*---------------------------------------------------------------------------*\
+ Class Parser Declaration
+\*---------------------------------------------------------------------------*/
+//! A Coco/R Parser
+class Parser
+{
+ enum {
+ _EOF=0,
+ _string=1,
+ _sqstring=2,
+ _package_name=3,
+ _package_dir=4,
+ maxT = 10 //<! max term (w/o pragmas)
+ };
+ static const int minErrDist = 2; //!< min. distance before reporting errors
+
+ Token *dummyToken;
+ bool deleteErrorsDestruct_; //!< delete the 'errors' member in destructor
+ int errDist;
+
+ void SynErr(int n); //!< Handle syntax error 'n'
+ void Get();
+ void Expect(int n);
+ bool StartOf(int s);
+ void ExpectWeak(int n, int follow);
+ bool WeakSeparator(int n, int syFol, int repFol);
+
+public:
+ Scanner *scanner;
+ Errors *errors;
+
+ Token *t; //!< last recognized token
+ Token *la; //!< lookahead token
+
+private:
+
+ //! Hash of files already visited
+ static StringHashSet visitedFiles_;
+
+ //! Hash of (java) directories already visited
+ static StringHashSet visitedDirs_;
+
+ //! Replace all '.' with '/'
+ static void dotToSlash(std::string& name);
+
+ //! Import (java) directories
+ static void importDir(const std::string& dirName);
+
+ //! Import (java) file
+ static void importFile(const std::string& name);
+
+public:
+ //! Include directories to search
+ static std::list<std::string> includeDirs;
+
+ //! The name of the top-level source file
+ static std::string sourceFile;
+
+ //! The name of the top-level dep file
+ static std::string depFile;
+
+ //! Add directory to list of visited dirs, thus effectively ignoring it
+ static void ignoreDir(const std::string& name);
+
+ //! Include file
+ static void includeFile(const std::string& name);
+
+/*---------------------------------------------------------------------------*/
+
+ //! Construct for the specified scanner
+ /*!
+ * Use the default error handling, or optionally provide an error
+ * handler, which will not be deleted upon destruction.
+ */
+ Parser(Scanner* scan, Errors* err = 0);
+ ~Parser();
+ void Parse(); //!< Execute the parse operation
+ void SemErr(const std::wstring& msg); //!< Handle semantic error
+ bool isUTF8() const; //!< Return true if scanner buffer is UTF8
+
+ void wmkdepend();
+
+}; // end Parser
+
+} // End namespace
+
+#endif // COCO_wmkdependPARSER_H__
+
+// ************************************************************************* //
diff --git a/wmake/src/wmkdependScanner.cpp b/wmake/src/wmkdependScanner.cpp
new file mode 100644
index 00000000000..72f082f7886
--- /dev/null
+++ b/wmake/src/wmkdependScanner.cpp
@@ -0,0 +1,1022 @@
+/*---------------------------------*- C++ -*---------------------------------*\
+ ========= |
+ \\ / F ield | OpenFOAM: The Open Source CFD Toolbox
+ \\ / O peration |
+ \\ / A nd | Copyright (C) 2010-2010 OpenCFD Ltd.
+ \\/ M anipulation |
+-------------------------------------------------------------------------------
+License
+ This file is part of OpenFOAM.
+
+ OpenFOAM is free software; you can redistribute it and/or modify it
+ under the terms of the GNU General Public License as published by the
+ Free Software Foundation; either version 2 of the License, or (at your
+ option) any later version.
+
+ OpenFOAM is distributed in the hope that it will be useful, but WITHOUT
+ ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License
+ for more details.
+
+ You should have received a copy of the GNU General Public License
+ along with OpenFOAM; if not, write to the Free Software Foundation,
+ Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA
+
+@file wmkdependParser.atg
+
+Description
+ An attributed Coco/R grammar to parse C/C++, Fortran and Java files
+ for include and import statements.
+
+SourceFiles
+ generated
+
+\*---------------------------------------------------------------------------*/
+// This file was generated with Coco/R C++ (7 Feb 2010)
+// http://www.ssw.uni-linz.ac.at/coco/
+// with these defines:
+// - FORCE_UTF8
+
+
+#include <sstream>
+
+// io.h and fcntl are used to ensure binary read from streams on windows
+#ifdef _WIN32
+# include <io.h>
+# include <fcntl.h>
+#endif
+
+#include <climits>
+
+#include "wmkdependScanner.h"
+
+// values for the file stream buffering
+#define MIN_BUFFER_LENGTH 1024 // 1KB
+#define MAX_BUFFER_LENGTH (64*MIN_BUFFER_LENGTH) // 64KB
+// value for the heap management
+#define HEAP_BLOCK_SIZE (64*1024) // 64KB
+
+
+namespace wmake {
+
+// * * * * * * * * * * * Miscellaneous String Routines * * * * * * * * * * * //
+
+//
+// string handling, byte character
+//
+
+std::string coco_stdString(const wchar_t* str)
+{
+ return str ? coco_stdString(str, 0, wcslen(str)) : std::string();
+}
+
+
+std::string coco_stdString(const wchar_t* str, unsigned length)
+{
+ return coco_stdString(str, 0, length);
+}
+
+
+std::string coco_stdString(const wchar_t* str, unsigned index, unsigned length)
+{
+ const unsigned len = (str && *str) ? length : 0;
+ std::string dst;
+ dst.reserve(len);
+
+ for (unsigned i = 0; i < len; ++i)
+ {
+ dst += char(str[index+i] & 0xFF);
+ }
+
+ return dst;
+}
+
+
+std::string coco_stdStringUTF8(const wchar_t* str)
+{
+ return str ? coco_stdStringUTF8(str, 0, wcslen(str)) : std::string();
+}
+
+
+std::string coco_stdStringUTF8(const wchar_t* str, unsigned length)
+{
+ return coco_stdStringUTF8(str, 0, length);
+}
+
+
+std::string coco_stdStringUTF8(const wchar_t* str, unsigned index, unsigned length)
+{
+ const unsigned len = (str && *str) ? length : 0;
+ std::string dst;
+ dst.reserve(len);
+
+ for (unsigned i = 0; i < len; ++i)
+ {
+ wchar_t wc = str[index+i];
+
+ if (!(wc & ~0x0000007F))
+ {
+ // 0x00000000 - 0x0000007F [min. 8bit storage, 1-byte encoding)
+ // 0aaaaaaa
+ dst += char(wc);
+ }
+ else if (!(wc & ~0x000007FF))
+ {
+ // 0x00000080 - 0x000007FF [min. 16bit storage, 2-byte encoding]
+ // 110bbbaa 10aaaaaa
+ dst += char(0xC0 | ((wc >> 6) & 0x1F));
+ dst += char(0x80 | ((wc) & 0x3F));
+ }
+ else if (!(wc & ~0x0000FFFF))
+ {
+ // 0x00000800 - 0x0000FFFF [min. 16bit storage, 3-byte encoding]
+ // 1110bbbb 10bbbbaa 10aaaaaa
+ dst += char(0xE0 | ((wc >> 12) & 0x0F));
+ dst += char(0x80 | ((wc >> 6) & 0x3F));
+ dst += char(0x80 | ((wc) & 0x3F));
+ }
+ else if (!(wc & ~0x001FFFFF))
+ {
+ // 0x00010000 - 0x001FFFFF [min. 24bit storage, 4-byte encoding]
+ // 11110ccc 10ccbbbb 10bbbbaa 10aaaaaa
+ dst += char(0xF0 | ((wc >> 18) & 0x07));
+ dst += char(0x80 | ((wc >> 12) & 0x3F));
+ dst += char(0x80 | ((wc >> 6) & 0x3F));
+ dst += char(0x80 | ((wc) & 0x3F));
+ }
+//
+// Not (yet) used - wchar_t storage is limited to 16bit on windows
+// This also corresponds to the unicode BMP (Basic Multilingual Plane)
+//
+// else if (!(wc & ~0x03FFFFFF))
+// {
+// // 0x00200000 - 0x03FFFFFF [min. 32bit storage, 5-byte encoding]
+// // 111110dd 10cccccc 10ccbbbb 10bbbbaa 10aaaaaa
+// dst += char(0xF8 | ((wc >> 24) & 0x03));
+// dst += char(0x80 | ((wc >> 18) & 0x3F));
+// dst += char(0x80 | ((wc >> 12) & 0x3F));
+// dst += char(0x80 | ((wc >> 6) & 0x3F));
+// dst += char(0x80 | ((wc) & 0x3F));
+// }
+// else if (!(wc & ~0x7FFFFFFF))
+// {
+// // 0x04000000 - 0x7FFFFFFF [min. 32bit storage, 6-byte encoding]
+// // 1111110d 10dddddd 10cccccc 10ccbbbb 10bbbbaa 10aaaaaa
+// dst += char(0xFC | ((wc >> 30) & 0x01));
+// dst += char(0x80 | ((wc >> 24) & 0x3F));
+// dst += char(0x80 | ((wc >> 18) & 0x3F));
+// dst += char(0x80 | ((wc >> 12) & 0x3F));
+// dst += char(0x80 | ((wc >> 6) & 0x3F));
+// dst += char(0x80 | ((wc) & 0x3F));
+// }
+//
+ else
+ {
+ // report anything unknown/invalid as replacement character U+FFFD
+ dst += char(0xEF);
+ dst += char(0xBF);
+ dst += char(0xBD);
+ }
+ }
+
+ return dst;
+}
+
+
+// * * * * * * * * * * * * End of String Routines * * * * * * * * * * * * * //
+
+
+Token::Token(wchar_t* value)
+:
+ kind(0),
+ pos(0),
+ col(0),
+ line(0),
+ val(value),
+ next(NULL)
+{}
+
+
+Token::~Token()
+{}
+
+
+int Token::length() const
+{
+ return val ? wcslen(val) : 0;
+}
+
+
+// ----------------------------------------------------------------------------
+// Buffer Implementation
+// ----------------------------------------------------------------------------
+
+Buffer::Buffer(Buffer* b)
+:
+ buf(b->buf),
+ bufCapacity(b->bufCapacity),
+ bufLen(b->bufLen),
+ bufPos(b->bufPos),
+ bufStart(b->bufStart),
+ fileLen(b->fileLen),
+ cStream(b->cStream),
+ stdStream(b->stdStream),
+ isUserStream_(b->isUserStream_)
+{
+ // avoid accidental deletion on any of these members
+ b->buf = NULL;
+ b->cStream = NULL;
+ b->stdStream = NULL;
+}
+
+
+Buffer::Buffer(const char* chars, int len)
+:
+ buf(new unsigned char[len]),
+ bufCapacity(len),
+ bufLen(len),
+ bufPos(0),
+ bufStart(0),
+ fileLen(len),
+ cStream(NULL),
+ stdStream(NULL),
+ isUserStream_(false)
+{
+ memcpy(this->buf, chars, len*sizeof(char));
+}
+
+
+Buffer::Buffer(const unsigned char* chars, int len)
+:
+ buf(new unsigned char[len]),
+ bufCapacity(len),
+ bufLen(len),
+ bufPos(0),
+ bufStart(0),
+ fileLen(len),
+ cStream(NULL),
+ stdStream(NULL),
+ isUserStream_(false)
+{
+ memcpy(this->buf, chars, len*sizeof(char));
+}
+
+
+Buffer::Buffer(FILE* ifh, bool isUserStream)
+:
+ buf(NULL),
+ bufCapacity(0),
+ bufLen(0),
+ bufPos(0),
+ bufStart(0),
+ fileLen(0),
+ cStream(ifh),
+ stdStream(NULL),
+ isUserStream_(isUserStream)
+{
+// ensure binary read on windows
+#ifdef _WIN32
+ _setmode(_fileno(cStream), _O_BINARY);
+#endif
+
+ if (CanSeek())
+ {
+ fseek(cStream, 0, SEEK_END);
+ fileLen = ftell(cStream);
+ fseek(cStream, 0, SEEK_SET);
+ bufLen = (fileLen < MAX_BUFFER_LENGTH) ? fileLen : MAX_BUFFER_LENGTH;
+ bufStart = INT_MAX; // nothing in the buffer so far
+ }
+
+ bufCapacity = (bufLen > 0) ? bufLen : MIN_BUFFER_LENGTH;
+ buf = new unsigned char[bufCapacity];
+ if (fileLen > 0) SetPos(0); // setup buffer to position 0 (start)
+ else bufPos = 0; // index 0 is already after the file, thus Pos = 0 is invalid
+ if (bufLen == fileLen && CanSeek()) Close();
+}
+
+
+Buffer::Buffer(std::istream* istr, bool isUserStream)
+:
+ buf(NULL),
+ bufCapacity(0),
+ bufLen(0),
+ bufPos(0),
+ bufStart(0),
+ fileLen(0),
+ cStream(NULL),
+ stdStream(istr),
+ isUserStream_(isUserStream)
+{
+#if _WIN32
+ // TODO: ensure binary read on windows?
+#endif
+}
+
+
+Buffer::~Buffer()
+{
+ Close();
+ if (buf)
+ {
+ delete[] buf;
+ buf = NULL;
+ }
+}
+
+
+void Buffer::Close()
+{
+ if (!isUserStream_)
+ {
+ if (cStream)
+ {
+ fclose(cStream);
+ cStream = NULL;
+ }
+ if (stdStream)
+ {
+ delete stdStream;
+ stdStream = 0;
+ }
+ }
+}
+
+
+int Buffer::Read()
+{
+ if (stdStream)
+ {
+ int ch = stdStream->get();
+ if (stdStream->eof())
+ {
+ return EoF;
+ }
+ return ch;
+ }
+
+ if (bufPos < bufLen) {
+ return buf[bufPos++];
+ }
+ else if (GetPos() < fileLen) {
+ SetPos(GetPos()); // shift buffer start to Pos
+ return buf[bufPos++];
+ }
+ else if (cStream && !CanSeek() && (ReadNextStreamChunk() > 0)) {
+ return buf[bufPos++];
+ }
+
+ return EoF;
+}
+
+bool Buffer::isUTF8() const
+{
+ return false;
+}
+
+int UTF8Buffer::Read()
+{
+ int ch;
+ do {
+ ch = Buffer::Read();
+ // until we find a utf8 start (0xxxxxxx or 11xxxxxx)
+ } while (ch != EoF && ch >= 128 && ((ch & 0xC0) != 0xC0));
+ if (ch < 128 || ch == EoF) {
+ // nothing to do, first 127 chars are identical in ASCII and UTF8
+ // 0xxxxxxx or end of file character
+ }
+ else if ((ch & 0xF0) == 0xF0) {
+ // 0x00010000 - 0x001FFFFF [min. 24bit storage, 4-byte encoding]
+ // 11110ccc 10ccbbbb 10bbbbaa 10aaaaaa
+ // CAUTION: this should probably be disallowed since it overflows
+ // wchar_t on windows and overflows the max (0xFFFF) used here
+ int c1 = ch & 0x07; ch = Buffer::Read();
+ int c2 = ch & 0x3F; ch = Buffer::Read();
+ int c3 = ch & 0x3F; ch = Buffer::Read();
+ int c4 = ch & 0x3F;
+ ch = (((((c1 << 6) | c2) << 6) | c3) << 6) | c4;
+ }
+ else if ((ch & 0xE0) == 0xE0) {
+ // 0x00000800 - 0x0000FFFF [min. 16bit storage, 3-byte encoding]
+ // 1110bbbb 10bbbbaa 10aaaaaa
+ int c1 = ch & 0x0F; ch = Buffer::Read();
+ int c2 = ch & 0x3F; ch = Buffer::Read();
+ int c3 = ch & 0x3F;
+ ch = (((c1 << 6) | c2) << 6) | c3;
+ }
+ else if ((ch & 0xC0) == 0xC0) {
+ // 0x00000080 - 0x000007FF [min. 16bit storage, 2-byte encoding]
+ // 110bbbaa 10aaaaaa
+ int c1 = ch & 0x1F; ch = Buffer::Read();
+ int c2 = ch & 0x3F;
+ ch = (c1 << 6) | c2;
+ }
+ return ch;
+}
+
+
+bool UTF8Buffer::isUTF8() const
+{
+ return true;
+}
+
+
+int Buffer::Peek()
+{
+ int curPos = GetPos();
+ int ch = Read();
+ SetPos(curPos);
+ return ch;
+}
+
+
+int Buffer::GetPos() const
+{
+ if (stdStream)
+ {
+ return stdStream->tellg();
+ }
+
+ return bufPos + bufStart;
+}
+
+
+void Buffer::SetPos(int value)
+{
+ if (stdStream)
+ {
+ stdStream->seekg(value, std::ios::beg);
+ return;
+ }
+
+ if ((value >= fileLen) && cStream && !CanSeek())
+ {
+ // Wanted position is after buffer and the stream
+ // is not seek-able e.g. network or console,
+ // thus we have to read the stream manually till
+ // the wanted position is in sight.
+ while ((value >= fileLen) && (ReadNextStreamChunk() > 0))
+ {}
+ }
+
+ if ((value < 0) || (value > fileLen))
+ {
+ fwprintf(stderr, L"--- buffer out of bounds access, position: %d\n", value);
+ ::exit(1);
+ }
+
+ if ((value >= bufStart) && (value < (bufStart + bufLen))) // already in buffer
+ {
+ bufPos = value - bufStart;
+ }
+ else if (cStream) // must be swapped in
+ {
+ fseek(cStream, value, SEEK_SET);
+ bufLen = fread(buf, sizeof(char), bufCapacity, cStream);
+ bufStart = value; bufPos = 0;
+ }
+ else
+ {
+ bufPos = fileLen - bufStart; // make Pos return fileLen
+ }
+}
+
+
+//
+// Read the next chunk of bytes from the stream, increases the buffer
+// if needed and updates the fields fileLen and bufLen.
+// Returns the number of bytes read.
+//
+int Buffer::ReadNextStreamChunk()
+{
+ int freeLen = bufCapacity - bufLen;
+ if (freeLen == 0)
+ {
+ // in the case of a growing input stream
+ // we can neither seek in the stream, nor can we
+ // foresee the maximum length, thus we must adapt
+ // the buffer size on demand.
+ bufCapacity = bufLen * 2;
+ unsigned char *newBuf = new unsigned char[bufCapacity];
+ memcpy(newBuf, buf, bufLen*sizeof(char));
+ delete[] buf;
+ buf = newBuf;
+ freeLen = bufLen;
+ }
+ int read = fread(buf + bufLen, sizeof(char), freeLen, cStream);
+ if (read > 0)
+ {
+ fileLen = bufLen = (bufLen + read);
+ return read;
+ }
+ // end of stream reached
+ return 0;
+}
+
+
+bool Buffer::CanSeek() const
+{
+ return cStream && (ftell(cStream) != -1);
+}
+
+// ----------------------------------------------------------------------------
+// Scanner Implementation
+// ----------------------------------------------------------------------------
+
+Scanner::Scanner(const char* buf, int len)
+:
+ buffer(new Buffer(buf, len))
+{
+ Init();
+}
+
+
+Scanner::Scanner(const unsigned char* buf, int len)
+:
+ buffer(new Buffer(buf, len))
+{
+ Init();
+}
+
+
+Scanner::Scanner(FILE* ifh)
+:
+ buffer(new Buffer(ifh, true))
+{
+ Init();
+}
+
+
+#ifdef _WIN32
+Scanner::Scanner(const std::wstring& fileName)
+{
+ FILE* ifh;
+
+ if ((ifh = _wfopen(fileName.c_str(), L"rb")) == NULL)
+ {
+ fwprintf(stderr, L"--- Cannot open file %ls\n", fileName.c_str());
+ ::exit(1);
+ }
+ buffer = new Buffer(ifh, false);
+ Init();
+}
+#endif
+
+
+Scanner::Scanner(const std::string& fileName)
+{
+ FILE* ifh;
+ if ((ifh = fopen(fileName.c_str(), "rb")) == NULL)
+ {
+ fwprintf(stderr, L"--- Cannot open file %s\n", fileName.c_str());
+ ::exit(1);
+ }
+ buffer = new Buffer(ifh, false);
+ Init();
+}
+
+
+Scanner::Scanner(std::istream& istr)
+:
+ buffer(new Buffer(&istr, true))
+{
+ Init();
+}
+
+
+Scanner::~Scanner()
+{
+ char* cur = reinterpret_cast<char*>(firstHeap);
+
+#ifdef COCO_DEBUG_HEAP
+ fwprintf(stderr, L"~Scanner:\n");
+#endif
+
+ while (cur)
+ {
+ cur = *(reinterpret_cast<char**>(cur + HEAP_BLOCK_SIZE));
+ free(firstHeap);
+#ifdef COCO_DEBUG_HEAP
+ fwprintf
+ (
+ stderr, L" free %p -> %p\n",
+ firstHeap,
+ reinterpret_cast<char*>(firstHeap) + HEAP_BLOCK_SIZE
+ );
+#endif
+ firstHeap = cur;
+ }
+ delete[] tval;
+ delete buffer;
+}
+
+
+void Scanner::Init()
+{
+ for (int i = 36; i <= 36; ++i) start.set(i, 7);
+ for (int i = 65; i <= 90; ++i) start.set(i, 7);
+ for (int i = 95; i <= 95; ++i) start.set(i, 7);
+ for (int i = 97; i <= 122; ++i) start.set(i, 7);
+ start.set(34, 1);
+ start.set(39, 4);
+ start.set(35, 11);
+ start.set(10, 12);
+ start.set(59, 13);
+ start.set(Buffer::EoF, -1);
+
+ keywords.set(L"include", 6);
+ keywords.set(L"import", 8);
+
+ tvalLength = 128;
+ tval = new wchar_t[tvalLength]; // text of current token
+ tlen = 0;
+ tval[tlen] = 0;
+
+ // HEAP_BLOCK_SIZE byte heap + pointer to next heap block
+ heap = malloc(HEAP_BLOCK_SIZE + sizeof(void*));
+ firstHeap = heap;
+ heapEnd =
+ reinterpret_cast<void**>
+ (reinterpret_cast<char*>(heap) + HEAP_BLOCK_SIZE);
+ *heapEnd = 0;
+ heapTop = heap;
+ if (sizeof(Token) > HEAP_BLOCK_SIZE)
+ {
+ fwprintf(stderr, L"--- Too small HEAP_BLOCK_SIZE\n");
+ ::exit(1);
+ }
+#ifdef COCO_DEBUG_HEAP
+ fwprintf
+ (
+ stderr, L"Scanner::init: firstHeap %p -> %p\n",
+ firstHeap,
+ reinterpret_cast<char*>(firstHeap) + HEAP_BLOCK_SIZE
+ );
+#endif
+
+ pos = -1; line = 1; col = 0;
+ oldEols = 0;
+ NextCh();
+ if (ch == 0xEF) // check optional byte order mark for UTF-8
+ { // Windows-specific magic
+ NextCh(); int ch1 = ch;
+ NextCh(); int ch2 = ch;
+ if (ch1 != 0xBB || ch2 != 0xBF)
+ {
+ fwprintf(stderr, L"Illegal byte order mark at start of file");
+ ::exit(1);
+ }
+ Buffer *oldBuf = buffer;
+ buffer = new UTF8Buffer(oldBuf); col = 0;
+ delete oldBuf; oldBuf = NULL;
+ NextCh();
+ }
+ else
+ {
+ // FORCE_UTF8 was defined
+ // use UTF8Buffer without relying on a byte order mark.
+ Buffer *oldBuf = buffer;
+ buffer = new UTF8Buffer(oldBuf); col = 0;
+ delete oldBuf; oldBuf = NULL;
+ }
+
+ pt = tokens = CreateToken(); // first token is a dummy
+}
+
+
+void Scanner::NextCh()
+{
+ if (oldEols > 0)
+ {
+ ch = EOL;
+ oldEols--;
+ }
+ else
+ {
+ pos = buffer->GetPos();
+ ch = buffer->Read(); col++;
+ // replace isolated '\r' by '\n' in order to make
+ // eol handling uniform across Windows, Unix and Mac
+ if (ch == '\r' && buffer->Peek() != '\n') ch = EOL;
+ if (ch == EOL) { line++; col = 0; }
+ }
+}
+
+
+void Scanner::AddCh()
+{
+ if (tlen >= tvalLength)
+ {
+ tvalLength *= 2;
+ wchar_t *newBuf = new wchar_t[tvalLength];
+ memcpy(newBuf, tval, tlen*sizeof(wchar_t));
+ delete[] tval;
+ tval = newBuf;
+ }
+ if (ch != Buffer::EoF)
+ {
+ tval[tlen++] = ch;
+ NextCh();
+ }
+}
+
+
+
+bool Scanner::Comment0() {
+ int level = 1, pos0 = pos, line0 = line, col0 = col;
+ NextCh();
+ if (ch == '/') {
+ NextCh();
+ while (true) {
+ if (ch == 10) {
+ level--;
+ if (level == 0) { oldEols = line - line0; NextCh(); return true; }
+ NextCh();
+ } else if (ch == buffer->EoF) return false;
+ else NextCh();
+ }
+ } else {
+ buffer->SetPos(pos0); NextCh(); line = line0; col = col0;
+ }
+ return false;
+}
+
+bool Scanner::Comment1() {
+ int level = 1, pos0 = pos, line0 = line, col0 = col;
+ NextCh();
+ if (ch == '*') {
+ NextCh();
+ while (true) {
+ if (ch == '*') {
+ NextCh();
+ if (ch == '/') {
+ level--;
+ if (level == 0) { oldEols = line - line0; NextCh(); return true; }
+ NextCh();
+ }
+ } else if (ch == '/') {
+ NextCh();
+ if (ch == '*') {
+ level++; NextCh();
+ }
+ } else if (ch == buffer->EoF) return false;
+ else NextCh();
+ }
+ } else {
+ buffer->SetPos(pos0); NextCh(); line = line0; col = col0;
+ }
+ return false;
+}
+
+void Scanner::CreateHeapBlock()
+{
+ char* cur = reinterpret_cast<char*>(firstHeap);
+
+#ifdef COCO_DEBUG_HEAP
+ fwprintf(stderr, L"CreateHeapBlock: tokens %p\n", tokens);
+#endif
+
+ // release unused blocks
+ while
+ (
+ (reinterpret_cast<char*>(tokens) < cur)
+ || (reinterpret_cast<char*>(tokens) > (cur + HEAP_BLOCK_SIZE))
+ )
+ {
+ cur = *(reinterpret_cast<char**>(cur + HEAP_BLOCK_SIZE));
+#ifdef COCO_DEBUG_HEAP
+ fwprintf
+ (
+ stderr, L" free %p -> %p\n",
+ firstHeap,
+ reinterpret_cast<char*>(firstHeap) + HEAP_BLOCK_SIZE
+ );
+#endif
+ free(firstHeap);
+ firstHeap = cur;
+ }
+
+ // HEAP_BLOCK_SIZE byte heap + pointer to next heap block
+ void* newHeap = malloc(HEAP_BLOCK_SIZE + sizeof(void*));
+ *heapEnd = newHeap;
+ heapEnd =
+ reinterpret_cast<void**>
+ (reinterpret_cast<char*>(newHeap) + HEAP_BLOCK_SIZE);
+ *heapEnd = 0;
+ heap = newHeap;
+ heapTop = heap;
+#ifdef COCO_DEBUG_HEAP
+ fwprintf
+ (
+ stderr, L" malloc %p -> %p\n",
+ newHeap,
+ reinterpret_cast<char*>(newHeap) + HEAP_BLOCK_SIZE
+ );
+#endif
+}
+
+
+Token* Scanner::CreateToken()
+{
+ const int reqMem = sizeof(Token);
+ if
+ (
+ (reinterpret_cast<char*>(heapTop) + reqMem)
+ >= reinterpret_cast<char*>(heapEnd)
+ )
+ {
+ CreateHeapBlock();
+ }
+ // token 'occupies' heap starting at heapTop
+ Token* tok = reinterpret_cast<Token*>(heapTop);
+ // increment past this part of the heap, which is now used
+ heapTop =
+ reinterpret_cast<void*>
+ (reinterpret_cast<char*>(heapTop) + reqMem);
+ tok->val = NULL;
+ tok->next = NULL;
+ return tok;
+}
+
+
+void Scanner::AppendVal(Token* tok)
+{
+ const int reqMem = (tlen + 1) * sizeof(wchar_t);
+ if
+ (
+ (reinterpret_cast<char*>(heapTop) + reqMem)
+ >= reinterpret_cast<char*>(heapEnd)
+ )
+ {
+ if (reqMem > HEAP_BLOCK_SIZE)
+ {
+ fwprintf(stderr, L"--- Too long token value\n");
+ ::exit(1);
+ }
+ CreateHeapBlock();
+ }
+
+ // add text value from heap
+ tok->val = reinterpret_cast<wchar_t*>(heapTop);
+
+ // increment past this part of the heap, which is now used
+ heapTop =
+ reinterpret_cast<void*>
+ (reinterpret_cast<char*>(heapTop) + reqMem);
+
+ // copy the currently parsed tval into the token
+ wcsncpy(tok->val, tval, tlen);
+ tok->val[tlen] = '\0';
+}
+
+
+Token* Scanner::NextToken()
+{
+ while
+ (
+ ch == ' '
+ || ch == 9
+ ) NextCh();
+ if ((ch == '/' && Comment0()) || (ch == '/' && Comment1())) return NextToken();
+ int recKind = noSym;
+ int recEnd = pos;
+ t = CreateToken();
+ t->pos = pos; t->col = col; t->line = line;
+ int state = start.state(ch);
+ tlen = 0; AddCh();
+
+ switch (state)
+ {
+ case -1: { t->kind = eofSym; break; } // NextCh already done
+ case 0: {
+ case_0:
+ if (recKind != noSym) {
+ tlen = recEnd - t->pos;
+ SetScannerBehindT();
+ }
+ t->kind = recKind; break;
+ } // NextCh already done
+ case 1:
+ case_1:
+ if (ch <= 9 || (ch >= 11 && ch <= 12) || (ch >= 14 && ch <= '!') || (ch >= '#' && ch <= '[') || (ch >= ']' && ch <= 65535)) {AddCh(); goto case_1;}
+ else if (ch == '"') {AddCh(); goto case_3;}
+ else if (ch == 92) {AddCh(); goto case_2;}
+ else {goto case_0;}
+ case 2:
+ case_2:
+ if ((ch >= ' ' && ch <= '~')) {AddCh(); goto case_1;}
+ else {goto case_0;}
+ case 3:
+ case_3:
+ {t->kind = 1; break;}
+ case 4:
+ case_4:
+ if (ch <= 9 || (ch >= 11 && ch <= 12) || (ch >= 14 && ch <= '!') || (ch >= '#' && ch <= '&') || (ch >= '(' && ch <= '[') || (ch >= ']' && ch <= 65535)) {AddCh(); goto case_4;}
+ else if (ch == 39) {AddCh(); goto case_8;}
+ else if (ch == 92) {AddCh(); goto case_5;}
+ else {goto case_0;}
+ case 5:
+ case_5:
+ if ((ch >= ' ' && ch <= '~')) {AddCh(); goto case_4;}
+ else {goto case_0;}
+ case 6:
+ case_6:
+ {t->kind = 4; break;}
+ case 7:
+ case_7:
+ recEnd = pos; recKind = 3;
+ if (ch == '$' || (ch >= '0' && ch <= '9') || (ch >= 'A' && ch <= 'Z') || ch == '_' || (ch >= 'a' && ch <= 'z')) {AddCh(); goto case_7;}
+ else if (ch == '.') {AddCh(); goto case_9;}
+ else {t->kind = 3; std::wstring literal(tval, tlen); t->kind = keywords.get(literal, t->kind); break;}
+ case 8:
+ case_8:
+ recEnd = pos; recKind = 2;
+ if (ch <= 9 || (ch >= 11 && ch <= 12) || (ch >= 14 && ch <= '!') || (ch >= '#' && ch <= '&') || (ch >= '(' && ch <= '[') || (ch >= ']' && ch <= 65535)) {AddCh(); goto case_4;}
+ else if (ch == 39) {AddCh(); goto case_8;}
+ else if (ch == 92) {AddCh(); goto case_5;}
+ else {t->kind = 2; break;}
+ case 9:
+ case_9:
+ if (ch == '$' || (ch >= 'A' && ch <= 'Z') || ch == '_' || (ch >= 'a' && ch <= 'z')) {AddCh(); goto case_10;}
+ else if (ch == '*') {AddCh(); goto case_6;}
+ else {goto case_0;}
+ case 10:
+ case_10:
+ recEnd = pos; recKind = 3;
+ if (ch == '$' || (ch >= '0' && ch <= '9') || (ch >= 'A' && ch <= 'Z') || ch == '_' || (ch >= 'a' && ch <= 'z')) {AddCh(); goto case_10;}
+ else if (ch == '.') {AddCh(); goto case_9;}
+ else {t->kind = 3; std::wstring literal(tval, tlen); t->kind = keywords.get(literal, t->kind); break;}
+ case 11:
+ {t->kind = 5; break;}
+ case 12:
+ {t->kind = 7; break;}
+ case 13:
+ {t->kind = 9; break;}
+ }
+ AppendVal(t);
+ return t;
+}
+
+
+void Scanner::SetScannerBehindT()
+{
+ buffer->SetPos(t->pos);
+ NextCh();
+ line = t->line; col = t->col;
+ for (int i = 0; i < tlen; i++) NextCh();
+}
+
+
+// get the next token (possibly a token already seen during peeking)
+Token* Scanner::Scan()
+{
+ if (tokens->next == NULL) {
+ pt = tokens = NextToken();
+ }
+ else {
+ pt = tokens = tokens->next;
+ }
+ return tokens;
+}
+
+
+// peek for the next token, ignore pragmas
+Token* Scanner::Peek()
+{
+ do
+ {
+ if (pt->next == NULL)
+ {
+ pt->next = NextToken();
+ }
+ pt = pt->next;
+ } while (pt->kind > maxT); // skip pragmas
+
+ return pt;
+}
+
+
+// make sure that peeking starts at the current scan position
+void Scanner::ResetPeek()
+{
+ pt = tokens;
+}
+
+
+int Scanner::Line() const
+{
+ return line;
+}
+
+
+void Scanner::Line(int lineNo)
+{
+ line = lineNo;
+}
+
+
+// * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * //
+
+} // End namespace
+
+// ************************************************************************* //
diff --git a/wmake/src/wmkdependScanner.h b/wmake/src/wmkdependScanner.h
new file mode 100644
index 00000000000..a72bf4b9684
--- /dev/null
+++ b/wmake/src/wmkdependScanner.h
@@ -0,0 +1,477 @@
+/*---------------------------------*- C++ -*---------------------------------*\
+ ========= |
+ \\ / F ield | OpenFOAM: The Open Source CFD Toolbox
+ \\ / O peration |
+ \\ / A nd | Copyright (C) 2010-2010 OpenCFD Ltd.
+ \\/ M anipulation |
+-------------------------------------------------------------------------------
+License
+ This file is part of OpenFOAM.
+
+ OpenFOAM is free software; you can redistribute it and/or modify it
+ under the terms of the GNU General Public License as published by the
+ Free Software Foundation; either version 2 of the License, or (at your
+ option) any later version.
+
+ OpenFOAM is distributed in the hope that it will be useful, but WITHOUT
+ ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License
+ for more details.
+
+ You should have received a copy of the GNU General Public License
+ along with OpenFOAM; if not, write to the Free Software Foundation,
+ Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA
+
+@file wmkdependParser.atg
+
+Description
+ An attributed Coco/R grammar to parse C/C++, Fortran and Java files
+ for include and import statements.
+
+SourceFiles
+ generated
+
+\*---------------------------------------------------------------------------*/
+// This file was generated with Coco/R C++ (7 Feb 2010)
+// http://www.ssw.uni-linz.ac.at/coco/
+// with these defines:
+// - FORCE_UTF8
+
+
+#ifndef COCO_wmkdependSCANNER_H__
+#define COCO_wmkdependSCANNER_H__
+
+#include <cstdio>
+#include <cstdlib>
+#include <cstring>
+#include <cwchar>
+#include <string>
+#include <fstream>
+#include <iostream>
+
+namespace wmake {
+
+// * * * * * * * * * * * Miscellaneous String Routines * * * * * * * * * * * //
+
+//! Simple lower-case string transformation
+template<class StringT>
+inline void coco_string_toLower(StringT& str)
+{
+ for
+ (
+ typename StringT::iterator iter = str.begin();
+ iter != str.end();
+ ++iter
+ )
+ {
+ if (*iter >= 'A' && *iter <= 'Z')
+ {
+ *iter += ('a' - 'A'); // lower-case
+ }
+ }
+}
+
+
+//! Simple string hashing function
+template<class StringT>
+inline int coco_string_hash(const StringT& str)
+{
+ int h = 0;
+ for
+ (
+ typename StringT::const_iterator iter = str.begin();
+ iter != str.end();
+ ++iter
+ )
+ {
+ h = (h * 7) ^ *iter;
+ }
+ return h < 0 ? -h : h;
+}
+
+
+//
+// String conversions
+// ~~~~~~~~~~~~~~~~~~
+
+//! Convert wide string to double
+inline double coco_string_toDouble(const wchar_t* str)
+{
+ return str ? wcstod(str, NULL) : 0;
+}
+
+
+//! Convert wide string to long
+inline long coco_string_toLong(const wchar_t* str)
+{
+ return str ? wcstol(str, NULL, 10) : 0;
+}
+
+
+//! A byte string (restricted to 8bit values) by copying str
+std::string coco_stdString(const wchar_t* str);
+
+//! A byte string (restricted to 8bit values) by copying str,
+//! up to length characters long
+std::string coco_stdString(const wchar_t* str, unsigned length);
+
+//! A byte substring (restricted to 8bit values) of str,
+//! starting at index and length characters long
+std::string coco_stdString(const wchar_t* str, unsigned index, unsigned length);
+
+//! A UTF8 byte string by copying str
+std::string coco_stdStringUTF8(const wchar_t* str);
+
+//! A UTF8 byte string by copying str, up to length characters long
+std::string coco_stdStringUTF8(const wchar_t* str, unsigned length);
+
+//! A UTF8 byte substring, starting at index and length characters long
+std::string coco_stdStringUTF8(const wchar_t* str, unsigned index, unsigned length);
+
+// * * * * * * * * * * * * End of String Routines * * * * * * * * * * * * * //
+
+
+
+/*---------------------------------------------------------------------------*\
+ Class Token Declaration
+\*---------------------------------------------------------------------------*/
+/*!
+ * @brief Scanner Token
+ *
+ * @note since each Token is allocated by the internal heap mechanism,
+ * the destructor does not clean up the val member.
+ */
+class Token
+{
+public:
+ int kind; //!< token kind
+ int pos; //!< token position in the source text (starting at 0)
+ int col; //!< token column (starting at 1)
+ int line; //!< token line (starting at 1)
+ wchar_t* val; //!< token value (normally allocated from the internal heap)
+ Token *next; //!< Peek tokens are kept in linked list
+
+ int length() const; //!< The length of val, or 0 if val is NULL
+
+ //! Construct null Token, optionally with pointer to a string value
+ Token(wchar_t* value = 0);
+ ~Token(); //!< Destructor - does not cleanup val member
+
+ //! Token val as byte string (restricted to 8bit values)
+ inline std::string toString() const
+ {
+ return coco_stdString(val);
+ }
+
+ //! Token val as byte string (restricted to 8bit values), up to length characters long
+ inline std::string toString(unsigned length) const
+ {
+ return coco_stdString(val, length);
+ }
+
+ //! Token val as byte string (restricted to 8bit values), starting at index and length characters long
+ inline std::string toString(unsigned index, unsigned length) const
+ {
+ return coco_stdString(val, index, length);
+ }
+
+ //! Token val as UTF8 byte string
+ inline std::string toStringUTF8() const
+ {
+ return coco_stdStringUTF8(val);
+ }
+
+ //! Token val as UTF8 byte string, up to length characters long
+ inline std::string toStringUTF8(unsigned length) const
+ {
+ return coco_stdStringUTF8(val, length);
+ }
+
+ //! Token val as UTF8 byte substring, starting at index and length characters long
+ inline std::string toStringUTF8(unsigned index, unsigned length) const
+ {
+ return coco_stdStringUTF8(this->val, index, length);
+ }
+
+};
+
+
+/*---------------------------------------------------------------------------*\
+ Class Buffer Declaration
+\*---------------------------------------------------------------------------*/
+/*!
+ * @brief Scanner Buffer Token
+ *
+ * This Buffer supports the following cases:
+ * -# seekable stream (file)
+ * -# whole stream in buffer
+ * -# part of stream in buffer
+ * -# non seekable stream (network, console)
+ */
+class Buffer
+{
+ unsigned char *buf; //!< input buffer
+ int bufCapacity; //!< capacity of buf
+ int bufLen; //!< length of buffer
+ int bufPos; //!< current position in buffer
+ int bufStart; //!< position of first byte in buffer relative to input stream
+ int fileLen; //!< length of input stream (may change if the stream is no file)
+ FILE* cStream; //!< input stdio stream (normally seekable)
+ std::istream* stdStream; //!< STL std stream (seekable)
+ bool isUserStream_; //!< was the stream opened by the user?
+
+ int ReadNextStreamChunk();
+ bool CanSeek() const; //!< true if stream can be seeked otherwise false
+
+protected:
+ Buffer(Buffer*); //!< for the UTF8Buffer
+
+public:
+ //! max unicode characters is 0xFFFF (16bit storage)
+ static const int MaxChar = 65535;
+ static const int EoF = MaxChar + 1;
+
+ //! Copy buffer contents from constant character string
+ Buffer(const char* chars, int len);
+
+ //! Copy buffer contents from constant character string
+ Buffer(const unsigned char* chars, int len);
+
+ //! @brief Attach buffer to a stdio stream.
+ //! User streams are not closed in the destructor
+ Buffer(FILE*, bool isUserStream = true);
+
+ //! @brief Attach buffer to an STL standard stream
+ //! User streams are not closed in the destructor
+ explicit Buffer(std::istream*, bool isUserStream = true);
+
+ //! Close stream (but not user streams) and free buf (if any)
+ virtual ~Buffer();
+
+ virtual void Close(); //!< Close stream (but not user streams)
+ virtual int Read(); //!< Get character from stream or buffer
+ virtual int Peek(); //!< Peek character from stream or buffer
+
+ virtual int GetPos() const;
+ virtual void SetPos(int value);
+ virtual bool isUTF8() const; //!< Return false - buffer is not UTF8
+};
+
+
+/*---------------------------------------------------------------------------*\
+ Class UTF8Buffer Declaration
+\*---------------------------------------------------------------------------*/
+//! A Scanner Buffer variant that decodes UTF-8 characters into 16bit unicode
+class UTF8Buffer : public Buffer
+{
+public:
+ UTF8Buffer(Buffer* b) : Buffer(b) {}
+ virtual int Read();
+ virtual bool isUTF8() const; //!< Return true - buffer is UTF8
+};
+
+
+/*---------------------------------------------------------------------------*\
+ Class StartStates Declaration
+\*---------------------------------------------------------------------------*/
+//! maps characters (integers) to start states of tokens as a HashTable
+class StartStates
+{
+ //! HashTable entry
+ struct Entry
+ {
+ int key; //<! The lookup key
+ int val; //<! The data
+ Entry *next; //<! Pointer next Entry in sub-list
+
+ Entry(int k, int v, Entry *n=0)
+ :
+ key(k), val(v), next(n)
+ {}
+ };
+
+ static const int size_ = 128; //<! fixed HashTable size
+ Entry **table_;
+
+public:
+ StartStates()
+ :
+ table_(new Entry*[size_])
+ {
+ memset(table_, 0, size_*sizeof(Entry*));
+ }
+
+ virtual ~StartStates()
+ {
+ for (int i = 0; i < size_; ++i)
+ {
+ Entry *e = table_[i];
+ while (e)
+ {
+ Entry *next = e->next;
+ delete e;
+ e = next;
+ }
+ }
+ delete[] table_;
+ }
+
+ void set(int key, int val)
+ {
+ const int hashIndex = unsigned(key) % size_;
+ table_[hashIndex] = new Entry(key, val, table_[hashIndex]);
+ }
+
+ int state(int key)
+ {
+ Entry *e = table_[unsigned(key) % size_];
+ while (e && e->key != key) e = e->next;
+ return e ? e->val : 0;
+ }
+};
+
+
+/*---------------------------------------------------------------------------*\
+ Class KeywordMap Declaration
+\*---------------------------------------------------------------------------*/
+//! maps strings to integers (identifiers to keyword kinds) as a HashTable
+class KeywordMap
+{
+ //! HashTable entry
+ struct Entry
+ {
+ const std::wstring key; //<! The lookup key
+ int val; //<! The data
+ Entry *next; //<! Pointer next Entry in sub-list
+
+ Entry(const std::wstring& k, int v, Entry *n=0)
+ :
+ key(k), val(v), next(n)
+ {}
+ };
+
+ static const int size_ = 128; //<! fixed HashTable size
+ Entry **table_;
+
+public:
+ KeywordMap()
+ :
+ table_(new Entry*[size_])
+ {
+ memset(table_, 0, size_*sizeof(Entry*));
+ }
+
+ virtual ~KeywordMap()
+ {
+ for (int i = 0; i < size_; ++i)
+ {
+ Entry *e = table_[i];
+ while (e)
+ {
+ Entry *next = e->next;
+ delete e;
+ e = next;
+ }
+ }
+ delete[] table_;
+ }
+
+ void set(const std::wstring& key, int val)
+ {
+ const int hashIndex = coco_string_hash(key) % size_;
+ table_[hashIndex] = new Entry(key, val, table_[hashIndex]);
+ }
+
+ int get(const std::wstring& key, int defaultVal)
+ {
+ Entry *e = table_[coco_string_hash(key) % size_];
+ while (e && e->key != key) e = e->next;
+ return e ? e->val : defaultVal;
+ }
+};
+
+
+/*---------------------------------------------------------------------------*\
+ Class Scanner Declaration
+\*---------------------------------------------------------------------------*/
+//! A Coco/R Scanner
+class Scanner
+{
+ static const int maxT = 10;
+ static const int noSym = 10;
+ static const int eofSym = 0; //!< end-of-file token id
+ static const char EOL = '\n'; //!< end-of-line character
+
+ void *firstHeap; //!< the start of the heap management
+ void *heap; //!< the currently active block
+ void *heapTop; //!< the top of the heap
+ void **heapEnd; //!< the end of the last heap block
+
+ StartStates start; //!< A map of start states for particular characters
+ KeywordMap keywords; //!< A hash of keyword literals to token kind
+
+ Token *t; //!< current token
+ wchar_t *tval; //!< text of current token
+ int tvalLength; //!< maximum capacity (length) for tval
+ int tlen; //!< length of tval
+
+ Token *tokens; //!< list of tokens already peeked (first token is a dummy)
+ Token *pt; //!< current peek token
+
+ int ch; //!< current input character
+ int pos; //!< byte position of current character
+ int line; //!< line number of current character
+ int col; //!< column number of current character
+ int oldEols; //!< the number of EOLs that appeared in a comment
+
+ void CreateHeapBlock(); //!< add a heap block, freeing unused ones
+ Token* CreateToken(); //!< fit token on the heap
+ void AppendVal(Token* tok); //!< adjust tok->val to point to the heap and copy tval into it
+ void SetScannerBehindT();
+
+ void Init(); //!< complete the initialization for the constructors
+ void NextCh(); //!< get the next input character into ch
+ void AddCh(); //!< append the character ch to tval
+ bool Comment0();
+ bool Comment1();
+ Token* NextToken(); //!< get the next token
+
+public:
+ //! The scanner buffer
+ Buffer *buffer;
+
+ //! Attach scanner to an existing character buffer
+ Scanner(const char* chars, int len);
+
+ //! Attach scanner to an existing character buffer
+ Scanner(const unsigned char* chars, int len);
+
+ //! Attach scanner to an existing open file handle
+ Scanner(FILE*);
+
+#ifdef _WIN32
+ //! Open a file for reading and attach scanner - Windows only
+ explicit Scanner(const std::wstring& fileName);
+#endif
+
+ //! Open a file for reading and attach scanner
+ explicit Scanner(const std::string& fileName);
+
+ //! Attach scanner to an existing open STL standard stream
+ explicit Scanner(std::istream&);
+
+ ~Scanner(); //!< free heap and allocated memory
+ Token* Scan(); //!< get the next token (possibly a token already seen during peeking)
+ Token* Peek(); //!< peek for the next token, ignore pragmas
+ void ResetPeek(); //!< ensure that peeking starts at the current scan position
+
+ int Line() const; //!< Return the current line
+ void Line(int lineNo); //!< Define the starting line for reporting errors
+
+}; // end Scanner
+
+} // End namespace
+
+#endif // COCO_wmkdependSCANNER_H__
+
+// ************************************************************************* //
--
GitLab