summaryrefslogtreecommitdiffstats
path: root/libksieve
diff options
context:
space:
mode:
Diffstat (limited to 'libksieve')
-rw-r--r--libksieve/Makefile.am17
-rw-r--r--libksieve/impl/lexer.h189
-rw-r--r--libksieve/impl/parser.h118
-rw-r--r--libksieve/impl/utf8validator.h42
-rw-r--r--libksieve/ksieve/Makefile.am8
-rw-r--r--libksieve/ksieve/error.h139
-rw-r--r--libksieve/ksieve/lexer.h108
-rw-r--r--libksieve/ksieve/parser.h72
-rw-r--r--libksieve/ksieve/scriptbuilder.h80
-rw-r--r--libksieve/parser/Makefile.am12
-rw-r--r--libksieve/parser/lexer.cpp666
-rw-r--r--libksieve/parser/parser.cpp651
-rw-r--r--libksieve/parser/utf8validator.cpp141
-rw-r--r--libksieve/shared/Makefile.am8
-rw-r--r--libksieve/shared/error.cpp247
-rw-r--r--libksieve/tests/Makefile.am13
-rw-r--r--libksieve/tests/lexertest.cpp484
-rw-r--r--libksieve/tests/parsertest.cpp667
18 files changed, 3662 insertions, 0 deletions
diff --git a/libksieve/Makefile.am b/libksieve/Makefile.am
new file mode 100644
index 000000000..fa7ff1a32
--- /dev/null
+++ b/libksieve/Makefile.am
@@ -0,0 +1,17 @@
+SUBDIRS = ksieve shared parser . tests
+
+INCLUDES = $(all_includes)
+
+lib_LTLIBRARIES = libksieve.la
+
+CLEANFILES = dummy.cpp
+
+libksieve_la_SOURCES = dummy.cpp
+libksieve_la_LIBADD = parser/libksieve_parser.la
+libksieve_la_LDFLAGS = $(all_libraries) -version-info 0:0:0 -no-undefined
+
+dummy.cpp:
+ echo > dummy.cpp
+
+messages:
+ $(XGETTEXT) shared/*.cpp parser/*.cpp impl/*.h ksieve/*.h -o $(podir)/libksieve.pot
diff --git a/libksieve/impl/lexer.h b/libksieve/impl/lexer.h
new file mode 100644
index 000000000..0eba0e585
--- /dev/null
+++ b/libksieve/impl/lexer.h
@@ -0,0 +1,189 @@
+/* -*- c++ -*-
+ impl/lexer.h
+
+ Internal header file. Subject to change without notice. DO NOT USE.
+
+ This file is part of KSieve,
+ the KDE internet mail/usenet news message filtering library.
+ Copyright (c) 2003 Marc Mutz <mutz@kde.org>
+
+ KSieve is free software; you can redistribute it and/or modify it
+ under the terms of the GNU General Public License, version 2, as
+ published by the Free Software Foundation.
+
+ KSieve 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 this program; if not, write to the Free Software
+ Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
+
+ In addition, as a special exception, the copyright holders give
+ permission to link the code of this program with any edition of
+ the Qt library by Trolltech AS, Norway (or with modified versions
+ of Qt that use the same license as Qt), and distribute linked
+ combinations including the two. You must obey the GNU General
+ Public License in all respects for all of the code used other than
+ Qt. If you modify this file, you may extend this exception to
+ your version of the file, but you are not obligated to do so. If
+ you do not wish to do so, delete this exception statement from
+ your version.
+*/
+
+#ifndef __KSIEVE_IMPL_LEXER_H__
+#define __KSIEVE_IMPL_LEXER_H__
+
+#include <ksieve/lexer.h>
+#include <ksieve/error.h>
+
+#include <qvaluestack.h>
+#include <qcstring.h>
+
+namespace KSieve {
+
+ class Lexer::Impl {
+ public:
+ Impl( const char * scursor, const char * send, int options );
+
+ bool ignoreComments() const {
+ return mIgnoreComments;
+ }
+
+ bool ignoreLineFeeds() const {
+ return mIgnoreLF;
+ }
+
+ const Error & error() const {
+ return mState.error;
+ }
+
+ bool atEnd() const {
+ return mState.cursor >= mEnd;
+ }
+
+ int column() const {
+ return mState.cursor - mState.beginOfLine;
+ }
+
+ int line() const {
+ return mState.line;
+ }
+
+ void save() {
+ mStateStack.push( mState );
+ }
+
+ void restore() {
+ mState = mStateStack.pop();
+ }
+
+ Lexer::Token nextToken( QString & tokenValue );
+
+ private:
+ /** Cursor must be positioned on the \r or the \n. */
+ bool eatCRLF();
+
+ /** Cursor must be positioned after the opening hash (#). If
+ parsing is successful, cursor is positioned behind the CRLF
+ that ended the comment's line (or past the end). */
+ bool parseHashComment( QString & result, bool reallySave=false );
+
+ /** Cursor must be positioned after the opening slash-asterisk */
+ bool parseBracketComment( QString & result, bool reallySave=false );
+
+ /** Cursor must be positioned on the opening '/'or '#' */
+ bool parseComment( QString & result, bool reallySave=false );
+
+ /** Eats whitespace, but not comments */
+ bool eatWS();
+
+ /** Eats comments and whitespace */
+ bool eatCWS();
+
+ /** Cursor must be positioned on the first character */
+ bool parseIdentifier( QString & result );
+
+ /** Cursor must be positioned after the initial ':' */
+ bool parseTag( QString & result );
+
+ /** Cursor must be positioned on the first digit */
+ bool parseNumber( QString & result );
+
+ /** Cursor must be positioned after the "text:" token. */
+ bool parseMultiLine( QString & result );
+
+ /** Cursor must be positioned after the initial " */
+ bool parseQuotedString( QString & result );
+
+ struct State {
+ State( const char * s=0 )
+ : cursor( s ), line( 0 ), beginOfLine( s ), error() {}
+ const char * cursor;
+ int line;
+ const char * beginOfLine;
+ Error error;
+ } mState;
+
+ const char * const mEnd;
+ const bool mIgnoreComments : 1;
+ const bool mIgnoreLF : 1;
+ QValueStack<State> mStateStack;
+
+ const char * beginOfLine() const { return mState.beginOfLine; }
+
+ int _strnicmp( const char * left, const char * right, size_t len ) const {
+ return charsLeft() >= len ? qstrnicmp( left, right, len ) : 1 ;
+ }
+
+ void clearErrors() { mState.error = Error(); }
+
+ unsigned int charsLeft() const {
+ return mEnd - mState.cursor < 0 ? 0 : mEnd - mState.cursor ;
+ }
+ void makeError( Error::Type e ) {
+ makeError( e, line(), column() );
+ }
+ void makeError( Error::Type e, int errorLine, int errorCol ) {
+ mState.error = Error( e, errorLine, errorCol );
+ }
+ void makeIllegalCharError( char ch );
+ void makeIllegalCharError() {
+ makeIllegalCharError( *mState.cursor );
+ }
+ /** Defines the current char to end a line.
+ Warning: increases @p mCursor!
+ **/
+ void newLine() {
+ ++mState.line;
+ mState.beginOfLine = ++mState.cursor;
+ }
+ bool skipTo( char c, bool acceptEnd=false ) {
+ while( !atEnd() ) {
+ if ( *mState.cursor == '\n' || *mState.cursor == '\r' ) {
+ if ( !eatCRLF() ) return false;
+ } else if ( *mState.cursor == c ) {
+ return true;
+ } else {
+ ++mState.cursor;
+ }
+ }
+ return acceptEnd;
+ }
+ bool skipToCRLF( bool acceptEnd=true ) {
+ for ( ; !atEnd() ; ++mState.cursor )
+ if ( *mState.cursor == '\n' || *mState.cursor == '\r' )
+ return eatCRLF();
+ return acceptEnd;
+ }
+ void skipTo8BitEnd() {
+ while ( !atEnd() && (signed char)*mState.cursor < 0 )
+ ++mState.cursor;
+ }
+ void skipToDelim();
+ };
+
+}
+
+#endif // __KSIEVE_IMPL_LEXER_H__
diff --git a/libksieve/impl/parser.h b/libksieve/impl/parser.h
new file mode 100644
index 000000000..736678b43
--- /dev/null
+++ b/libksieve/impl/parser.h
@@ -0,0 +1,118 @@
+/* -*- c++ -*-
+ impl/parser.h
+
+ Internal header file. Subject to change without notice. DO NOT USE.
+
+ This file is part of KSieve,
+ the KDE internet mail/usenet news message filtering library.
+ Copyright (c) 2003 Marc Mutz <mutz@kde.org>
+
+ KSieve is free software; you can redistribute it and/or modify it
+ under the terms of the GNU General Public License, version 2, as
+ published by the Free Software Foundation.
+
+ KSieve 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 this program; if not, write to the Free Software
+ Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
+
+ In addition, as a special exception, the copyright holders give
+ permission to link the code of this program with any edition of
+ the Qt library by Trolltech AS, Norway (or with modified versions
+ of Qt that use the same license as Qt), and distribute linked
+ combinations including the two. You must obey the GNU General
+ Public License in all respects for all of the code used other than
+ Qt. If you modify this file, you may extend this exception to
+ your version of the file, but you are not obligated to do so. If
+ you do not wish to do so, delete this exception statement from
+ your version.
+*/
+
+#ifndef __KSIEVE_IMPL_PARSER_H__
+#define __KSIEVE_IMPL_PARSER_H__
+
+#include <ksieve/parser.h>
+
+#include <ksieve/error.h>
+#include <ksieve/lexer.h>
+#include <impl/lexer.h>
+
+#include <ksieve/scriptbuilder.h>
+
+
+namespace KSieve {
+
+ class Parser::Impl {
+ friend class Parser;
+ private:
+ Impl( const char * scursor, const char * const send, int options=0 );
+
+ void setScriptBuilder( ScriptBuilder * builder ) {
+ mBuilder = builder;
+ }
+ ScriptBuilder * scriptBuilder() const {
+ return mBuilder;
+ }
+
+ bool parse();
+
+ const Error & error() const { return mError == Error::None ? lexer.error() : mError ; }
+
+ bool parseCommandList();
+
+ bool parseCommand();
+
+ bool parseArgumentList();
+
+ bool parseArgument();
+
+ bool parseTestList();
+
+ bool parseTest();
+
+ bool parseBlock();
+
+ bool parseStringList();
+
+ bool parseNumber();
+
+
+ Lexer::Token token() const { return mToken; }
+ QString tokenValue() const { return mTokenValue; }
+
+ bool atEnd() const {
+ return !mToken && lexer.atEnd() ;
+ }
+ bool obtainToken();
+ void consumeToken() {
+ mToken = Lexer::None;
+ mTokenValue = QString::null;
+ }
+ void makeError( Error::Type e, int line, int col ) {
+ mError = Error( e, line, col );
+ if ( scriptBuilder() )
+ scriptBuilder()->error( mError );
+ }
+ void makeError( Error::Type e ) {
+ makeError( e, lexer.line(), lexer.column() );
+ }
+ void makeUnexpectedTokenError( Error::Type e ) {
+ makeError( e ); // ### save wrong token...
+ }
+ bool isArgumentToken() const;
+ bool isStringToken() const;
+
+ Error mError;
+ Lexer::Token mToken;
+ QString mTokenValue;
+ Lexer::Impl lexer;
+ ScriptBuilder * mBuilder;
+ };
+
+}
+
+#endif // __KSIEVE_IMPL_PARSER_H__
diff --git a/libksieve/impl/utf8validator.h b/libksieve/impl/utf8validator.h
new file mode 100644
index 000000000..aa10bad2a
--- /dev/null
+++ b/libksieve/impl/utf8validator.h
@@ -0,0 +1,42 @@
+/* -*- c++ -*-
+ impl/utf8validator.h
+
+ This file is part of KSieve,
+ the KDE internet mail/usenet news message filtering library.
+ Copyright (c) 2003 Marc Mutz <mutz@kde.org>
+
+ KSieve is free software; you can redistribute it and/or modify it
+ under the terms of the GNU General Public License, version 2, as
+ published by the Free Software Foundation.
+
+ KSieve 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 this program; if not, write to the Free Software
+ Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
+
+ In addition, as a special exception, the copyright holders give
+ permission to link the code of this program with any edition of
+ the Qt library by Trolltech AS, Norway (or with modified versions
+ of Qt that use the same license as Qt), and distribute linked
+ combinations including the two. You must obey the GNU General
+ Public License in all respects for all of the code used other than
+ Qt. If you modify this file, you may extend this exception to
+ your version of the file, but you are not obligated to do so. If
+ you do not wish to do so, delete this exception statement from
+ your version.
+*/
+
+#ifndef __KSIEVE_UTF8VALIDATOR_H__
+#define __KSIEVE_UTF8VALIDATOR_H__
+
+namespace KSieve {
+
+ extern bool isValidUtf8( const char * s, unsigned int len );
+
+} // namespace KSieve
+
+#endif // __KSIEVE_ERROR_H__
diff --git a/libksieve/ksieve/Makefile.am b/libksieve/ksieve/Makefile.am
new file mode 100644
index 000000000..9d086dd2a
--- /dev/null
+++ b/libksieve/ksieve/Makefile.am
@@ -0,0 +1,8 @@
+# here are header files that are part of the public api:
+ksievedir = $(includedir)/ksieve
+
+ksieve_HEADERS = \
+ error.h \
+ lexer.h \
+ parser.h \
+ scriptbuilder.h
diff --git a/libksieve/ksieve/error.h b/libksieve/ksieve/error.h
new file mode 100644
index 000000000..2dbed32c8
--- /dev/null
+++ b/libksieve/ksieve/error.h
@@ -0,0 +1,139 @@
+/* -*- c++ -*-
+ ksieve/error.h
+
+ This file is part of KSieve,
+ the KDE internet mail/usenet news message filtering library.
+ Copyright (c) 2002-2003 Marc Mutz <mutz@kde.org>
+
+ KSieve is free software; you can redistribute it and/or modify it
+ under the terms of the GNU General Public License, version 2, as
+ published by the Free Software Foundation.
+
+ KSieve 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 this program; if not, write to the Free Software
+ Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
+
+ In addition, as a special exception, the copyright holders give
+ permission to link the code of this program with any edition of
+ the Qt library by Trolltech AS, Norway (or with modified versions
+ of Qt that use the same license as Qt), and distribute linked
+ combinations including the two. You must obey the GNU General
+ Public License in all respects for all of the code used other than
+ Qt. If you modify this file, you may extend this exception to
+ your version of the file, but you are not obligated to do so. If
+ you do not wish to do so, delete this exception statement from
+ your version.
+*/
+
+#ifndef __KSIEVE_ERROR_H__
+#define __KSIEVE_ERROR_H__
+
+#include <qstring.h>
+
+#include <kdepimmacros.h>
+
+#ifdef None // X headers
+# undef None
+#endif
+
+namespace KSieve {
+
+ class KDE_EXPORT Error {
+ public:
+ enum Type {
+ None = 0,
+ Custom,
+ // parse (well-formedness in XML speak) errors:
+ FirstParseError,
+
+ CRWithoutLF = FirstParseError,
+ SlashWithoutAsterisk,
+ IllegalCharacter,
+ UnexpectedCharacter,
+ NoLeadingDigits,
+ NonCWSAfterTextColon,
+
+ NumberOutOfRange,
+ InvalidUTF8,
+
+ UnfinishedBracketComment,
+ PrematureEndOfMultiLine,
+ PrematureEndOfQuotedString,
+ PrematureEndOfStringList,
+ PrematureEndOfTestList,
+ PrematureEndOfBlock,
+ MissingWhitespace,
+ MissingSemicolonOrBlock,
+
+ ExpectedBlockOrSemicolon,
+ ExpectedCommand,
+ ConsecutiveCommasInStringList,
+ ConsecutiveCommasInTestList,
+ MissingCommaInTestList,
+ MissingCommaInStringList,
+ NonStringInStringList,
+ NonCommandInCommandList,
+ NonTestInTestList,
+ LastParseError = NonTestInTestList,
+ // validity errors:
+ FirstValidityError,
+ RequireNotFirst = FirstValidityError, // rfc3028, 3.2
+ RequireMissingForCommand,
+ RequireMissingForTest,
+ RequireMissingForComparator,
+ UnsupportedCommand,
+ UnsupportedTest,
+ UnsupportedComparator,
+ TestNestingTooDeep, // site policy
+ BlockNestingTooDeep, // site policy
+ InvalidArgument,
+ ConflictingArguments, // e.g. rfc3028, 2.7.{1,3}
+ ArgumentsRepeated, // similar to ConflictingArguments, e.g. :is :is
+ CommandOrderingConstraintViolation, // e.g. else w/o if, rfc3028, 3.1
+ LastValidityError = CommandOrderingConstraintViolation,
+ // runtime errors:
+ FirstRuntimeError,
+ IncompatibleActionsRequested = FirstRuntimeError,
+ MailLoopDetected,
+ TooManyActions,
+ LastRuntimeError = TooManyActions
+ };
+
+ static const char * typeToString( Type type );
+
+ Error( Type type=None,
+ const QString & s1=QString::null, const QString & s2=QString::null,
+ int line=-1, int col=-1 )
+ : mType( type ), mLine( line ), mCol( col ),
+ mStringOne( s1 ), mStringTwo( s2 ) {}
+ Error( Type type, int line, int col )
+ : mType( type ), mLine( line ), mCol( col ) {}
+
+ QString asString() const;
+
+ /** So you can write <pre>if( error() )</pre> with e.g. @ref Lexer */
+ operator bool() const {
+ return type() != None;
+ }
+
+ Type type() const { return mType; }
+ int line() const { return mLine; }
+ int column() const { return mCol; }
+ QString firstString() const { return mStringOne; }
+ QString secondString() const { return mStringTwo; }
+
+ protected:
+ Type mType;
+ int mLine;
+ int mCol;
+ QString mStringOne, mStringTwo;
+ };
+
+} // namespace KSieve
+
+#endif // __KSIEVE_ERROR_H__
diff --git a/libksieve/ksieve/lexer.h b/libksieve/ksieve/lexer.h
new file mode 100644
index 000000000..d5bb1fc3b
--- /dev/null
+++ b/libksieve/ksieve/lexer.h
@@ -0,0 +1,108 @@
+/* -*- c++ -*-
+ ksieve/lexer.h
+
+ This file is part of KSieve,
+ the KDE internet mail/usenet news message filtering library.
+ Copyright (c) 2003 Marc Mutz <mutz@kde.org>
+
+ KSieve is free software; you can redistribute it and/or modify it
+ under the terms of the GNU General Public License, version 2, as
+ published by the Free Software Foundation.
+
+ KSieve 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 this program; if not, write to the Free Software
+ Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
+
+ In addition, as a special exception, the copyright holders give
+ permission to link the code of this program with any edition of
+ the Qt library by Trolltech AS, Norway (or with modified versions
+ of Qt that use the same license as Qt), and distribute linked
+ combinations including the two. You must obey the GNU General
+ Public License in all respects for all of the code used other than
+ Qt. If you modify this file, you may extend this exception to
+ your version of the file, but you are not obligated to do so. If
+ you do not wish to do so, delete this exception statement from
+ your version.
+*/
+
+#ifndef __KSIEVE_LEXER_H__
+#define __KSIEVE_LEXER_H__
+
+class QString;
+
+namespace KSieve {
+
+ class Error;
+
+ class Lexer {
+ public:
+ enum Options {
+ IncludeComments = 0,
+ IgnoreComments = 1,
+ IncludeLineFeeds = 0,
+ IgnoreLineFeeds = 2
+ };
+
+ Lexer( const char * scursor, const char * send, int options=0 );
+ ~Lexer();
+
+ /** Return whether comments are returned by @ref
+ nextToken. Default is to not ignore comments. Ignoring them
+ can speed up script parsing a bit, and can be used when the
+ internal representation of the script won't be serialized into
+ string form again (or if you simply want to delete all
+ comments)
+ **/
+ bool ignoreComments() const;
+
+ /** Return whether line feeds are returned by @ref
+ nextToken. Default is to not ignore line feeds. Ignoring them
+ can speed up script parsing a bit, and can be used when the
+ internal representation of the script won't be serialized into
+ string form again.
+ **/
+ bool ignoreLineFeeds() const;
+
+ const Error & error() const;
+
+ bool atEnd() const;
+ int column() const;
+ int line() const;
+
+ enum Token {
+ None = 0,
+ Number, // 1, 100, 1M, 10k, 1G, 2g, 3m
+ Identifier, // atom
+ Tag, // :tag
+ Special, // {} [] () ,;
+ QuotedString, // "foo\"bar" -> foo"bar
+ MultiLineString, // text: \nfoo\n. -> foo
+ HashComment, // # foo
+ BracketComment, // /* foo */
+ LineFeeds // the number of line feeds encountered
+ };
+
+ /** Parse the next token and return it's type. @p result will contain
+ the value of the token. */
+ Token nextToken( QString & result );
+
+ void save();
+ void restore();
+
+ class Impl;
+ private:
+ Impl * i;
+
+ private:
+ const Lexer & operator=( const Lexer & );
+ Lexer( const Lexer & );
+ };
+
+} // namespace KSieve
+
+#endif // __KSIEVE_LEXER_H__
diff --git a/libksieve/ksieve/parser.h b/libksieve/ksieve/parser.h
new file mode 100644
index 000000000..e70e1db4d
--- /dev/null
+++ b/libksieve/ksieve/parser.h
@@ -0,0 +1,72 @@
+/* -*- c++ -*-
+ ksieve/parser.h
+
+ This file is part of KSieve,
+ the KDE internet mail/usenet news message filtering library.
+ Copyright (c) 2002-2003 Marc Mutz <mutz@kde.org>
+
+ KSieve is free software; you can redistribute it and/or modify it
+ under the terms of the GNU General Public License, version 2, as
+ published by the Free Software Foundation.
+
+ KSieve 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 this program; if not, write to the Free Software
+ Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
+
+ In addition, as a special exception, the copyright holders give
+ permission to link the code of this program with any edition of
+ the Qt library by Trolltech AS, Norway (or with modified versions
+ of Qt that use the same license as Qt), and distribute linked
+ combinations including the two. You must obey the GNU General
+ Public License in all respects for all of the code used other than
+ Qt. If you modify this file, you may extend this exception to
+ your version of the file, but you are not obligated to do so. If
+ you do not wish to do so, delete this exception statement from
+ your version.
+*/
+
+#ifndef __KSIEVE_PARSING_H__
+#define __KSIEVE_PARSING_H__
+
+#include <kdepimmacros.h>
+
+class QString;
+
+namespace KSieve {
+
+ class ScriptBuilder;
+ class Error;
+
+ /** @short Parser for the Sieve grammar.
+ @author Marc Mutz <mutz@kde.org>
+ **/
+ class KDE_EXPORT Parser {
+ public:
+
+ Parser( const char * scursor, const char * const send, int options=0 );
+ ~Parser();
+
+ void setScriptBuilder( ScriptBuilder * builder );
+ ScriptBuilder * scriptBuilder() const;
+
+ bool parse();
+
+ const Error & error() const;
+
+ class Impl;
+ private:
+ Impl * i;
+
+ private:
+ const Parser & operator=( const Parser & );
+ Parser( const Parser & );
+ };
+
+} // namespace KSieve
+
+#endif // __KSIEVE_PARSING_H__
diff --git a/libksieve/ksieve/scriptbuilder.h b/libksieve/ksieve/scriptbuilder.h
new file mode 100644
index 000000000..5e0a955bb
--- /dev/null
+++ b/libksieve/ksieve/scriptbuilder.h
@@ -0,0 +1,80 @@
+/* -*- c++ -*-
+ ksieve/interfaces/scriptbuilder.h
+
+ This file is part of KSieve,
+ the KDE internet mail/usenet news message filtering library.
+ Copyright (c) 2002-2003 Marc Mutz <mutz@kde.org>
+
+ KSieve is free software; you can redistribute it and/or modify it
+ under the terms of the GNU General Public License, version 2, as
+ published by the Free Software Foundation.
+
+ KSieve 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 this program; if not, write to the Free Software
+ Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
+
+ In addition, as a special exception, the copyright holders give
+ permission to link the code of this program with any edition of
+ the Qt library by Trolltech AS, Norway (or with modified versions
+ of Qt that use the same license as Qt), and distribute linked
+ combinations including the two. You must obey the GNU General
+ Public License in all respects for all of the code used other than
+ Qt. If you modify this file, you may extend this exception to
+ your version of the file, but you are not obligated to do so. If
+ you do not wish to do so, delete this exception statement from
+ your version.
+*/
+
+#ifndef __KSIEVE_INTERFACES_SCRIPTBUILDER_H__
+#define __KSIEVE_INTERFACES_SCRIPTBUILDER_H__
+
+class QString;
+
+namespace KSieve {
+
+ class Error;
+
+ class ScriptBuilder {
+ public:
+ virtual ~ScriptBuilder() {}
+
+ virtual void taggedArgument( const QString & tag ) = 0;
+ virtual void stringArgument( const QString & string, bool multiLine, const QString & embeddedHashComment ) = 0;
+ virtual void numberArgument( unsigned long number, char quantifier ) = 0;
+
+ virtual void stringListArgumentStart() = 0;
+ virtual void stringListEntry( const QString & string, bool multiLine, const QString & embeddedHashComment ) = 0;
+ virtual void stringListArgumentEnd() = 0;
+
+ virtual void commandStart( const QString & identifier ) = 0;
+ virtual void commandEnd() = 0;
+
+ virtual void testStart( const QString & identifier ) = 0;
+ virtual void testEnd() = 0;
+
+ virtual void testListStart() = 0;
+ virtual void testListEnd() = 0;
+
+ virtual void blockStart() = 0;
+ virtual void blockEnd() = 0;
+
+ /** A hash comment always includes an implicit lineFeed() at it's end. */
+ virtual void hashComment( const QString & comment ) = 0;
+ /** Bracket comments inclde explicit lineFeed()s in their content */
+ virtual void bracketComment( const QString & comment ) = 0;
+
+ virtual void lineFeed() = 0;
+
+ virtual void error( const Error & error ) = 0;
+
+ virtual void finished() = 0;
+ };
+
+} // namespace KSieve
+
+#endif // __KSIEVE_INTERFACES_SCRIPTBUILDER_H__
diff --git a/libksieve/parser/Makefile.am b/libksieve/parser/Makefile.am
new file mode 100644
index 000000000..044d045cf
--- /dev/null
+++ b/libksieve/parser/Makefile.am
@@ -0,0 +1,12 @@
+# final breaks static use:
+# If you feel like "fixing" it, better talk to mutz@kde.org first :)
+KDE_OPTIONS = nofinal
+
+INCLUDES = -I$(top_srcdir)/libksieve $(all_includes)
+
+noinst_LTLIBRARIES = libksieve_parser.la
+
+libksieve_parser_la_SOURCES = utf8validator.cpp lexer.cpp parser.cpp
+libksieve_parser_la_LIBADD = ../shared/libksieve_shared.la
+libksieve_parser_la_LDFLAGS = $(all_libraries) -no-undefined
+
diff --git a/libksieve/parser/lexer.cpp b/libksieve/parser/lexer.cpp
new file mode 100644
index 000000000..d8b76da71
--- /dev/null
+++ b/libksieve/parser/lexer.cpp
@@ -0,0 +1,666 @@
+/* -*- c++ -*-
+ parser/lexer.cpp
+
+ This file is part of KSieve,
+ the KDE internet mail/usenet news message filtering library.
+ Copyright (c) 2002-2003 Marc Mutz <mutz@kde.org>
+
+ KSieve is free software; you can redistribute it and/or modify it
+ under the terms of the GNU General Public License, version 2, as
+ published by the Free Software Foundation.
+
+ KSieve 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 this program; if not, write to the Free Software
+ Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
+
+ In addition, as a special exception, the copyright holders give
+ permission to link the code of this program with any edition of
+ the Qt library by Trolltech AS, Norway (or with modified versions
+ of Qt that use the same license as Qt), and distribute linked
+ combinations including the two. You must obey the GNU General
+ Public License in all respects for all of the code used other than
+ Qt. If you modify this file, you may extend this exception to
+ your version of the file, but you are not obligated to do so. If
+ you do not wish to do so, delete this exception statement from
+ your version.
+*/
+
+#include <config.h>
+
+#include <ksieve/lexer.h>
+#include <impl/lexer.h>
+
+#include <impl/utf8validator.h>
+#include <ksieve/error.h>
+
+#include <qstring.h>
+#include <qstringlist.h>
+#include <qtextcodec.h>
+
+#include <memory> // std::auto_ptr
+
+#include <assert.h>
+#include <ctype.h> // isdigit
+
+#ifdef STR_DIM
+# undef STR_DIM
+#endif
+#define STR_DIM(x) (sizeof(x) - 1)
+
+namespace KSieve {
+
+ //
+ //
+ // Lexer Bridge implementation
+ //
+ //
+
+ Lexer::Lexer( const char * scursor, const char * send, int options )
+ : i( 0 )
+ {
+ i = new Impl( scursor, send, options );
+ }
+
+ Lexer::~Lexer() {
+ delete i; i = 0;
+ }
+
+ bool Lexer::ignoreComments() const {
+ assert( i );
+ return i->ignoreComments();
+ }
+
+ const Error & Lexer::error() const {
+ assert( i );
+ return i->error();
+ }
+
+ bool Lexer::atEnd() const {
+ assert( i );
+ return i->atEnd();
+ }
+
+ int Lexer::column() const {
+ assert( i );
+ return i->column();
+ }
+
+ int Lexer::line() const {
+ assert( i );
+ return i->line();
+ }
+
+ void Lexer::save() {
+ assert( i );
+ i->save();
+ }
+
+ void Lexer::restore() {
+ assert( i );
+ i->restore();
+ }
+
+ Lexer::Token Lexer::nextToken( QString & result ) {
+ assert( i );
+ return i->nextToken( result );
+ }
+
+} // namespace KSieve
+
+
+// none except a-zA-Z0-9_
+static const unsigned char iTextMap[16] = {
+ 0x00, 0x00, 0x00, 0x00, // CTLs: none
+ 0x00, 0x00, 0xFF, 0xC0, // SP ... '?': 0-9
+ 0x7F, 0xFF, 0xFF, 0xE1, // '@' ... '_': A-Z_
+ 0x7F, 0xFF, 0xFF, 0xE0 // '`' ... DEL: a-z
+};
+
+// SP, HT, CR, LF, {}[]();,#/
+// ### exclude '['? Why would one want to write identifier["foo"]?
+static const unsigned char delimMap[16] = {
+ 0x00, 0x64, 0x00, 0x00, // CTLs: CR, HT, LF
+ 0x90, 0xC9, 0x00, 0x10, // SP ... '?': SP, #(),;
+ 0x00, 0x00, 0x00, 0x16, // '@' ... '_': []
+ 0x00, 0x00, 0x00, 0x16 // '`' ... DEL: {}
+};
+
+// All except iText, delim, "*:
+static const unsigned char illegalMap[16] = {
+ 0xFF, 0x9B, 0xFF, 0xFF,
+ 0x4F, 0x16, 0x00, 0x0F,
+ 0x80, 0x00, 0x00, 0x0A,
+ 0x80, 0x00, 0x00, 0x0A
+};
+
+static inline bool isOfSet( const unsigned char map[16], unsigned char ch ) {
+ assert( ch < 128 );
+ return ( map[ ch/8 ] & 0x80 >> ch%8 );
+}
+
+static inline bool isIText( unsigned char ch ) {
+ return ch <= 'z' && isOfSet( iTextMap, ch );
+}
+
+static inline bool isDelim( unsigned char ch ) {
+ return ch <= '}' && isOfSet( delimMap, ch );
+}
+
+static inline bool isIllegal( unsigned char ch ) {
+ return ch >= '~' || isOfSet( illegalMap, ch );
+}
+
+static inline bool is8Bit( signed char ch ) {
+ return ch < 0;
+}
+
+static QString removeCRLF( const QString & s ) {
+ const bool CRLF = s.endsWith( "\r\n" );
+ const bool LF = !CRLF && s.endsWith( "\n" );
+
+ const int e = CRLF ? 2 : LF ? 1 : 0 ; // what to chop off at the end
+
+ return s.left( s.length() - e );
+}
+
+static QString removeDotStuff( const QString & s ) {
+ return s.startsWith( ".." ) ? s.mid( 1 ) : s ;
+}
+
+namespace KSieve {
+
+ //
+ //
+ // Lexer Implementation
+ //
+ //
+
+ Lexer::Impl::Impl( const char * scursor, const char * send, int options )
+ : mState( scursor ? scursor : send ),
+ mEnd( send ? send : scursor ),
+ mIgnoreComments( options & IgnoreComments ),
+ mIgnoreLF( options & IgnoreLineFeeds )
+ {
+ if ( !scursor || !send )
+ assert( atEnd() );
+ }
+
+ Lexer::Token Lexer::Impl::nextToken( QString & result ) {
+ assert( !atEnd() );
+ result = QString::null;
+ //clearErrors();
+
+ const int oldLine = line();
+
+ const bool eatingWSSucceeded = ignoreComments() ? eatCWS() : eatWS() ;
+
+ if ( !ignoreLineFeeds() && oldLine != line() ) {
+ result.setNum( line() - oldLine ); // return number of linefeeds encountered
+ return LineFeeds;
+ }
+
+ if ( !eatingWSSucceeded )
+ return None;
+
+ if ( atEnd() )
+ return None;
+
+ switch ( *mState.cursor ) {
+ case '#': // HashComment
+ assert( !ignoreComments() );
+ ++mState.cursor;
+ if ( !atEnd() )
+ parseHashComment( result, true );
+ return HashComment;
+ case '/': // BracketComment
+ assert( !ignoreComments() );
+ ++mState.cursor; // eat slash
+ if ( atEnd() || *mState.cursor != '*' ) {
+ makeError( Error::SlashWithoutAsterisk );
+ return BracketComment;
+ }
+ ++mState.cursor; // eat asterisk
+ if ( atEnd() ) {
+ makeError( Error::UnfinishedBracketComment );
+ return BracketComment;
+ }
+ parseBracketComment( result, true );
+ return BracketComment;
+ case ':': // Tag
+ ++mState.cursor;
+ if ( atEnd() ) {
+ makeError( Error::UnexpectedCharacter, line(), column() - 1 );
+ return Tag;
+ }
+ if ( !isIText( *mState.cursor ) ) {
+ makeIllegalCharError( *mState.cursor );
+ return Tag;
+ }
+ parseTag( result );
+ return Tag;
+ case '"': // QuotedString
+ ++mState.cursor;
+ parseQuotedString( result );
+ return QuotedString;
+ case '{':
+ case '}':
+ case '[':
+ case ']':
+ case '(':
+ case ')':
+ case ';':
+ case ',': // Special
+ result = *mState.cursor++;
+ return Special;
+ case '0':
+ case '1':
+ case '2':
+ case '3':
+ case '4':
+ case '5':
+ case '6':
+ case '7':
+ case '8':
+ case '9': // Number
+ parseNumber( result );
+ return Number;
+ case 't': // maybe MultiLineString, else Identifier
+ if ( _strnicmp( mState.cursor, "text:", STR_DIM("text:") ) == 0 ) {
+ // MultiLineString
+ mState.cursor += STR_DIM("text:");
+ parseMultiLine( result );
+ // ### FIXME: There can be a hash-comment between "text:"
+ // and CRLF! That should be preserved somehow...
+ return MultiLineString;
+ }
+ // else fall through:
+ default: // Identifier (first must not be 0-9, and can't (caught by Number above))
+ if ( !isIText( *mState.cursor ) ) {
+ makeError( Error::IllegalCharacter );
+ return None;
+ }
+ parseIdentifier( result );
+ return Identifier;
+ }
+ }
+
+ bool Lexer::Impl::eatWS() {
+ while ( !atEnd() )
+ switch ( *mState.cursor ) {
+ case '\r':
+ case '\n':
+ if ( !eatCRLF() )
+ return false;
+ break;
+ case ' ':
+ case '\t':
+ ++mState.cursor;
+ break;
+ default:
+ return true;
+ }
+
+ // at end:
+ return true;
+ }
+
+ bool Lexer::Impl::eatCRLF() {
+ assert( !atEnd() );
+ assert( *mState.cursor == '\n' || *mState.cursor == '\r' );
+
+ if ( *mState.cursor == '\r' ) {
+ ++mState.cursor;
+ if ( atEnd() || *mState.cursor != '\n' ) {
+ // CR w/o LF -> error
+ makeError( Error::CRWithoutLF );
+ return false;
+ } else {
+ // good CRLF
+ newLine();
+ return true;
+ }
+ } else /* *mState.cursor == '\n' */ {
+ // good, LF only
+ newLine();
+ return true;
+ }
+ }
+
+
+ bool Lexer::Impl::parseHashComment( QString & result, bool reallySave ) {
+ // hash-comment := "#" *CHAR-NOT-CRLF CRLF
+
+ // check that the caller plays by the rules:
+ assert( *(mState.cursor-1) == '#' );
+
+ const char * const commentStart = mState.cursor;
+
+ // find next CRLF:
+ while ( !atEnd() ) {
+ if ( *mState.cursor == '\n' || *mState.cursor == '\r' ) break;
+ ++mState.cursor;
+ }
+
+ const char * const commentEnd = mState.cursor - 1;
+
+ if ( commentEnd == commentStart ) return true; // # was last char in script...
+
+ if ( atEnd() || eatCRLF() ) {
+ const int commentLength = commentEnd - commentStart + 1;
+ if ( commentLength > 0 ) {
+ if ( !isValidUtf8( commentStart, commentLength ) ) {
+ makeError( Error::InvalidUTF8 );
+ return false;
+ }
+ if ( reallySave )
+ result += QString::fromUtf8( commentStart, commentLength );
+ }
+ return true;
+ }
+
+ return false;
+ }
+
+ bool Lexer::Impl::parseBracketComment( QString & result, bool reallySave ) {
+ // bracket-comment := "/*" *(CHAR-NOT-STAR / ("*" CHAR-NOT-SLASH )) "*/"
+
+ // check that caller plays by the rules:
+ assert( *(mState.cursor-2) == '/' );
+ assert( *(mState.cursor-1) == '*' );
+
+ const char * const commentStart = mState.cursor;
+ const int commentCol = column() - 2;
+ const int commentLine = line();
+
+ // find next asterisk:
+ do {
+ if ( !skipTo( '*' ) ) {
+ if ( !error() )
+ makeError( Error::UnfinishedBracketComment, commentLine, commentCol );
+ return false;
+ }
+ } while ( !atEnd() && *++mState.cursor != '/' );
+
+ if ( atEnd() ) {
+ makeError( Error::UnfinishedBracketComment, commentLine, commentCol );
+ return false;
+ }
+
+ assert( *mState.cursor == '/' );
+
+ const int commentLength = mState.cursor - commentStart - 1;
+ if ( commentLength > 0 ) {
+ if ( !isValidUtf8( commentStart, commentLength ) ) {
+ makeError( Error::InvalidUTF8 );
+ return false;
+ }
+ if ( reallySave ) {
+ QString tmp = QString::fromUtf8( commentStart, commentLength );
+ result += tmp.remove( '\r' ); // get rid of CR in CRLF pairs
+ }
+ }
+
+ ++mState.cursor; // eat '/'
+ return true;
+ }
+
+ bool Lexer::Impl::parseComment( QString & result, bool reallySave ) {
+ // comment := hash-comment / bracket-comment
+
+ switch( *mState.cursor ) {
+ case '#':
+ ++mState.cursor;
+ return parseHashComment( result, reallySave );
+ case '/':
+ if ( charsLeft() < 2 || mState.cursor[1] != '*' ) {
+ makeError( Error::IllegalCharacter );
+ return false;
+ } else {
+ mState.cursor += 2; // eat "/*"
+ return parseBracketComment( result, reallySave );
+ }
+ default:
+ return false; // don't set an error here - there was no comment
+ }
+ }
+
+ bool Lexer::Impl::eatCWS() {
+ // white-space := 1*(SP / CRLF / HTAB / comment )
+
+ while ( !atEnd() ) {
+ switch( *mState.cursor ) {
+ case ' ':
+ case '\t': // SP / HTAB
+ ++mState.cursor;
+ break;;
+ case '\n':
+ case '\r': // CRLF
+ if ( !eatCRLF() )
+ return false;
+ break;
+ case '#':
+ case '/': // comments
+ {
+ QString dummy;
+ if ( !parseComment( dummy ) )
+ return false;
+ }
+ break;
+ default:
+ return true;
+ }
+ }
+ return true;
+ }
+
+ bool Lexer::Impl::parseIdentifier( QString & result ) {
+ // identifier := (ALPHA / "_") *(ALPHA DIGIT "_")
+
+ assert( isIText( *mState.cursor ) );
+
+ const char * const identifierStart = mState.cursor;
+
+ // first char:
+ if ( isdigit( *mState.cursor ) ) { // no digits for the first
+ makeError( Error::NoLeadingDigits );
+ return false;
+ }
+
+ // rest of identifier chars ( now digits are allowed ):
+ for ( ++mState.cursor ; !atEnd() && isIText( *mState.cursor ) ; ++mState.cursor );
+
+ const int identifierLength = mState.cursor - identifierStart;
+
+ // Can use the fast fromLatin1 here, since identifiers are always
+ // in the us-ascii subset:
+ result += QString::fromLatin1( identifierStart, identifierLength );
+
+ if ( atEnd() || isDelim( *mState.cursor ) )
+ return true;
+
+ makeIllegalCharError( *mState.cursor );
+ return false;
+ }
+
+ bool Lexer::Impl::parseTag( QString & result ) {
+ // tag := ":" identifier
+
+ // check that the caller plays by the rules:
+ assert( *(mState.cursor-1) == ':' );
+ assert( !atEnd() );
+ assert( isIText( *mState.cursor ) );
+
+ return parseIdentifier( result );
+ }
+
+ bool Lexer::Impl::parseNumber( QString & result ) {
+ // number := 1*DIGIT [QUANTIFIER]
+ // QUANTIFIER := "K" / "M" / "G"
+
+ assert( isdigit( *mState.cursor ) );
+
+ while ( !atEnd() && isdigit( *mState.cursor ) )
+ result += *mState.cursor++;
+
+ if ( atEnd() || isDelim( *mState.cursor ) )
+ return true;
+
+ switch ( *mState.cursor ) {
+ case 'G':
+ case 'g':
+ case 'M':
+ case 'm':
+ case 'K':
+ case 'k':
+ result += *mState.cursor++;
+ break;
+ default:
+ makeIllegalCharError();
+ return false;
+ }
+
+ // quantifier found. Check for delimiter:
+ if ( atEnd() || isDelim( *mState.cursor ) )
+ return true;
+ makeIllegalCharError();
+ return false;
+ }
+
+ bool Lexer::Impl::parseMultiLine( QString & result ) {
+ // multi-line := "text:" *(SP / HTAB) (hash-comment / CRLF)
+ // *(multi-line-literal / multi-line-dotstuff)
+ // "." CRLF
+ // multi-line-literal := [CHAR-NOT-DOT *CHAR-NOT-CRLF] CRLF
+ // multi-line-dotstuff := "." 1*CHAR-NOT-CRLF CRLF
+ // ;; A line containing only "." ends the multi-line.
+ // ;; Remove a leading '.' if followed by another '.'.
+
+ assert( _strnicmp( mState.cursor - 5, "text:", STR_DIM("text:") ) == 0 );
+
+ const int mlBeginLine = line();
+ const int mlBeginCol = column() - 5;
+
+ while ( !atEnd() ) {
+ switch ( *mState.cursor ) {
+ case ' ':
+ case '\t':
+ ++mState.cursor;
+ break;
+ case '#':
+ {
+ ++mState.cursor;
+ QString dummy;
+ if ( !parseHashComment( dummy ) )
+ return false;
+ goto MultiLineStart; // break from switch _and_ while
+ }
+ case '\n':
+ case '\r':
+ if ( !eatCRLF() ) return false;
+ goto MultiLineStart; // break from switch _and_ while
+ default:
+ makeError( Error::NonCWSAfterTextColon );
+ return false;
+ }
+ }
+
+ MultiLineStart:
+ if ( atEnd() ) {
+ makeError( Error::PrematureEndOfMultiLine, mlBeginLine, mlBeginCol );
+ return false;
+ }
+
+ // Now, collect the single lines until one with only a single dot is found:
+ QStringList lines;
+ while ( !atEnd() ) {
+ const char * const oldBeginOfLine = beginOfLine();
+ if ( !skipToCRLF() )
+ return false;
+ const int lineLength = mState.cursor - oldBeginOfLine;
+ if ( lineLength > 0 ) {
+ if ( !isValidUtf8( oldBeginOfLine, lineLength ) ) {
+ makeError( Error::InvalidUTF8 );
+ return false;
+ }
+ const QString line = removeCRLF( QString::fromUtf8( oldBeginOfLine, lineLength ) );
+ lines.push_back( removeDotStuff( line ) );
+ if ( line == "." )
+ break;
+ } else {
+ lines.push_back( QString::null );
+ }
+ }
+
+ if ( lines.back() != "." ) {
+ makeError( Error::PrematureEndOfMultiLine, mlBeginLine, mlBeginCol );
+ return false;
+ }
+
+ assert( !lines.empty() );
+ lines.erase( --lines.end() ); // don't include the lone dot.
+ result = lines.join("\n");
+ return true;
+ }
+
+ bool Lexer::Impl::parseQuotedString( QString & result ) {
+ // quoted-string := DQUOTE *CHAR DQUOTE
+
+ // check that caller plays by the rules:
+ assert( *(mState.cursor-1) == '"' );
+
+ const int qsBeginCol = column() - 1;
+ const int qsBeginLine = line();
+
+ const QTextCodec * const codec = QTextCodec::codecForMib( 106 ); // UTF-8
+ assert( codec );
+ const std::auto_ptr<QTextDecoder> dec( codec->makeDecoder() );
+ assert( dec.get() );
+
+ while ( !atEnd() )
+ switch ( *mState.cursor ) {
+ case '"':
+ ++mState.cursor;
+ return true;
+ case '\r':
+ case '\n':
+ if ( !eatCRLF() )
+ return false;
+ result += '\n';
+ break;
+ case '\\':
+ ++mState.cursor;
+ if ( atEnd() )
+ break;
+ // else fall through:
+ default:
+ if ( !is8Bit( *mState.cursor ) )
+ result += *mState.cursor++;
+ else { // probably UTF-8
+ const char * const eightBitBegin = mState.cursor;
+ skipTo8BitEnd();
+ const int eightBitLen = mState.cursor - eightBitBegin;
+ assert( eightBitLen > 0 );
+ if ( isValidUtf8( eightBitBegin, eightBitLen ) )
+ result += dec->toUnicode( eightBitBegin, eightBitLen );
+ else {
+ assert( column() >= eightBitLen );
+ makeError( Error::InvalidUTF8, line(), column() - eightBitLen );
+ return false;
+ }
+ }
+ }
+
+ makeError( Error::PrematureEndOfQuotedString, qsBeginLine, qsBeginCol );
+ return false;
+ }
+
+ void Lexer::Impl::makeIllegalCharError( char ch ) {
+ makeError( isIllegal( ch ) ? Error::IllegalCharacter : Error::UnexpectedCharacter );
+ }
+
+} // namespace KSieve
diff --git a/libksieve/parser/parser.cpp b/libksieve/parser/parser.cpp
new file mode 100644
index 000000000..8c2db050e
--- /dev/null
+++ b/libksieve/parser/parser.cpp
@@ -0,0 +1,651 @@
+/* -*- c++ -*-
+ parser/parser.cpp
+
+ This file is part of KSieve,
+ the KDE internet mail/usenet news message filtering library.
+ Copyright (c) 2002-2003 Marc Mutz <mutz@kde.org>
+
+ KSieve is free software; you can redistribute it and/or modify it
+ under the terms of the GNU General Public License, version 2, as
+ published by the Free Software Foundation.
+
+ KSieve 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 this program; if not, write to the Free Software
+ Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
+
+ In addition, as a special exception, the copyright holders give
+ permission to link the code of this program with any edition of
+ the Qt library by Trolltech AS, Norway (or with modified versions
+ of Qt that use the same license as Qt), and distribute linked
+ combinations including the two. You must obey the GNU General
+ Public License in all respects for all of the code used other than
+ Qt. If you modify this file, you may extend this exception to
+ your version of the file, but you are not obligated to do so. If
+ you do not wish to do so, delete this exception statement from
+ your version.
+*/
+
+#include <config.h>
+
+#include <ksieve/parser.h>
+#include <impl/parser.h>
+
+#include <ksieve/error.h>
+
+#include <qstring.h>
+
+#include <assert.h>
+#include <limits.h> // ULONG_MAX
+#include <ctype.h> // isdigit
+
+namespace KSieve {
+
+ //
+ //
+ // Parser Bridge implementation
+ //
+ //
+
+ Parser::Parser( const char * scursor, const char * const send, int options )
+ : i( 0 )
+ {
+ i = new Impl( scursor, send, options );
+ }
+
+ Parser::~Parser() {
+ delete i; i = 0;
+ }
+
+ void Parser::setScriptBuilder( ScriptBuilder * builder ) {
+ assert( i );
+ i->mBuilder = builder;
+ }
+
+ ScriptBuilder * Parser::scriptBuilder() const {
+ assert( i );
+ return i->mBuilder;
+ }
+
+ const Error & Parser::error() const {
+ assert( i );
+ return i->error();
+ }
+
+ bool Parser::parse() {
+ assert( i );
+ return i->parse();
+ }
+
+}
+
+static inline unsigned long factorForQuantifier( char ch ) {
+ switch ( ch ) {
+ case 'g':
+ case 'G':
+ return 1024*1024*1024;
+ case 'm':
+ case 'M':
+ return 1024*1024;
+ case 'k':
+ case 'K':
+ return 1024;
+ default:
+ assert( 0 ); // lexer should prohibit this
+ return 1; // make compiler happy
+ }
+}
+
+static inline bool willOverflowULong( unsigned long result, unsigned long add ) {
+ static const unsigned long maxULongByTen = (unsigned long)(ULONG_MAX / 10.0) ;
+ return result > maxULongByTen || ULONG_MAX - 10 * result < add ;
+}
+
+namespace KSieve {
+
+ //
+ //
+ // Parser Implementation
+ //
+ //
+
+ Parser::Impl::Impl( const char * scursor, const char * const send, int options )
+ : mToken( Lexer::None ),
+ lexer( scursor, send, options ),
+ mBuilder( 0 )
+ {
+
+ }
+
+ bool Parser::Impl::isStringToken() const {
+ return token() == Lexer::QuotedString ||
+ token() == Lexer::MultiLineString ;
+ }
+
+
+ bool Parser::Impl::isArgumentToken() const {
+ return isStringToken() ||
+ token() == Lexer::Number ||
+ token() == Lexer::Tag ||
+ token() == Lexer::Special && mTokenValue == "[" ;
+ }
+
+ bool Parser::Impl::obtainToken() {
+ while ( !mToken && !lexer.atEnd() && !lexer.error() ) {
+ mToken = lexer.nextToken( mTokenValue );
+ if ( lexer.error() )
+ break;
+ // comments and line feeds are semantically invisible and may
+ // appear anywhere, so we handle them here centrally:
+ switch ( token() ) {
+ case Lexer::HashComment:
+ if ( scriptBuilder() )
+ scriptBuilder()->hashComment( tokenValue() );
+ consumeToken();
+ break;
+ case Lexer::BracketComment:
+ if ( scriptBuilder() )
+ scriptBuilder()->bracketComment( tokenValue() );
+ consumeToken();
+ break;
+ case Lexer::LineFeeds:
+ for ( unsigned int i = 0, end = tokenValue().toUInt() ; i < end ; ++i )
+ if ( scriptBuilder() ) // better check every iteration, b/c
+ // we call out to ScriptBuilder,
+ // where nasty things might happen!
+ scriptBuilder()->lineFeed();
+ consumeToken();
+ break;
+ default: ; // make compiler happy
+ }
+ }
+ if ( lexer.error() && scriptBuilder() )
+ scriptBuilder()->error( lexer.error() );
+ return !lexer.error();
+ }
+
+ bool Parser::Impl::parse() {
+ // this is the entry point: START := command-list
+ if ( !parseCommandList() )
+ return false;
+ if ( !atEnd() ) {
+ makeUnexpectedTokenError( Error::ExpectedCommand );
+ return false;
+ }
+ if ( scriptBuilder() )
+ scriptBuilder()->finished();
+ return true;
+ }
+
+
+ bool Parser::Impl::parseCommandList() {
+ // our ABNF:
+ // command-list := *comand
+
+ while ( !atEnd() ) {
+ if ( !obtainToken() )
+ return false;
+ if ( token() == Lexer::None )
+ continue;
+ if ( token() != Lexer::Identifier )
+ return true;
+ if ( !parseCommand() ) {
+ assert( error() );
+ return false;
+ }
+ }
+ return true;
+ }
+
+
+ bool Parser::Impl::parseCommand() {
+ // command := identifier arguments ( ";" / block )
+ // arguments := *argument [ test / test-list ]
+ // block := "{" *command "}"
+ // our ABNF:
+ // block := "{" [ command-list ] "}"
+
+ if ( atEnd() )
+ return false;
+
+ //
+ // identifier
+ //
+
+ if ( !obtainToken() || token() != Lexer::Identifier )
+ return false;
+
+ if ( scriptBuilder() )
+ scriptBuilder()->commandStart( tokenValue() );
+ consumeToken();
+
+ //
+ // *argument
+ //
+
+ if ( !obtainToken() )
+ return false;
+
+ if ( atEnd() ) {
+ makeError( Error::MissingSemicolonOrBlock );
+ return false;
+ }
+
+ if ( isArgumentToken() && !parseArgumentList() ) {
+ assert( error() );
+ return false;
+ }
+
+ //
+ // test / test-list
+ //
+
+ if ( !obtainToken() )
+ return false;
+
+ if ( atEnd() ) {
+ makeError( Error::MissingSemicolonOrBlock );
+ return false;
+ }
+
+ if ( token() == Lexer::Special && tokenValue() == "(" ) { // test-list
+ if ( !parseTestList() ) {
+ assert( error() );
+ return false;
+ }
+ } else if ( token() == Lexer::Identifier ) { // should be test:
+ if ( !parseTest() ) {
+ assert( error() );
+ return false;
+ }
+ }
+
+ //
+ // ";" / block
+ //
+
+ if ( !obtainToken() )
+ return false;
+
+ if ( atEnd() ) {
+ makeError( Error::MissingSemicolonOrBlock );
+ return false;
+ }
+
+ if ( token() != Lexer::Special ) {
+ makeUnexpectedTokenError( Error::ExpectedBlockOrSemicolon );
+ return false;
+ }
+
+ if ( tokenValue() == ";" )
+ consumeToken();
+ else if ( tokenValue() == "{" ) { // block
+ if ( !parseBlock() )
+ return false; // it's an error since we saw '{'
+ } else {
+ makeError( Error::MissingSemicolonOrBlock );
+ return false;
+ }
+
+ if ( scriptBuilder() )
+ scriptBuilder()->commandEnd();
+ return true;
+ }
+
+
+ bool Parser::Impl::parseArgumentList() {
+ // our ABNF:
+ // argument-list := *argument
+
+ while ( !atEnd() ) {
+ if ( !obtainToken() )
+ return false;
+ if ( !isArgumentToken() )
+ return true;
+ if ( !parseArgument() )
+ return !error();
+ }
+ return true;
+ }
+
+
+ bool Parser::Impl::parseArgument() {
+ // argument := string-list / number / tag
+
+ if ( !obtainToken() || atEnd() )
+ return false;
+
+ if ( token() == Lexer::Number ) {
+ if ( !parseNumber() ) {
+ assert( error() );
+ return false;
+ }
+ return true;
+ } else if ( token() == Lexer::Tag ) {
+ if ( scriptBuilder() )
+ scriptBuilder()->taggedArgument( tokenValue() );
+ consumeToken();
+ return true;
+ } else if ( isStringToken() ) {
+ if ( scriptBuilder() )
+ scriptBuilder()->stringArgument( tokenValue(), token() == Lexer::MultiLineString, QString::null );
+ consumeToken();
+ return true;
+ } else if ( token() == Lexer::Special && tokenValue() == "[" ) {
+ if ( !parseStringList() ) {
+ assert( error() );
+ return false;
+ }
+ return true;
+ }
+
+ return false;
+ }
+
+
+ bool Parser::Impl::parseTestList() {
+ // test-list := "(" test *("," test) ")"
+
+ if ( !obtainToken() || atEnd() )
+ return false;
+
+ if ( token() != Lexer::Special || tokenValue() != "(" )
+ return false;
+ if ( scriptBuilder() )
+ scriptBuilder()->testListStart();
+ consumeToken();
+
+ // generic while/switch construct for comma-separated lists. See
+ // parseStringList() for another one. Any fix here is like to apply there, too.
+ bool lastWasComma = true;
+ while ( !atEnd() ) {
+ if ( !obtainToken() )
+ return false;
+
+ switch ( token() ) {
+ case Lexer::None:
+ break;
+ case Lexer::Special:
+ assert( tokenValue().length() == 1 );
+ assert( tokenValue()[0].latin1() );
+ switch ( tokenValue()[0].latin1() ) {
+ case ')':
+ consumeToken();
+ if ( lastWasComma ) {
+ makeError( Error::ConsecutiveCommasInTestList );
+ return false;
+ }
+ if ( scriptBuilder() )
+ scriptBuilder()->testListEnd();
+ return true;
+ case ',':
+ consumeToken();
+ if( lastWasComma ) {
+ makeError( Error::ConsecutiveCommasInTestList );
+ return false;
+ }
+ lastWasComma = true;
+ break;
+ default:
+ makeError( Error::NonStringInStringList );
+ return false;
+ }
+ break;
+
+ case Lexer::Identifier:
+ if ( !lastWasComma ) {
+ makeError( Error::MissingCommaInTestList );
+ return false;
+ } else {
+ lastWasComma = false;
+ if ( !parseTest() ) {
+ assert( error() );
+ return false;
+ }
+ }
+ break;
+
+ default:
+ makeUnexpectedTokenError( Error::NonTestInTestList );
+ return false;
+ }
+ }
+
+ makeError( Error::PrematureEndOfTestList );
+ return false;
+ }
+
+
+ bool Parser::Impl::parseTest() {
+ // test := identifier arguments
+ // arguments := *argument [ test / test-list ]
+
+ //
+ // identifier
+ //
+
+ if ( !obtainToken() || atEnd() )
+ return false;
+
+ if ( token() != Lexer::Identifier )
+ return false;
+
+ if ( scriptBuilder() )
+ scriptBuilder()->testStart( tokenValue() );
+ consumeToken();
+
+ //
+ // *argument
+ //
+
+ if ( !obtainToken() )
+ return false;
+
+ if ( atEnd() ) // a test w/o args
+ goto TestEnd;
+
+ if ( isArgumentToken() && !parseArgumentList() ) {
+ assert( error() );
+ return false;
+ }
+
+ //
+ // test / test-list
+ //
+
+ if ( !obtainToken() )
+ return false;
+
+ if ( atEnd() ) // a test w/o nested tests
+ goto TestEnd;
+
+ if ( token() == Lexer::Special && tokenValue() == "(" ) { // test-list
+ if ( !parseTestList() ) {
+ assert( error() );
+ return false;
+ }
+ } else if ( token() == Lexer::Identifier ) { // should be test:
+ if ( !parseTest() ) {
+ assert( error() );
+ return false;
+ }
+ }
+
+ TestEnd:
+ if ( scriptBuilder() )
+ scriptBuilder()->testEnd();
+ return true;
+ }
+
+
+ bool Parser::Impl::parseBlock() {
+ // our ABNF:
+ // block := "{" [ command-list ] "}"
+
+ if ( !obtainToken() || atEnd() )
+ return false;
+
+ if ( token() != Lexer::Special || tokenValue() != "{" )
+ return false;
+ if ( scriptBuilder() )
+ scriptBuilder()->blockStart();
+ consumeToken();
+
+ if ( !obtainToken() )
+ return false;
+
+ if ( atEnd() ) {
+ makeError( Error::PrematureEndOfBlock );
+ return false;
+ }
+
+ if ( token() == Lexer::Identifier ) {
+ if ( !parseCommandList() ) {
+ assert( error() );
+ return false;
+ }
+ }
+
+ if ( !obtainToken() )
+ return false;
+
+ if ( atEnd() ) {
+ makeError( Error::PrematureEndOfBlock );
+ return false;
+ }
+
+ if ( token() != Lexer::Special || tokenValue() != "}" ) {
+ makeError( Error::NonCommandInCommandList );
+ return false;
+ }
+ if ( scriptBuilder() )
+ scriptBuilder()->blockEnd();
+ consumeToken();
+ return true;
+ }
+
+ bool Parser::Impl::parseStringList() {
+ // string-list := "[" string *("," string) "]" / string
+ // ;; if there is only a single string, the brackets are optional
+ //
+ // However, since strings are already handled separately from
+ // string lists in parseArgument(), our ABNF is modified to:
+ // string-list := "[" string *("," string) "]"
+
+ if ( !obtainToken() || atEnd() )
+ return false;
+
+ if ( token() != Lexer::Special || tokenValue() != "[" )
+ return false;
+
+ if ( scriptBuilder() )
+ scriptBuilder()->stringListArgumentStart();
+ consumeToken();
+
+ // generic while/switch construct for comma-separated lists. See
+ // parseTestList() for another one. Any fix here is like to apply there, too.
+ bool lastWasComma = true;
+ while ( !atEnd() ) {
+ if ( !obtainToken() )
+ return false;
+
+ switch ( token() ) {
+ case Lexer::None:
+ break;
+ case Lexer::Special:
+ assert( tokenValue().length() == 1 );
+ switch ( tokenValue()[0].latin1() ) {
+ case ']':
+ consumeToken();
+ if ( lastWasComma ) {
+ makeError( Error::ConsecutiveCommasInStringList );
+ return false;
+ }
+ if ( scriptBuilder() )
+ scriptBuilder()->stringListArgumentEnd();
+ return true;
+ case ',':
+ consumeToken();
+ if ( lastWasComma ) {
+ makeError( Error::ConsecutiveCommasInStringList );
+ return false;
+ }
+ lastWasComma = true;
+ break;
+ default:
+ makeError( Error::NonStringInStringList );
+ return false;
+ }
+ break;
+
+ case Lexer::QuotedString:
+ case Lexer::MultiLineString:
+ if ( !lastWasComma ) {
+ makeError( Error::MissingCommaInStringList );
+ return false;
+ }
+ lastWasComma = false;
+ if ( scriptBuilder() )
+ scriptBuilder()->stringListEntry( tokenValue(), token() == Lexer::MultiLineString, QString::null );
+ consumeToken();
+ break;
+
+ default:
+ makeError( Error::NonStringInStringList );
+ return false;
+ }
+ }
+
+ makeError( Error::PrematureEndOfStringList );
+ return false;
+ }
+
+ bool Parser::Impl::parseNumber() {
+ // The lexer returns the number including the quantifier as a
+ // single token value. Here, we split is an check that the number
+ // is not out of range:
+
+ if ( !obtainToken() || atEnd() )
+ return false;
+
+ if ( token() != Lexer::Number )
+ return false;
+
+ // number:
+ unsigned long result = 0;
+ unsigned int i = 0;
+ const QCString s = tokenValue().latin1();
+ for ( const unsigned int len = s.length() ; i < len && isdigit( s[i] ) ; ++i ) {
+ const unsigned long digitValue = s[i] - '0' ;
+ if ( willOverflowULong( result, digitValue ) ) {
+ makeError( Error::NumberOutOfRange );
+ return false;
+ } else {
+ result *= 10 ; result += digitValue ;
+ }
+ }
+
+ // optional quantifier:
+ char quantifier = '\0';
+ if ( i < s.length() ) {
+ assert( i + 1 == s.length() );
+ quantifier = s[i];
+ const unsigned long factor = factorForQuantifier( quantifier );
+ if ( result > double(ULONG_MAX) / double(factor) ) {
+ makeError( Error::NumberOutOfRange );
+ return false;
+ }
+ result *= factor;
+ }
+
+ if ( scriptBuilder() )
+ scriptBuilder()->numberArgument( result, quantifier );
+ consumeToken();
+ return true;
+ }
+
+} // namespace KSieve
diff --git a/libksieve/parser/utf8validator.cpp b/libksieve/parser/utf8validator.cpp
new file mode 100644
index 000000000..248a1f5e9
--- /dev/null
+++ b/libksieve/parser/utf8validator.cpp
@@ -0,0 +1,141 @@
+/* -*- c++ -*-
+ utf8validator.cpp
+
+ This file is part of KSieve,
+ the KDE internet mail/usenet news message filtering library.
+ Copyright (c) 2003 Marc Mutz <mutz@kde.org>
+
+ KSieve is free software; you can redistribute it and/or modify it
+ under the terms of the GNU General Public License, version 2, as
+ published by the Free Software Foundation.
+
+ KSieve 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 this program; if not, write to the Free Software
+ Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
+
+ In addition, as a special exception, the copyright holders give
+ permission to link the code of this program with any edition of
+ the Qt library by Trolltech AS, Norway (or with modified versions
+ of Qt that use the same license as Qt), and distribute linked
+ combinations including the two. You must obey the GNU General
+ Public License in all respects for all of the code used other than
+ Qt. If you modify this file, you may extend this exception to
+ your version of the file, but you are not obligated to do so. If
+ you do not wish to do so, delete this exception statement from
+ your version.
+*/
+
+#include <impl/utf8validator.h>
+
+#include <qglobal.h>
+#include <qcstring.h>
+
+static inline bool is8Bit( signed char ch ) {
+ return ch < 0;
+}
+
+static inline bool isUtf8TupelIndicator( unsigned char ch ) {
+ return (ch & 0xE0) == 0xC0; // 110x xxxx
+}
+
+static inline bool isUtf8OverlongTupel( unsigned char ch ) {
+ return (ch & 0xFE) == 0xC0;
+}
+
+static inline bool isUtf8TripleIndicator( unsigned char ch ) {
+ return (ch & 0xF0) == 0xE0; // 1110 xxxx
+}
+
+static inline bool isUtf8OverlongTriple( unsigned char ch1, unsigned char ch2 ) {
+ return (ch1 & 0xFF) == 0xE0 && (ch2 & 0xE0) == 0x80 ;
+}
+
+static inline bool isUtf8QuartetIndicator( unsigned char ch ) {
+ return (ch & 0xF8) == 0xF0; // 1111 0xxx
+}
+
+static inline bool isUtf8OverlongQuartet( unsigned char ch1, unsigned char ch2 ) {
+ return (ch1 & 0xFF) == 0xF0 && (ch2 & 0xF0) == 0x80 ;
+}
+
+static inline bool isUtf8QuintetIndicator( unsigned char ch ) {
+ return (ch & 0xFC) == 0xF8; // 1111 10xx
+}
+
+static inline bool isUtf8OverlongQuintet( unsigned char ch1, unsigned char ch2 ) {
+ return (ch1 & 0xFF) == 0xF8 && (ch2 & 0xF8) == 0x80 ;
+}
+
+static inline bool isUtf8SextetIndicator( unsigned char ch ) {
+ return (ch & 0xFE) == 0xFC; // 1111 110x
+}
+
+static inline bool isUtf8OverlongSextet( unsigned char ch1, unsigned char ch2 ) {
+ return (ch1 & 0xFF) == 0xFC && (ch2 & 0xFC) == 0x80 ;
+}
+
+static inline bool isUtf8Continuation( unsigned char ch ) {
+ return (ch & 0xC0) == 0x80;
+}
+
+bool KSieve::isValidUtf8( const char * s, unsigned int len ) {
+ for ( unsigned int i = 0 ; i < len ; ++i ) {
+ const unsigned char ch = s[i];
+ if ( !is8Bit( ch ) )
+ continue;
+ if ( isUtf8TupelIndicator( ch ) ) {
+ if ( len - i < 1 ) // too short
+ return false;
+ if ( isUtf8OverlongTupel( ch ) ) // not minimally encoded
+ return false;
+ if ( !isUtf8Continuation( s[i+1] ) ) // not followed by 10xx xxxx
+ return false;
+ i += 1;
+ } else if ( isUtf8TripleIndicator( ch ) ) {
+ if ( len - i < 2 ) // too short
+ return false;
+ if ( isUtf8OverlongTriple( ch, s[i+1] ) ) // not minimally encoded
+ return false;
+ if ( !isUtf8Continuation( s[i+2] ) ) // not followed by 10xx xxxx
+ return false;
+ i += 2;
+ } else if ( isUtf8QuartetIndicator( ch ) ) {
+ if ( len - i < 3 ) // too short
+ return false;
+ if ( isUtf8OverlongQuartet( ch, s[i+1] ) ) // not minimally encoded
+ return false;
+ if ( !isUtf8Continuation( s[i+2] ) ||
+ !isUtf8Continuation( s[i+3] ) ) // not followed by 2x 10xx xxxx
+ return false;
+ i += 3;
+ } else if ( isUtf8QuintetIndicator( ch ) ) {
+ if ( len - i < 4 ) // too short
+ return false;
+ if ( isUtf8OverlongQuintet( ch, s[i+1] ) ) // not minimally encoded
+ return false;
+ if ( !isUtf8Continuation( s[i+2] ) ||
+ !isUtf8Continuation( s[i+3] ) ||
+ !isUtf8Continuation( s[i+4] ) ) // not followed by 3x 10xx xxxx
+ return false;
+ i += 4;
+ } else if ( isUtf8SextetIndicator( ch ) ) {
+ if ( len - i < 5 ) // too short
+ return false;
+ if ( isUtf8OverlongSextet( ch, s[i+1] ) ) // not minimally encoded
+ return false;
+ if ( !isUtf8Continuation( s[i+2] ) ||
+ !isUtf8Continuation( s[i+3] ) ||
+ !isUtf8Continuation( s[i+4] ) ||
+ !isUtf8Continuation( s[i+5] ) ) // not followed by 4x 10xx xxxx
+ return false;
+ i += 5;
+ } else
+ return false;
+ }
+ return true;
+}
diff --git a/libksieve/shared/Makefile.am b/libksieve/shared/Makefile.am
new file mode 100644
index 000000000..a2999f687
--- /dev/null
+++ b/libksieve/shared/Makefile.am
@@ -0,0 +1,8 @@
+INCLUDES = -I$(top_srcdir)/libksieve $(all_includes)
+
+noinst_LTLIBRARIES = libksieve_shared.la
+
+libksieve_shared_la_SOURCES = error.cpp
+libksieve_shared_la_LIBADD = $(LIB_KDECORE)
+libksieve_shared_la_LDFLAGS = $(all_libraries) -no-undefined
+
diff --git a/libksieve/shared/error.cpp b/libksieve/shared/error.cpp
new file mode 100644
index 000000000..e53b0c252
--- /dev/null
+++ b/libksieve/shared/error.cpp
@@ -0,0 +1,247 @@
+/* -*- c++ -*-
+ error.cpp
+
+ This file is part of KSieve,
+ the KDE internet mail/usenet news message filtering library.
+ Copyright (c) 2002-2003 Marc Mutz <mutz@kde.org>
+
+ KSieve is free software; you can redistribute it and/or modify it
+ under the terms of the GNU General Public License, version 2, as
+ published by the Free Software Foundation.
+
+ KSieve 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 this program; if not, write to the Free Software
+ Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
+
+ In addition, as a special exception, the copyright holders give
+ permission to link the code of this program with any edition of
+ the Qt library by Trolltech AS, Norway (or with modified versions
+ of Qt that use the same license as Qt), and distribute linked
+ combinations including the two. You must obey the GNU General
+ Public License in all respects for all of the code used other than
+ Qt. If you modify this file, you may extend this exception to
+ your version of the file, but you are not obligated to do so. If
+ you do not wish to do so, delete this exception statement from
+ your version.
+*/
+
+#include <config.h>
+
+#include <ksieve/error.h>
+
+#include <klocale.h> // i18n
+
+#include <climits> // UINT_MAX
+
+namespace KSieve {
+
+ const char * Error::typeToString( Type t ) {
+ switch ( t ) {
+#define CASE(x) case x: return #x
+ CASE( None );
+ CASE( Custom );
+
+ CASE( CRWithoutLF );
+ CASE( SlashWithoutAsterisk );
+ CASE( IllegalCharacter );
+ CASE( UnexpectedCharacter );
+ CASE( NoLeadingDigits );
+ CASE( NonCWSAfterTextColon );
+
+ CASE( NumberOutOfRange );
+ CASE( InvalidUTF8 );
+
+ CASE( UnfinishedBracketComment );
+ CASE( PrematureEndOfMultiLine );
+ CASE( PrematureEndOfQuotedString );
+ CASE( PrematureEndOfStringList );
+ CASE( PrematureEndOfTestList );
+ CASE( PrematureEndOfBlock );
+ CASE( MissingWhitespace );
+ CASE( MissingSemicolonOrBlock );
+
+ CASE( ExpectedBlockOrSemicolon );
+ CASE( ExpectedCommand );
+ CASE( ConsecutiveCommasInStringList );
+ CASE( ConsecutiveCommasInTestList );
+ CASE( MissingCommaInTestList );
+ CASE( MissingCommaInStringList );
+ CASE( NonStringInStringList );
+ CASE( NonCommandInCommandList );
+ CASE( NonTestInTestList );
+
+ CASE( RequireNotFirst );
+ CASE( RequireMissingForCommand );
+ CASE( RequireMissingForTest );
+ CASE( RequireMissingForComparator );
+ CASE( UnsupportedCommand );
+ CASE( UnsupportedTest );
+ CASE( UnsupportedComparator );
+ CASE( TestNestingTooDeep );
+ CASE( BlockNestingTooDeep );
+ CASE( InvalidArgument );
+ CASE( ConflictingArguments );
+ CASE( ArgumentsRepeated );
+ CASE( CommandOrderingConstraintViolation );
+
+ CASE( IncompatibleActionsRequested );
+ CASE( MailLoopDetected );
+ CASE( TooManyActions );
+#undef CASE
+ default:
+ return "<unknown>";
+ }
+ }
+
+ QString Error::asString() const {
+
+ QString err;
+ switch( type() ) {
+ case None:
+ return QString::null;
+ case Custom:
+ return mStringOne;
+
+ // Parse errors:
+ case CRWithoutLF:
+ err = i18n("Parse error: Carriage Return (CR) without Line Feed (LF)");
+ break;
+ case SlashWithoutAsterisk:
+ err = i18n("Parse error: Unquoted Slash ('/') without Asterisk ('*'). "
+ "Broken Comment?");
+ break;
+ case IllegalCharacter:
+ err = i18n("Parse error: Illegal Character");
+ break;
+ case UnexpectedCharacter:
+ err = i18n("Parse error: Unexpected Character, probably a missing space?");
+ break;
+ case NoLeadingDigits:
+ err = i18n("Parse error: Tag Name has leading Digits");
+ break;
+ case NonCWSAfterTextColon:
+ err = i18n("Parse error: Only whitespace and #comments may "
+ "follow \"text:\" on the same line");
+ break;
+ case NumberOutOfRange:
+ err = i18n("Parse error: Number out of Range (must be smaller than %1)").arg(UINT_MAX);
+ break;
+ case InvalidUTF8:
+ err = i18n("Parse error: Invalid UTF-8 sequence");
+ break;
+ case PrematureEndOfMultiLine:
+ err = i18n("Parse error: Premature end of Multiline String (did you forget the '.'?)");
+ break;
+ case PrematureEndOfQuotedString:
+ err = i18n("Parse error: Premature end of Quoted String (missing closing '\"')");
+ break;
+ case PrematureEndOfStringList:
+ err = i18n("Parse error: Premature end of String List (missing closing ']')");
+ break;
+ case PrematureEndOfTestList:
+ err = i18n("Parse error: Premature end of Test List (missing closing ')')");
+ break;
+ case PrematureEndOfBlock:
+ err = i18n("Parse error: Premature end of Block (missing closing '}')");
+ break;
+ case MissingWhitespace:
+ err = i18n("Parse error: Missing Whitespace");
+ break;
+ case MissingSemicolonOrBlock:
+ err = i18n("Parse error: Missing ';' or Block");
+ break;
+ case ExpectedBlockOrSemicolon:
+ err = i18n("Parse error: Expected ';' or '{', got something else");
+ break;
+ case ExpectedCommand:
+ err = i18n("Parse error: Expected Command, got something else");
+ break;
+ case ConsecutiveCommasInStringList:
+ err = i18n("Parse error: Trailing, Leading or Duplicate Commas in String List");
+ break;
+ case ConsecutiveCommasInTestList:
+ err = i18n("Parse error: Trailing, Leading or Duplicate Commas in Test List");
+ break;
+ case MissingCommaInStringList:
+ err = i18n("Parse error: Missing ',' between Strings in String List");
+ break;
+ case MissingCommaInTestList:
+ err = i18n("Parse error: Missing ',' between Tests in Test List");
+ break;
+ case NonCommandInCommandList:
+ err = i18n("Parse error: Expected Command, got something else");
+ break;
+ case NonStringInStringList:
+ err = i18n("Parse error: Only Strings allowed in String Lists");
+ break;
+ case NonTestInTestList:
+ err = i18n("Parse error: Only Tests allowed in Test Lists");
+ break;
+
+ // validity errors:
+ case RequireNotFirst:
+ err = i18n("\"require\" must be first command");
+ break;
+ case RequireMissingForCommand:
+ err = i18n("\"require\" missing for command \"%1\"").arg(mStringOne);
+ break;
+ case RequireMissingForTest:
+ err = i18n("\"require\" missing for test \"%1\"").arg(mStringOne);
+ break;
+ case RequireMissingForComparator:
+ err = i18n("\"require\" missing for comparator \"%1\"").arg(mStringOne);
+ break;
+ case UnsupportedCommand:
+ err = i18n("Command \"%1\" not supported").arg(mStringOne);
+ break;
+ case UnsupportedTest:
+ err = i18n("Test \"%1\" not supported").arg(mStringOne);
+ break;
+ case UnsupportedComparator:
+ err = i18n("Comparator \"%1\" not supported").arg(mStringOne);
+ break;
+ case TestNestingTooDeep:
+ err = i18n("Site Policy Limit Violation: Test nesting too deep (max. %1)").arg( mStringOne.toUInt() );
+ break;
+ case BlockNestingTooDeep:
+ err = i18n("Site Policy Limit Violation: Block nesting too deep (max. %1)").arg( mStringOne.toUInt() );
+ break;
+ case InvalidArgument:
+ err = i18n("Invalid Argument \"%1\" to \"%2\"").arg(mStringOne).arg(mStringTwo);
+ break;
+ case ConflictingArguments:
+ err = i18n("Conflicting Arguments: \"%1\" and \"%2\"").arg(mStringOne).arg(mStringTwo);
+ break;
+ case ArgumentsRepeated:
+ err = i18n("Argument \"%1\" Repeated").arg(mStringOne);
+ break;
+ case CommandOrderingConstraintViolation:
+ err = i18n("Command \"%1\" violates command ordering constraints").arg(mStringOne);
+ break;
+
+ // runtime errors:
+ case IncompatibleActionsRequested:
+ err = i18n("Incompatible Actions \"%1\" and \"%2\" requested").arg(mStringOne).arg(mStringTwo);
+ break;
+ case MailLoopDetected:
+ err = i18n("Mail Loop detected");
+ break;
+ case TooManyActions:
+ err = i18n("Site Policy Limit Violation: Too many Actions requested (max. %1)").arg( mStringOne.toUInt() );
+ break;
+ default:
+ err = i18n("Unknown error");
+ break;
+ }
+
+ return err;
+ }
+
+
+} // namespace KSieve
+
diff --git a/libksieve/tests/Makefile.am b/libksieve/tests/Makefile.am
new file mode 100644
index 000000000..36b538408
--- /dev/null
+++ b/libksieve/tests/Makefile.am
@@ -0,0 +1,13 @@
+
+INCLUDES = -I$(top_srcdir)/libksieve $(all_includes)
+LDADD = ../libksieve.la
+
+# test programs:
+check_PROGRAMS = \
+ lexertest \
+ parsertest
+
+TESTS = $(check_PROGRAMS)
+
+lexertest_SOURCES = lexertest.cpp
+parsertest_SOURCES = parsertest.cpp
diff --git a/libksieve/tests/lexertest.cpp b/libksieve/tests/lexertest.cpp
new file mode 100644
index 000000000..461499501
--- /dev/null
+++ b/libksieve/tests/lexertest.cpp
@@ -0,0 +1,484 @@
+/* -*- c++ -*-
+ tests/lexertest.cpp
+
+ This file is part of the testsuite of KSieve,
+ the KDE internet mail/usenet news message filtering library.
+ Copyright (c) 2003 Marc Mutz <mutz@kde.org>
+
+ KSieve is free software; you can redistribute it and/or modify it
+ under the terms of the GNU General Public License, version 2, as
+ published by the Free Software Foundation.
+
+ KSieve 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 this program; if not, write to the Free Software
+ Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
+
+ In addition, as a special exception, the copyright holders give
+ permission to link the code of this program with any edition of
+ the Qt library by Trolltech AS, Norway (or with modified versions
+ of Qt that use the same license as Qt), and distribute linked
+ combinations including the two. You must obey the GNU General
+ Public License in all respects for all of the code used other than
+ Qt. If you modify this file, you may extend this exception to
+ your version of the file, but you are not obligated to do so. If
+ you do not wish to do so, delete this exception statement from
+ your version.
+*/
+#include <config.h>
+#include <ksieve/lexer.h>
+using KSieve::Lexer;
+
+#include <ksieve/error.h>
+using KSieve::Error;
+
+#include <qcstring.h> // qstrlen
+#include <qstring.h>
+
+#include <iostream>
+using std::cout;
+using std::cerr;
+using std::endl;
+
+static const char * token2string( Lexer::Token t ) {
+ switch ( t ) {
+#define CASE(x) case Lexer::x: return #x
+ CASE( None );
+ CASE( HashComment );
+ CASE( BracketComment );
+ CASE( Identifier );
+ CASE( Tag );
+ CASE( Number );
+ CASE( MultiLineString );
+ CASE( QuotedString );
+ CASE( Special );
+ CASE( LineFeeds );
+ }
+ return "";
+#undef CASE
+}
+
+struct TestCase {
+ const char * name;
+ const char * string;
+ struct {
+ Lexer::Token token;
+ const char * result;
+ } expected[16]; // end with { None, 0 }
+ Error::Type expectedError;
+ int errorLine, errorCol;
+};
+
+static const TestCase testcases[] = {
+ //
+ // Whitespace:
+ //
+
+ { "Null script", 0,
+ { { Lexer::None, 0 } },
+ Error::None, 0, 0
+ },
+
+ { "Empty script", "",
+ { { Lexer::None, 0 } },
+ Error::None, 0, 0
+ },
+
+ { "Whitespace-only script", " \t\n\t \n",
+ { { Lexer::LineFeeds, "2" }, { Lexer::None, 0 } },
+ Error::None, 0, 0
+ },
+
+ { "Lone CR", "\r",
+ { { Lexer::None, 0 } },
+ Error::CRWithoutLF, 0, 1
+ },
+
+ { "CR+Space", "\r ",
+ { { Lexer::None, 0 } },
+ Error::CRWithoutLF, 0, 1
+ },
+
+ { "CRLF alone", "\r\n",
+ { { Lexer::LineFeeds, "1" }, { Lexer::None, 0 } },
+ Error::None, 0, 0
+ },
+
+ //
+ // hash comments:
+ //
+
+ { "Basic hash comment (no newline)", "#comment",
+ { { Lexer::HashComment, "comment" }, { Lexer::None, 0 } },
+ Error::None, 0, 0
+ },
+
+ { "Basic hash comment (LF)", "#comment\n",
+ { { Lexer::HashComment, "comment" }, { Lexer::None, 0 } },
+ Error::None, 0, 0
+ },
+
+ { "Basic hash comment (CRLF)", "#comment\r\n",
+ { { Lexer::HashComment, "comment" }, { Lexer::None, 0 } },
+ Error::None, 0, 0
+ },
+
+ { "Basic hash comment (CR)", "#comment\r",
+ { { Lexer::HashComment, 0 } },
+ Error::CRWithoutLF, 0, 9
+ },
+
+ { "Non-UTF-8 in hash comment", "#\xA9 copyright",
+ { { Lexer::HashComment, 0 } },
+ Error::InvalidUTF8, 0, 12
+ },
+
+ //
+ // bracket comments:
+ //
+
+ { "Basic bracket comment", "/* comment */",
+ { { Lexer::BracketComment, " comment " }, { Lexer::None, 0 } },
+ Error::None, 0, 0
+ },
+
+ { "Basic bracket comment - missing trailing slash", "/* comment *",
+ { { Lexer::BracketComment, 0 } },
+ Error::UnfinishedBracketComment, 0, 0
+ },
+
+ { "Basic bracket comment - missing trailing asterisk + slash", "/* comment ",
+ { { Lexer::BracketComment, 0 } },
+ Error::UnfinishedBracketComment, 0, 0
+ },
+
+ { "Basic bracket comment - missing leading slash", "* comment */",
+ { { Lexer::None, 0 } },
+ Error::IllegalCharacter, 0, 0
+ },
+
+ { "Basic bracket comment - missing leading asterisk + slash", "comment */",
+ { { Lexer::Identifier, "comment" }, { Lexer::None, 0 } },
+ Error::IllegalCharacter, 0, 8
+ },
+
+ { "Basic multiline bracket comment (LF)", "/* comment\ncomment */",
+ { { Lexer::BracketComment, " comment\ncomment " }, { Lexer::None, 0 } },
+ Error::None, 0, 0
+ },
+
+ { "Basic multiline bracket comment (CRLF)", "/* comment\r\ncomment */",
+ { { Lexer::BracketComment, " comment\ncomment " }, { Lexer::None, 0 } },
+ Error::None, 0, 0
+ },
+
+ { "Basic multiline bracket comment (CR)", "/* comment\rcomment */",
+ { { Lexer::BracketComment, 0 } },
+ Error::CRWithoutLF, 0, 11
+ },
+
+ { "Non-UTF-8 in bracket comment", "/*\xA9 copyright*/",
+ { { Lexer::BracketComment, 0 } },
+ Error::InvalidUTF8, 0, 14
+ },
+
+ //
+ // numbers:
+ //
+ { "Basic number 1", "1",
+ { { Lexer::Number, "1" }, { Lexer::None, 0 } },
+ Error::None, 0, 0
+ },
+ { "Basic number 01", "01",
+ { { Lexer::Number, "01" }, { Lexer::None, 0 } },
+ Error::None, 0, 0
+ },
+ { "Qualified number 1k", "1k",
+ { { Lexer::Number, "1k" }, { Lexer::None, 0 } },
+ Error::None, 0, 0
+ },
+ { "Qualified number 1M", "1M",
+ { { Lexer::Number, "1M" }, { Lexer::None, 0 } },
+ Error::None, 0, 0
+ },
+ { "Qualified number 1G", "1G",
+ { { Lexer::Number, "1G" }, { Lexer::None, 0 } },
+ Error::None, 0, 0
+ },
+ //
+ // identifiers:
+ //
+ { "Basic identifier \"id\"", "id",
+ { { Lexer::Identifier, "id" }, { Lexer::None, 0 } },
+ Error::None, 0, 0
+ },
+ { "Basic identifier \"_id\"", "_id",
+ { { Lexer::Identifier, "_id" }, { Lexer::None, 0 } },
+ Error::None, 0, 0
+ },
+ //
+ // tags:
+ //
+ { "Basic tag \":tag\"", ":tag",
+ { { Lexer::Tag, "tag" }, { Lexer::None, 0 } },
+ Error::None, 0, 0
+ },
+ { "Basic tag \":_tag\"", ":_tag",
+ { { Lexer::Tag, "_tag" }, { Lexer::None, 0 } },
+ Error::None, 0, 0
+ },
+ //
+ // specials:
+ //
+ { "Basic special \"{}[]();,\"", "{}[]();,",
+ { { Lexer::Special, "{" }, { Lexer::Special, "}" },
+ { Lexer::Special, "[" }, { Lexer::Special, "]" },
+ { Lexer::Special, "(" }, { Lexer::Special, ")" },
+ { Lexer::Special, ";" }, { Lexer::Special, "," }, { Lexer::None, 0 } },
+ Error::None, 0, 0
+ },
+ //
+ // quoted-string:
+ //
+ { "Basic quoted string \"foo\"", "\"foo\"",
+ { { Lexer::QuotedString, "foo" }, { Lexer::None, 0 } },
+ Error::None, 0, 0
+ },
+ { "Basic quoted string, UTF-8", "\"foo\xC3\xB1" "foo\"", // fooäfoo
+ { { Lexer::QuotedString, "foo\xC3\xB1" "foo" }, { Lexer::None, 0 } },
+ Error::None, 0, 0
+ },
+ { "Quoted string, escaped '\"'", "\"foo\\\"bar\"",
+ { { Lexer::QuotedString, "foo\"bar" }, { Lexer::None, 0 } },
+ Error::None, 0, 0
+ },
+ { "Quoted string, escaped '\\'", "\"foo\\\\bar\"",
+ { { Lexer::QuotedString, "foo\\bar" }, { Lexer::None, 0 } },
+ Error::None, 0, 0
+ },
+ { "Quoted string, excessive escapes", "\"\\fo\\o\"",
+ { { Lexer::QuotedString, "foo" }, { Lexer::None, 0 } },
+ Error::None, 0, 0
+ },
+ { "Quoted string across lines (LF)", "\"foo\nbar\"",
+ { { Lexer::QuotedString, "foo\nbar" }, { Lexer::None, 0 } },
+ Error::None, 0, 0
+ },
+ { "Quoted string across lines (CRLF)", "\"foo\r\nbar\"",
+ { { Lexer::QuotedString, "foo\nbar" }, { Lexer::None, 0 } },
+ Error::None, 0, 0
+ },
+ //
+ // multiline strings:
+ //
+ { "Basic multiline string I (LF)", "text:\nfoo\n.",
+ { { Lexer::MultiLineString, "foo" /* "foo\n" ? */ }, { Lexer::None, 0 } },
+ Error::None, 0, 0
+ },
+ { "Basic multiline string I (CRLF)", "text:\r\nfoo\r\n.",
+ { { Lexer::MultiLineString, "foo" /* "foo\n" ? */ }, { Lexer::None, 0 } },
+ Error::None, 0, 0
+ },
+ { "Basic multiline string II (LF)", "text:\nfoo\n.\n",
+ { { Lexer::MultiLineString, "foo" /* "foo\n" ? */ }, { Lexer::None, 0 } },
+ Error::None, 0, 0
+ },
+ { "Basic multiline string II (CRLF)", "text:\r\nfoo\r\n.\r\n",
+ { { Lexer::MultiLineString, "foo" /* "foo\n" ? */ }, { Lexer::None, 0 } },
+ Error::None, 0, 0
+ },
+ { "Dotstuffed multiline string (LF)", "text:\n..foo\n.",
+ { { Lexer::MultiLineString, ".foo" /* ".foo\n" ? */ }, { Lexer::None, 0 } },
+ Error::None, 0, 0
+ },
+ { "Dotstuffed multiline string (CRLF)", "text:\r\n..foo\r\n.",
+ { { Lexer::MultiLineString, ".foo" /* ".foo\n" ? */ }, { Lexer::None, 0 } },
+ Error::None, 0, 0
+ },
+ { "Incompletely dotstuffed multiline string (LF)", "text:\n.foo\n.",
+ { { Lexer::MultiLineString, ".foo" /* ".foo\n" ? */ }, { Lexer::None, 0 } },
+ Error::None, 0, 0
+ },
+ { "Incompletely dotstuffed multiline string (CRLF)", "text:\r\n.foo\r\n.",
+ { { Lexer::MultiLineString, ".foo" /* ".foo\n" ? */ }, { Lexer::None, 0 } },
+ Error::None, 0, 0
+ },
+ { "Mutiline with a line with only one '.'","text:\r\nfoo\r\n..\r\nbar\r\n.",
+ { { Lexer::MultiLineString, "foo\n.\nbar" }, { Lexer::None, 0 } },
+ Error::None, 0, 0
+ },
+
+
+ //
+ // Errors in single tokens:
+ //
+
+ //
+ // numbers:
+ //
+ { "Number, unknown qualifier", "100f",
+ { { Lexer::Number, "100" } },
+ Error::UnexpectedCharacter, 0, 3
+ },
+ { "Negative number", "-100",
+ { { Lexer::None, 0 } },
+ Error::IllegalCharacter, 0, 0
+ },
+ //
+ // identifiers:
+ //
+ { "Identifier, leading digits", "0id",
+ { { Lexer::Number, "0" } },
+ Error::UnexpectedCharacter, 0, 1
+ },
+ { "Identifier, embedded umlaut", "idäid",
+ { { Lexer::Identifier, "id" } },
+ Error::IllegalCharacter, 0, 2
+ },
+ //
+ // tags:
+ //
+ { "Lone ':' (at end)", ":",
+ { { Lexer::Tag, 0 } },
+ Error::UnexpectedCharacter, 0, 0
+ },
+ { "Lone ':' (in stream)", ": ",
+ { { Lexer::Tag, 0 } },
+ Error::UnexpectedCharacter, 0, 1
+ },
+ { "Tag, leading digits", ":0tag",
+ { { Lexer::Tag, 0 } },
+ Error::NoLeadingDigits, 0, 1
+ },
+ { "Tag, embedded umlaut", ":tagätag",
+ { { Lexer::Tag, "tag" } },
+ Error::IllegalCharacter, 0, 4
+ },
+ //
+ // specials: (none)
+ // quoted string:
+ //
+ { "Premature end of quoted string", "\"foo",
+ { { Lexer::QuotedString, "foo" } },
+ Error::PrematureEndOfQuotedString, 0, 0
+ },
+ { "Invalid UTF-8 in quoted string", "\"foo\xC0\xA0" "foo\"",
+ { { Lexer::QuotedString, "foo" } },
+ Error::InvalidUTF8, 0, 4
+ },
+
+ //
+ // Whitespace / token separation: valid
+ //
+
+ { "Two identifiers with linebreaks", "foo\nbar\n",
+ { { Lexer::Identifier, "foo" },
+ { Lexer::LineFeeds, "1" },
+ { Lexer::Identifier, "bar" },
+ { Lexer::LineFeeds, "1" },
+ { Lexer::None, 0 } },
+ Error::None, 0, 0
+ },
+
+ //
+ // Whitespace / token separation: invalid
+ //
+
+};
+
+static const int numTestCases = sizeof testcases / sizeof *testcases ;
+
+int main( int argc, char * argv[] ) {
+
+ if ( argc == 2 ) { // manual test
+
+ const char * scursor = argv[1];
+ const char * const send = argv[1] + qstrlen( argv[1] );
+
+ Lexer lexer( scursor, send );
+
+ cout << "Begin" << endl;
+ while ( !lexer.atEnd() ) {
+ QString result;
+ Lexer::Token token = lexer.nextToken( result );
+ if ( lexer.error() ) {
+ cout << "Error " << token2string( token ) << ": \""
+ << lexer.error().asString().latin1() << "\" at ("
+ << lexer.error().line() << "," << lexer.error().column()
+ << ")" << endl;
+ break;
+ } else
+ cout << "Got " << token2string( token ) << ": \""
+ << result.utf8().data() << "\" at ("
+ << lexer.line() << "," << lexer.column() << ")" << endl;
+ }
+ cout << "End" << endl;
+
+ } else if ( argc == 1 ) { // automated test
+ bool success = true;
+ for ( int i = 0 ; i < numTestCases ; ++i ) {
+ bool ok = true;
+ const TestCase & t = testcases[i];
+ const char * const send = t.string + qstrlen( t.string );
+ Lexer lexer( t.string, send, Lexer::IncludeComments );
+ cerr << t.name << ":";
+ for ( int j = 0 ; !lexer.atEnd() ; ++j ) {
+ QString result;
+ Lexer::Token token = lexer.nextToken( result );
+ Error error = lexer.error();
+ if ( t.expected[j].token != token ) {
+ ok = false;
+ cerr << " expected token " << token2string( t.expected[j].token )
+ << ", got " << token2string( token );
+ }
+ if ( QString::fromUtf8( t.expected[j].result ) != result ) {
+ ok = false;
+ if ( t.expected[j].result )
+ cerr << " expected string \"" << t.expected[j].result << "\"";
+ else
+ cerr << " expected null string";
+ if ( !result.utf8().isNull() )
+ cerr << ", got \"" << result.utf8().data() << "\"";
+ else
+ cerr << ", got null string";
+ }
+ if ( error && error.type() != t.expectedError ) {
+ ok = false;
+ cerr << " expected error #" << (int)t.expectedError
+ << ", got #" << (int)error.type();
+ }
+ if ( error && ( error.line() != t.errorLine || error.column() != t.errorCol ) ) {
+ ok = false;
+ cerr << " expected position (" << t.errorLine << "," << t.errorCol
+ << "), got (" << error.line() << "," << error.column() << ")";
+ }
+ if ( error )
+ goto ErrorOut;
+ if ( t.expected[j].token == Lexer::None &&
+ t.expected[j].result == 0 )
+ break;
+ }
+ if ( !lexer.atEnd() ) {
+ ok = false;
+ cerr << " premature end of expected token list";
+ }
+ ErrorOut:
+ if ( ok )
+ cerr << " ok";
+ cerr << endl;
+ if ( !ok )
+ success = false;
+ }
+ if ( !success )
+ return 1;
+ } else { // usage error
+ cerr << "usage: lexertest [ <string> ]" << endl;
+ exit( 1 );
+ }
+
+ return 0;
+}
diff --git a/libksieve/tests/parsertest.cpp b/libksieve/tests/parsertest.cpp
new file mode 100644
index 000000000..e2ea0fd39
--- /dev/null
+++ b/libksieve/tests/parsertest.cpp
@@ -0,0 +1,667 @@
+/* -*- c++ -*-
+ tests/parsertest.cpp
+
+ This file is part of the testsuite of KSieve,
+ the KDE internet mail/usenet news message filtering library.
+ Copyright (c) 2003 Marc Mutz <mutz@kde.org>
+
+ KSieve is free software; you can redistribute it and/or modify it
+ under the terms of the GNU General Public License, version 2, as
+ published by the Free Software Foundation.
+
+ KSieve 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 this program; if not, write to the Free Software
+ Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
+
+ In addition, as a special exception, the copyright holders give
+ permission to link the code of this program with any edition of
+ the Qt library by Trolltech AS, Norway (or with modified versions
+ of Qt that use the same license as Qt), and distribute linked
+ combinations including the two. You must obey the GNU General
+ Public License in all respects for all of the code used other than
+ Qt. If you modify this file, you may extend this exception to
+ your version of the file, but you are not obligated to do so. If
+ you do not wish to do so, delete this exception statement from
+ your version.
+*/
+#include <config.h>
+#include <ksieve/parser.h>
+using KSieve::Parser;
+
+#include <ksieve/error.h>
+#include <ksieve/scriptbuilder.h>
+
+#include <qcstring.h> // qstrlen
+#include <qstring.h>
+
+#include <iostream>
+using std::cout;
+using std::cerr;
+using std::endl;
+
+#include <cassert>
+
+enum BuilderMethod {
+ TaggedArgument,
+ StringArgument,
+ NumberArgument,
+ CommandStart,
+ CommandEnd,
+ TestStart,
+ TestEnd,
+ TestListStart,
+ TestListEnd,
+ BlockStart,
+ BlockEnd,
+ StringListArgumentStart,
+ StringListEntry,
+ StringListArgumentEnd,
+ HashComment,
+ BracketComment,
+ Error,
+ Finished
+};
+
+static const unsigned int MAX_RESPONSES = 100;
+
+struct TestCase {
+ const char * name;
+ const char * script;
+ struct Response {
+ BuilderMethod method;
+ const char * string;
+ bool boolean;
+ } responses[MAX_RESPONSES];
+} testCases[] = {
+
+ //
+ // single commands:
+ //
+
+ { "Null script",
+ 0,
+ { { Finished, 0, false } }
+ },
+
+ { "Empty script",
+ "",
+ { { Finished, 0, false } }
+ },
+
+ { "WS-only script",
+ " \t\n\r\n",
+ { { Finished, 0, false } }
+ },
+
+ { "Bare hash comment",
+ "#comment",
+ { { HashComment, "comment", false },
+ { Finished, 0, false } }
+ },
+
+ { "Bare bracket comment",
+ "/*comment*/",
+ { { BracketComment, "comment", false },
+ { Finished, 0, false } }
+ },
+
+ { "Bare command",
+ "command;",
+ { { CommandStart, "command", false },
+ { CommandEnd, 0, false },
+ { Finished, 0, false } }
+ },
+
+ { "Bare command - missing semicolon",
+ "command",
+ { { CommandStart, "command", false },
+ { Error, "MissingSemicolonOrBlock", false } }
+ },
+
+ { "surrounded by bracket comments",
+ "/*comment*/command/*comment*/;/*comment*/",
+ { { BracketComment, "comment", false },
+ { CommandStart, "command", false },
+ { BracketComment, "comment", false },
+ { CommandEnd, 0, false },
+ { BracketComment, "comment", false },
+ { Finished, 0, false } }
+ },
+
+ { "surrounded by hash comments",
+ "#comment\ncommand#comment\n;#comment",
+ { { HashComment, "comment", false },
+ { CommandStart, "command", false },
+ { HashComment, "comment", false },
+ { CommandEnd, 0, false },
+ { HashComment, "comment", false },
+ { Finished, 0, false } }
+ },
+
+ { "single tagged argument",
+ "command :tag;",
+ { { CommandStart, "command", false },
+ { TaggedArgument, "tag", false },
+ { CommandEnd, 0, false },
+ { Finished, 0, false } }
+ },
+
+ { "single tagged argument - missing semicolon",
+ "command :tag",
+ { { CommandStart, "command", false },
+ { TaggedArgument, "tag", false },
+ { Error, "MissingSemicolonOrBlock", false } }
+ },
+
+ { "single string argument - quoted string",
+ "command \"string\";",
+ { { CommandStart, "command", false },
+ { StringArgument, "string", false /*quoted*/ },
+ { CommandEnd, 0, false },
+ { Finished, 0, false } }
+ },
+
+ { "single string argument - multi-line string",
+ "command text:\nstring\n.\n;",
+ { { CommandStart, "command", false },
+ { StringArgument, "string", true /*multiline*/ },
+ { CommandEnd, 0, false },
+ { Finished, 0, false } }
+ },
+
+ { "single number argument - 100",
+ "command 100;",
+ { { CommandStart, "command", false },
+ { NumberArgument, "100 ", false },
+ { CommandEnd, 0, false },
+ { Finished, 0, false } }
+ },
+
+ { "single number argument - 100k",
+ "command 100k;",
+ { { CommandStart, "command", false },
+ { NumberArgument, "102400k", false },
+ { CommandEnd, 0, false },
+ { Finished, 0, false } }
+ },
+
+ { "single number argument - 100M",
+ "command 100M;",
+ { { CommandStart, "command", false },
+ { NumberArgument, "104857600M", false },
+ { CommandEnd, 0, false },
+ { Finished, 0, false } }
+ },
+
+ { "single number argument - 2G",
+ "command 2G;",
+ { { CommandStart, "command", false },
+ { NumberArgument, "2147483648G", false },
+ { CommandEnd, 0, false },
+ { Finished, 0, false } }
+ },
+
+#if SIZEOF_UNSIGNED_LONG == 8
+# define ULONG_MAX_STRING "18446744073709551615"
+# define ULONG_MAXP1_STRING "18446744073709551616"
+#elif SIZEOF_UNSIGNED_LONG == 4
+# define ULONG_MAX_STRING "4294967295"
+# define ULONG_MAXP1_STRING "4G"
+#else
+# error sizeof( unsigned long ) != 4 && sizeof( unsigned long ) != 8 ???
+#endif
+
+ { "single number argument - ULONG_MAX + 1",
+ "command " ULONG_MAXP1_STRING ";",
+ { { CommandStart, "command", false },
+ { Error, "NumberOutOfRange", false } }
+ },
+
+ { "single number argument - ULONG_MAX",
+ "command " ULONG_MAX_STRING ";",
+ { { CommandStart, "command", false },
+ { NumberArgument, ULONG_MAX_STRING " ", false },
+ { CommandEnd, 0, false },
+ { Finished, 0, false } }
+ },
+
+ { "single one-element string list argument - quoted string",
+ "command [\"string\"];",
+ { { CommandStart, "command", false },
+ { StringListArgumentStart, 0, false },
+ { StringListEntry, "string", false /*quoted*/ },
+ { StringListArgumentEnd, 0, false },
+ { CommandEnd, 0, false },
+ { Finished, 0, false } }
+ },
+
+ { "single one-element string list argument - multi-line string",
+ "command [text:\nstring\n.\n];",
+ { { CommandStart, "command", false },
+ { StringListArgumentStart, 0, false },
+ { StringListEntry, "string", true /*multiline*/ },
+ { StringListArgumentEnd, 0, false },
+ { CommandEnd, 0, false },
+ { Finished, 0, false } }
+ },
+
+ { "single two-element string list argument - quoted strings",
+ "command [\"string\",\"string\"];",
+ { { CommandStart, "command", false },
+ { StringListArgumentStart, 0, false },
+ { StringListEntry, "string", false /*quoted*/ },
+ { StringListEntry, "string", false /*quoted*/ },
+ { StringListArgumentEnd, 0, false },
+ { CommandEnd, 0, false },
+ { Finished, 0, false } }
+ },
+
+ { "single two-element string list argument - multi-line strings",
+ "command [text:\nstring\n.\n,text:\nstring\n.\n];",
+ { { CommandStart, "command", false },
+ { StringListArgumentStart, 0, false },
+ { StringListEntry, "string", true /*multiline*/ },
+ { StringListEntry, "string", true /*multiline*/ },
+ { StringListArgumentEnd, 0, false },
+ { CommandEnd, 0, false },
+ { Finished, 0, false } }
+ },
+
+ { "single two-element string list argument - quoted + multi-line strings",
+ "command [\"string\",text:\nstring\n.\n];",
+ { { CommandStart, "command", false },
+ { StringListArgumentStart, 0, false },
+ { StringListEntry, "string", false /*quoted*/ },
+ { StringListEntry, "string", true /*multiline*/ },
+ { StringListArgumentEnd, 0, false },
+ { CommandEnd, 0, false },
+ { Finished, 0, false } }
+ },
+
+ { "single two-element string list argument - multi-line + quoted strings",
+ "command [text:\nstring\n.\n,\"string\"];",
+ { { CommandStart, "command", false },
+ { StringListArgumentStart, 0, false },
+ { StringListEntry, "string", true /*multiline*/ },
+ { StringListEntry, "string", false /*quoted*/ },
+ { StringListArgumentEnd, 0, false },
+ { CommandEnd, 0, false },
+ { Finished, 0, false } }
+ },
+
+ { "single bare test argument",
+ "command test;",
+ { { CommandStart, "command", false },
+ { TestStart, "test", false },
+ { TestEnd, 0, false },
+ { CommandEnd, 0, false },
+ { Finished, 0, false } }
+ },
+
+ { "one-element test list argument",
+ "command(test);",
+ { { CommandStart, "command", false },
+ { TestListStart, 0, false },
+ { TestStart, "test", false },
+ { TestEnd, 0, false },
+ { TestListEnd, 0, false },
+ { CommandEnd, 0, false },
+ { Finished, 0, false } }
+ },
+
+ { "two-element test list argument",
+ "command(test,test);",
+ { { CommandStart, "command", false },
+ { TestListStart, 0, false },
+ { TestStart, "test", false },
+ { TestEnd, 0, false },
+ { TestStart, "test", false },
+ { TestEnd, 0, false },
+ { TestListEnd, 0, false },
+ { CommandEnd, 0, false },
+ { Finished, 0, false } }
+ },
+
+ { "zero-element block",
+ "command{}",
+ { { CommandStart, "command", false },
+ { BlockStart, 0, false },
+ { BlockEnd, 0, false },
+ { CommandEnd, 0, false },
+ { Finished, 0, false } }
+ },
+
+ { "one-element block",
+ "command{command;}",
+ { { CommandStart, "command", false },
+ { BlockStart, 0, false },
+ { CommandStart, "command", false },
+ { CommandEnd, 0, false },
+ { BlockEnd, 0, false },
+ { CommandEnd, 0, false },
+ { Finished, 0, false } }
+ },
+
+ { "two-element block",
+ "command{command;command;}",
+ { { CommandStart, "command", false },
+ { BlockStart, 0, false },
+ { CommandStart, "command", false },
+ { CommandEnd, 0, false },
+ { CommandStart, "command", false },
+ { CommandEnd, 0, false },
+ { BlockEnd, 0, false },
+ { CommandEnd, 0, false },
+ { Finished, 0, false } }
+ },
+
+ { "command with a test with a test with a test",
+ "command test test test;",
+ { { CommandStart, "command", false },
+ { TestStart, "test", false },
+ { TestStart, "test", false },
+ { TestStart, "test", false },
+ { TestEnd, 0, false },
+ { TestEnd, 0, false },
+ { TestEnd, 0, false },
+ { CommandEnd, 0, false },
+ { Finished, 0, false } }
+ },
+
+};
+
+static const int numTestCases = sizeof testCases / sizeof *testCases ;
+
+// Prints out the parse tree in XML-like format. For visual inspection
+// (manual tests).
+class PrintingScriptBuilder : public KSieve::ScriptBuilder {
+public:
+ PrintingScriptBuilder()
+ : KSieve::ScriptBuilder(), indent( 0 )
+ {
+ write( "<script type=\"application/sieve\">" );
+ ++indent;
+ }
+ virtual ~PrintingScriptBuilder() {}
+
+ void taggedArgument( const QString & tag ) {
+ write( "tag", tag );
+ }
+ void stringArgument( const QString & string, bool multiLine, const QString & /*fixme*/ ) {
+ write( multiLine ? "string type=\"multiline\"" : "string type=\"quoted\"", string );
+ }
+ void numberArgument( unsigned long number, char quantifier ) {
+ const QString txt = "number" + ( quantifier ? QString(" quantifier=\"%1\"").arg( quantifier ) : QString::null ) ;
+ write( txt.latin1(), QString::number( number ) );
+ }
+ void commandStart( const QString & identifier ) {
+ write( "<command>" );
+ ++indent;
+ write( "identifier", identifier );
+ }
+ void commandEnd() {
+ --indent;
+ write( "</command>" );
+ }
+ void testStart( const QString & identifier ) {
+ write( "<test>" );
+ ++indent;
+ write( "identifier", identifier );
+ }
+ void testEnd() {
+ --indent;
+ write( "</test>" );
+ }
+ void testListStart() {
+ write( "<testlist>" );
+ ++indent;
+ }
+ void testListEnd() {
+ --indent;
+ write( "</testlist>" );
+ }
+ void blockStart() {
+ write( "<block>" );
+ ++indent;
+ }
+ void blockEnd() {
+ --indent;
+ write( "</block>" );
+ }
+ void stringListArgumentStart() {
+ write( "<stringlist>" );
+ ++indent;
+ }
+ void stringListArgumentEnd() {
+ --indent;
+ write( "</stringlist>" );
+ }
+ void stringListEntry( const QString & string, bool multiline, const QString & hashComment ) {
+ stringArgument( string, multiline, hashComment );
+ }
+ void hashComment( const QString & comment ) {
+ write( "comment type=\"hash\"", comment );
+ }
+ void bracketComment( const QString & comment ) {
+ write( "comment type=\"bracket\"", comment );
+ }
+
+ void lineFeed() {
+ write( "<crlf/>" );
+ }
+
+ void error( const KSieve::Error & error ) {
+ indent = 0;
+ write( ("Error: " + error.asString()).latin1() );
+ }
+ void finished() {
+ --indent;
+ write( "</script>" );
+ }
+private:
+ int indent;
+ void write( const char * msg ) {
+ for ( int i = 2*indent ; i > 0 ; --i )
+ cout << " ";
+ cout << msg << endl;
+ }
+ void write( const QCString & key, const QString & value ) {
+ if ( value.isEmpty() ) {
+ write( "<" + key + "/>" );
+ return;
+ }
+ write( "<" + key + ">" );
+ ++indent;
+ write( value.utf8().data() );
+ --indent;
+ write( "</" + key + ">" );
+ }
+};
+
+
+// verifes that methods get called with expected arguments (and in
+// expected sequence) as specified by the TestCase. For automated
+// tests.
+class VerifyingScriptBuilder : public KSieve::ScriptBuilder {
+public:
+ VerifyingScriptBuilder( const TestCase & testCase )
+ : KSieve::ScriptBuilder(),
+ mNextResponse( 0 ), mTestCase( testCase ), mOk( true )
+ {
+ }
+ virtual ~VerifyingScriptBuilder() {}
+
+ bool ok() const { return mOk; }
+
+ void taggedArgument( const QString & tag ) {
+ checkIs( TaggedArgument );
+ checkEquals( tag );
+ ++mNextResponse;
+ }
+ void stringArgument( const QString & string, bool multiline, const QString & /*fixme*/ ) {
+ checkIs( StringArgument );
+ checkEquals( string );
+ checkEquals( multiline );
+ ++mNextResponse;
+ }
+ void numberArgument( unsigned long number, char quantifier ) {
+ checkIs( NumberArgument );
+ checkEquals( QString::number( number ) + ( quantifier ? quantifier : ' ' ) );
+ ++mNextResponse;
+ }
+ void commandStart( const QString & identifier ) {
+ checkIs( CommandStart );
+ checkEquals( identifier );
+ ++mNextResponse;
+ }
+ void commandEnd() {
+ checkIs( CommandEnd );
+ ++mNextResponse;
+ }
+ void testStart( const QString & identifier ) {
+ checkIs( TestStart );
+ checkEquals( identifier );
+ ++mNextResponse;
+ }
+ void testEnd() {
+ checkIs( TestEnd );
+ ++mNextResponse;
+ }
+ void testListStart() {
+ checkIs( TestListStart );
+ ++mNextResponse;
+ }
+ void testListEnd() {
+ checkIs( TestListEnd );
+ ++mNextResponse;
+ }
+ void blockStart() {
+ checkIs( BlockStart );
+ ++mNextResponse;
+ }
+ void blockEnd() {
+ checkIs( BlockEnd );
+ ++mNextResponse;
+ }
+ void stringListArgumentStart() {
+ checkIs( StringListArgumentStart );
+ ++mNextResponse;
+ }
+ void stringListEntry( const QString & string, bool multiLine, const QString & /*fixme*/ ) {
+ checkIs( StringListEntry );
+ checkEquals( string );
+ checkEquals( multiLine );
+ ++mNextResponse;
+ }
+ void stringListArgumentEnd() {
+ checkIs( StringListArgumentEnd );
+ ++mNextResponse;
+ }
+ void hashComment( const QString & comment ) {
+ checkIs( HashComment );
+ checkEquals( comment );
+ ++mNextResponse;
+ }
+ void bracketComment( const QString & comment ) {
+ checkIs( BracketComment );
+ checkEquals( comment );
+ ++mNextResponse;
+ }
+ void lineFeed() {
+ // FIXME
+ }
+ void error( const KSieve::Error & error ) {
+ checkIs( Error );
+ checkEquals( QString( KSieve::Error::typeToString( error.type() ) ) );
+ ++mNextResponse;
+ }
+ void finished() {
+ checkIs( Finished );
+ //++mNextResponse (no!)
+ }
+
+private:
+ const TestCase::Response & currentResponse() const {
+ assert( mNextResponse <= MAX_RESPONSES );
+ return mTestCase.responses[mNextResponse];
+ }
+
+ void checkIs( BuilderMethod m ) {
+ if ( currentResponse().method != m ) {
+ cerr << " expected method " << (int)currentResponse().method
+ << ", got " << (int)m;
+ mOk = false;
+ }
+ }
+
+ void checkEquals( const QString & s ) {
+ if ( s != QString::fromUtf8( currentResponse().string ) ) {
+ cerr << " expected string arg \""
+ << ( currentResponse().string ? currentResponse().string : "<null>" )
+ << "\", got \"" << ( s.isNull() ? "<null>" : s.utf8().data() ) << "\"";
+ mOk = false;
+ }
+ }
+ void checkEquals( bool b ) {
+ if ( b != currentResponse().boolean ) {
+ cerr << " expected boolean arg <" << currentResponse().boolean
+ << ">, got <" << b << ">";
+ mOk = false;
+ }
+ }
+
+ unsigned int mNextResponse;
+ const TestCase & mTestCase;
+ bool mOk;
+};
+
+
+int main( int argc, char * argv[] ) {
+
+ if ( argc == 2 ) { // manual test
+
+ const char * scursor = argv[1];
+ const char * const send = argv[1] + qstrlen( argv[1] );
+
+ Parser parser( scursor, send );
+ PrintingScriptBuilder psb;
+ parser.setScriptBuilder( &psb );
+ if ( parser.parse() )
+ cout << "ok" << endl;
+ else
+ cout << "bad" << endl;
+
+
+ } else if ( argc == 1 ) { // automated test
+ bool success = true;
+ for ( int i = 0 ; i < numTestCases ; ++i ) {
+ const TestCase & t = testCases[i];
+ cerr << t.name << ":";
+ VerifyingScriptBuilder v( t );
+ Parser p( t.script, t.script + qstrlen( t.script ) );
+ p.setScriptBuilder( &v );
+ const bool ok = p.parse();
+ if ( v.ok() )
+ if ( ok )
+ cerr << " ok";
+ else
+ cerr << " xfail";
+ else
+ success = false;
+ cerr << endl;
+ }
+ if ( !success )
+ exit( 1 );
+
+ } else { // usage error
+ cerr << "usage: parsertest [ <string> ]" << endl;
+ exit( 1 );
+ }
+
+ return 0;
+}