int Scanner::consumeWhiteSpaceAndComments (const char *text) {
regex_t* whiteSpace=makeRegex("^[\n\t\r ]+") ;
regex_t* blockComment=makeRegex("^/\\*([^\\*]|\\*+[^\\*/])*\\*+/");
regex_t* lineComment=makeRegex("^//[^\n]*\n");
int numMatchedChars = 0 ;
int totalNumMatchedChars = 0 ;
int stillConsumingWhiteSpace ;
do {
stillConsumingWhiteSpace = 0 ; // exit loop if not reset by a match
// Try to match white space
numMatchedChars = matchRegex (whiteSpace, text) ;
totalNumMatchedChars += numMatchedChars ;
if (numMatchedChars > 0) {
text = text + numMatchedChars ;
stillConsumingWhiteSpace = 1 ;
}
// Try to match block comments
numMatchedChars = matchRegex (blockComment, text) ;
totalNumMatchedChars += numMatchedChars ;
if (numMatchedChars > 0) {
text = text + numMatchedChars ;
stillConsumingWhiteSpace = 1 ;
}
// Try to match single-line comments
numMatchedChars = matchRegex (lineComment, text) ;
totalNumMatchedChars += numMatchedChars ;
if (numMatchedChars > 0) {
text = text + numMatchedChars ;
stillConsumingWhiteSpace = 1 ;
}
}
while ( stillConsumingWhiteSpace ) ;
return totalNumMatchedChars ;
}
bool MakeLeftRules()
{
RULELIST * Rules = klf.getRules();
//vtERs.clear();
for (EXPENDEDRULELIST::iterator it = vtERs.begin(); it != vtERs.end(); ++it) {
structExpendedRule * r = &*it;
delete r->match_pattern;
delete r->vk;
delete r->switches;
delete r->regex;
}
vtERs.clear();
for (RULELIST::iterator it = Rules->begin(); it != Rules->end(); it++) {
std::wstring * str_in = new std::wstring;
VIRTUALKEYS * vks = new VIRTUALKEYS;
SWITCHES * sws = new SWITCHES;
if (makeRegex(str_in, vks, sws, it->strInRule, wcslen((wchar_t*)it->strInRule))) {
structExpendedRule er;
er.match_pattern = str_in;
er.vk = vks;
er.switches = sws;
er.estimated_length = estimate_length(str_in->c_str());
extractCharClasses(str_in->c_str(), &er.cc);
er.regex = new boost::wregex (str_in->c_str());
er.rule_it = it;
vtERs.push_back(er);
}
}
vtERs.sort();
vtERs.reverse();
return true;
}
bool getOutputAndSend(InputRuleInfo * ir_info, WORD wVk, LPBYTE KeyStates)
{
boost::wcmatch matches;
const wchar_t * srcString = InternalEditor.GetTextBackward(ir_info->it->estimated_length);
if (!srcString)
return false;
boost::regex_match(srcString, matches, *ir_info->it->regex);
int len = wcslen((wchar_t*)ir_info->it->rule_it->strOutRule);
std::wstring outStr ;
VIRTUALKEYS vks;
SWITCHES sws;
/*for (SWITCHES::iterator sws_it = ir_info->it->switches->begin();
sws_it != ir_info->it->switches->end();
sws_it++) {
InternalEditor.invertSwitch(*sws_it);
}*/
if (!makeRegex(&outStr, &vks, &sws, ir_info->it->rule_it->strOutRule, len, false, &matches, &ir_info->it->cc))
return false;
sendSingleKey(VK_CONTROL, KEYEVENTF_KEYUP);
sendSingleKey(VK_MENU, KEYEVENTF_KEYUP);
//sendSingleKey(VK_SHIFT, KEYEVENTF_KEYUP);
if (!outStr.length())
{
if (ir_info->deleted == false)
// Delete recent added temporary input
InternalEditor.Delete();
sendBackspace(wcslen(srcString));
if (mksOldState.CTRL)
sendSingleKey(VK_CONTROL, 0);
if (mksOldState.ALT)
sendSingleKey(VK_MENU, 0);
//if (mksOldState.SHIFT)
// sendSingleKey(VK_SHIFT, 0);
return true;
}
std::wstring resultStr;
//std::wstring * temp = new std::wstring(srcString);
resultStr = boost::regex_replace((std::wstring)srcString, *ir_info->it->regex, outStr);
//temp->clear();
if (!resultStr.size()){
//rise error
//continue;
return false;
}
if (ir_info->deleted == false)
// Delete recently added temporary input
InternalEditor.Delete();
int srcStringStringLen = wcslen(srcString);
outStr.assign(resultStr);
if (srcStringStringLen) {
int matchLength = getMatchLength(outStr.c_str(), srcString);
if (matchLength < 0){
sendBackspace(0 - matchLength);
matchLength = getMatchLength(outStr.c_str(), srcString);
sendBackspace(outStr.length() - matchLength);
}
else if (srcStringStringLen > matchLength){
sendBackspace(srcStringStringLen - matchLength);
}
outStr.erase(0, matchLength);
}
sendKeyStrokes(&outStr);
int strOutLength = outStr.length();
if ( strOutLength == 1 && (outStr.at(0)>0x20 && outStr.at(0)<0x7F) ){} else
{
// KeyUp match VKs
for (VIRTUALKEYS::iterator vk_it = ir_info->it->vk->begin();
vk_it != ir_info->it->vk->end();
vk_it++){
KeyStates[*vk_it] = 0;
}
// Do match again
matchRules(0, 0, KeyStates, false);
}
if (mksOldState.CTRL)
sendSingleKey(VK_CONTROL, 0);
if (mksOldState.ALT)
sendSingleKey(VK_MENU, 0);
//if (mksOldState.SHIFT)
// sendSingleKey(VK_SHIFT, 0);
return true;
}
void Scanner::buildDataBase(){
// malloc typeDatabase, this will be an array of pointers to regexes, there are 46 types of tokens, but we wont need all of them in regexes. this array will be easily extensible
typeDatabase=(regex_t**)malloc(43*(sizeof(regex_t*)));
typeDatabase[intKwd]=makeRegex("^Int");//intkwd
typeDatabase[floatKwd]=makeRegex("^Float");//floatkwd
typeDatabase[boolKwd]=makeRegex("^Bool");//boolkwd
typeDatabase[trueKwd]=makeRegex("^True");//truekwd
typeDatabase[falseKwd]=makeRegex("^False");//falsekwd
typeDatabase[stringKwd]=makeRegex("^Str");//stringkwd
typeDatabase[matrixKwd]=makeRegex("^Matrix");//matrixkwd
typeDatabase[letKwd]=makeRegex("^let");//letkwd
typeDatabase[inKwd]=makeRegex("^in");//inkwd
typeDatabase[endKwd]=makeRegex("^end");//endkwd
typeDatabase[ifKwd]=makeRegex("^if");//ifkwd
typeDatabase[thenKwd]=makeRegex("^then");//thenkwd
typeDatabase[elseKwd]=makeRegex("^else");//elsekwd
typeDatabase[forKwd]=makeRegex("^for");//forkwd
typeDatabase[whileKwd]=makeRegex("^while");//whilekwd
typeDatabase[printKwd]=makeRegex("^print");//printkwd
typeDatabase[intConst]=makeRegex("^[0-9]+");//intconst
typeDatabase[floatConst]=makeRegex("(^[0-9]+\\.[0-9]+)");
//|(^[0-9]*\\.[0-9]+)");//floatconst
typeDatabase[stringConst]=makeRegex("^\"[^\"]*\"");//stringconst
typeDatabase[variableName]=makeRegex("^[a-zA-Z_]+[0-9_]*");//variableName
typeDatabase[leftParen]=makeRegex("^\\(");//leftParen
typeDatabase[rightParen]=makeRegex("^\\)");//rightParen
typeDatabase[leftCurly]=makeRegex("^\\{");//leftCurly
typeDatabase[rightCurly]=makeRegex("^\\}");//rightCurly
typeDatabase[leftSquare]=makeRegex("^\\[");//leftsquare
typeDatabase[rightSquare]=makeRegex("^\\]");//rightsquare
typeDatabase[comma]=makeRegex("^,");//comma
typeDatabase[semiColon]=makeRegex("^;");//semicolon
typeDatabase[colon]=makeRegex("^:");//colon
typeDatabase[assign]=makeRegex("^=");//assign
typeDatabase[plusSign]=makeRegex("^\\+");//plusSign
typeDatabase[star]=makeRegex("^\\*");//star
typeDatabase[dash]=makeRegex("^\\-");//dash
typeDatabase[forwardSlash]=makeRegex("^/");//forwardslash
typeDatabase[lessThan]=makeRegex("^<");//lessThan
typeDatabase[lessThanEqual]=makeRegex("^<=");//lessThanEqual
typeDatabase[greaterThan]=makeRegex("^>");//greaterThan
typeDatabase[greaterThanEqual]=makeRegex("^>=");//greaterThanEqual
typeDatabase[equalsEquals]=makeRegex("^==");//equalsEquals
typeDatabase[notEquals]=makeRegex("^!=");//notEquals
typeDatabase[andOp]=makeRegex("^&&");//andOp
typeDatabase[orOp]=makeRegex("^\\|\\|");//orOp
typeDatabase[notOp]=makeRegex("^!");//notOp
}
int main(int argc, char **argv) {
char *text = readInput(argc, argv) ;
// If reading in input failed, exit with return code of 1.
if (text==NULL) {
return 1 ;
}
// Create the compiled regular expressions.
regex_t whiteSpace ;
makeRegex (&whiteSpace, "^[\n\t\r ]+") ;
regex_t blockComment ;
makeRegex (&blockComment, "^/\\*([^\\*]|\\*+[^\\*/])*\\*+/");
regex_t lineComment ;
makeRegex (&lineComment, "^//[^\n]*\n");
regex_t word ;
makeRegex (&word, "^([a-zA-Z]+)") ;
regex_t integerConst ;
//makeRegex (&integerConst, "^[0-9]+") ;
// modified to include the count of floating point numbers
// somehow using / instead of [] complains, ^ is needed to show front
makeRegex (&integerConst, "^[0-9]*[.]*[0-9]+");
//Add: Boot Regex is not ^[Boot]+ as that matches nonwords
regex_t boot ;
makeRegex (&boot, "^(Boot)") ;
/* This enumerated type is used to keep track of what kind of
construct was matched.
*/
enum MatchType { numMatch, wordMatch, bootMatch, noMatch } matchType ;
int numMatchedChars = 0 ;
// Consume leading white space and comments
numMatchedChars = consumeWhiteSpaceAndComments (&whiteSpace, &blockComment, &lineComment,
text) ;
/* text is a character pointer that points to the current
beginning of the array of characters in the input. Adding an
integer value to it advances the pointer that many elements in
the array. Thus, text is increased so that it points to the
current location in the input.
*/
text = text + numMatchedChars ;
int maxNumMatchedChars = 0 ;
int numWords = 0, numNumericConsts = 0 ;
int numBoot = 0;
while ( text[0] != '\0' ) {
maxNumMatchedChars = 0 ; matchType = noMatch ;
/* maxNumMatchedChars is used to ensure that the regular
expression that matched the longest string is the one that
we use.
The regexs for word and integerConst cannot match the
same text, but if we extend this program to search for
specific keywords, then the keyword regex and the
word-regex may, in some cases, match the same input text.
If two regexs match the same number of characters
then the tie has to be broken. To break the tie, priority
is given to the first one that was tried. Thus the
comparison
(numMatchedChars > maxNumMatchedChars)
is strictly greater than. Not greater than or equal to.
*/
// Add: BootMatch, which has to be above the normal wordMatch
// this is more important than word as it has to take precedence over the word regex
numMatchedChars = matchRegex (&boot, text) ;
if (numMatchedChars > maxNumMatchedChars) {
maxNumMatchedChars = numMatchedChars ;
matchType = bootMatch ;
}
// Try to match a word
numMatchedChars = matchRegex (&word, text) ;
if (numMatchedChars > maxNumMatchedChars) {
maxNumMatchedChars = numMatchedChars ;
matchType = wordMatch ;
}
// Try to match an integer constant
numMatchedChars = matchRegex (&integerConst, text) ;
if (numMatchedChars > maxNumMatchedChars) {
maxNumMatchedChars = numMatchedChars ;
matchType = numMatch ;
}
switch (matchType)
{
case bootMatch: ++numBoot; break;
case wordMatch: ++numWords; break;
case numMatch: ++numNumericConsts; break;
case noMatch: ;
}
if (matchType == noMatch) {
// If we didn't match anything, then just skip the first character.
text = text + 1 ;
}
else {
// Consume the characters that were matched.
text = text + maxNumMatchedChars ;
}
// Consume white space and comments before trying again for
// another word or integer.
numMatchedChars = consumeWhiteSpaceAndComments (&whiteSpace, &blockComment, &lineComment,
text) ;
text = text + numMatchedChars ;
}
/* In this application the only information we collect is the
number of words and number of integer constants. In a scanner
we would need to accumulate the list of tokens. */
printf ("%d\n", numWords) ;
printf ("%d\n", numNumericConsts) ;
printf ("%d\n", numBoot) ;
/* You will add another printf statement to print the number of
"John" keywords. All of these numbers should be on separate
lines. In assessing your work we will require that your output
exactly match ours: no extra spaces and each number on a
separate line. */
}