/**
Parses URL from a token. Is used by SearchUrlL method and if a URL
was found it's appended to item array. Note that parsing for generic URIs
is done with SearchGenericUriL -method.
@param aType a Type of URL to seach, i.e.
www.
wap.
IP e.g.127.0.0.1
@param aTokenPtr Pointer to token that will be parsed
@param aTextOffset Offset of the token (start position in the whole text)
@leave KErrNone, if successful; otherwise one of the other system-wide error codes.
@return ETrue if the parameter for phone number is valid, else returns EFalse
*/
TBool CTulAddressStringTokenizer::ParseUrlL(const TDesC& aType, const TPtrC& aTokenPtr, TInt aTextOffset)
{
TBool wasValidUrl = EFalse;
TLex url;
TInt position = aTokenPtr.FindF( aType );
if ( position != KErrNotFound )
{ // address start found
url = aTokenPtr.Right( aTokenPtr.Length() - position );
url.Inc( aType.Length() );
while( IsValidUrlChar( url.Peek() ) && !(url.Eos()) )
{
if( url.Peek() == ':' )
{
url.Inc();
if ( !url.Peek().IsDigit() )
{
url.UnGet();
break;
}
}
else
url.Inc();
}
// If a period or question mark was followed by a whitespace remove it
if ( url.Eos() ) // Can't be followed by white space if it's
{ // the last character at token
url.UnGet();
if ( url.Peek() != '.' && url.Peek() != '?' && url.Peek() != ',' ) // If it wasn't a period or question mark
url.Inc();
}
url.Mark();
wasValidUrl = ETrue;
}
if ( wasValidUrl && ( url.MarkedOffset() > aType.Length() ) )
{
AddItemL( aTextOffset - aTokenPtr.Length() + position, url.MarkedOffset(), EFindItemSearchURLBin );
return ETrue;
}
return EFalse;
}
TPtrC CScriptFile::ParseValue(const TDesC& aText, const TLex& aInput, TInt aCurrentItemStart) const
{
TInt mid = aCurrentItemStart;
TInt len = aInput.MarkedOffset() - mid;
TPtrC ret(KNullDesC);
if (len > 0)
ret.Set(aText.Mid(mid, len));
return ret;
}
/**
Search algorithm for searching e-mail addresses
@param aText Text that will be parsed
@return ETrue if any EMail items were found else returns EFalse
@leave KErrNone, if successful; otherwise one of the other system-wide error codes.
@panic ETulPanicDescriptorLength in debug build if item's position
and/or length is out of the document's range.
*/
TBool CTulAddressStringTokenizer::SearchMailAddressL( const TDesC& aText )
{
TInt searchStart = 0;
TInt searchResult = 0;
const TInt end = aText.Length(); // end of document
do
{
TPtrC segment = aText.Right( end - searchStart );
searchResult = segment.LocateF('@');
if (searchResult != KErrNotFound)
{ // @ found
// There should be valid characters (not a period) before and after the @ character
if ( searchResult == 0 // first char
|| (searchResult >= segment.Length() - 1) // last char
|| !(IsValidEmailChar(segment[searchResult - 1]))
|| !(IsValidEmailHostChar(segment[searchResult + 1]))
|| segment[searchResult - 1] == '.'
|| segment[searchResult + 1] == '.'
)
{
searchStart += searchResult + 1;
continue;
}
TBool wasPeriod = EFalse; // To prevent sequential periods
// Get TLex from the pointer to get a better API for parsing
TLexMark startPos;
TLexMark endPos;
TLex token = segment;
// Go to searchResult and un-get until the beginning of e-mail address is reached
token.Inc( searchResult );
token.Mark();
do
{
token.UnGet();
if ( token.Peek() == '.' )
{ // If it was a period
if (wasPeriod) // and if the former was also -> break
break;
else // else mark that this one was a period
wasPeriod = ETrue;
}
else
wasPeriod = EFalse;
}
while (token.Offset() > 0 && IsValidEmailChar(token.Peek()));
if (token.Offset() != 0 || !IsValidEmailChar(token.Peek()))
token.Inc();
// Get rid of periods from the start of address
// Does it have to start with a number or char(abc...).
// If it does, the loop should check that it gets rid of all special chars also.
while (token.Peek() == '.')
token.Inc();
token.Mark( startPos ); // Mark the beginning of address
token.UnGetToMark();
wasPeriod = EFalse;
do // Go forward until a nonvalid character
{
token.Inc();
if ( token.Peek() == '.' )
{ // If it was a period
if ( wasPeriod ) // and if the former was also -> break
break;
else // else mark that this one was a period
wasPeriod = ETrue;
}
else
wasPeriod = EFalse;
}
while ( !token.Eos() && IsValidEmailHostChar( token.Peek() ) );
// If address ends with a period take it away
token.UnGet();
if (token.Peek() != '.')
token.Inc();
token.Mark( endPos ); // Mark the beginning of address
// Append the found string to the array
__ASSERT_DEBUG( searchStart + token.MarkedOffset( startPos )
+ token.MarkedOffset( endPos )
- token.MarkedOffset( startPos ) <= aText.Length(),
Panic(ETulPanicDescriptorLength) );
AddItemL( searchStart + token.MarkedOffset( startPos ),
token.MarkedOffset( endPos ) - token.MarkedOffset( startPos ),
EFindItemSearchMailAddressBin);
searchStart += token.MarkedOffset( endPos ) + 1;
}
}
while ( searchResult != KErrNotFound && searchStart < end );
return (iFoundItems->Count() > 0);
}
/**
Search algorithm for searching phone numbers
@param aText Text that will be parsed
@return ETrue if any Phone Number items were found else returns EFalse
@leave KErrNone, if successful; otherwise one of the other system-wide error codes.
@panic ETulPanicDescriptorLength in debug build if item's position
and/or length is out of the document's range.
*/
TBool CTulAddressStringTokenizer::SearchPhoneNumberL( const TDesC& aText )
{
TLexMark startMark; // Points to the start of the found phone number
TLexMark endMark; // Points to the end of the found phone number
TLexMark mark;
const TInt end = aText.Length();
TLex number = aText;
while ( !(number.Eos()) )
{
TInt numberCount = 0; // How many real numbers (1234567890)
TInt bracketsOpen = 0; // How many brackets are currently open
TInt brackets = 0; // How many brackets overall
TChar charac = number.Peek();
while( (!(IsValidPhoneNumberChar( charac ) || charac == '+'
|| charac == '(' ) || charac == '-' || charac == '.' || charac == '/')
&& !(number.Eos()) && number.Offset() < end )
{
number.Inc();
charac = number.Peek();
}
if ( number.Offset() >= end )
break;
if ( number.Peek() == '#' )
{
number.Inc();
if (number.Peek() == '.' )
continue;
number.UnGet();
}
if ( number.Peek() == '+' )
{ // '+' has to be followed by a number (not # or * ...)
number.Inc();
if ( !(number.Peek().IsDigit()) )
continue;
number.UnGet();
}
if ( number.Peek() == '(' )
{ // '(' has to be followed by valid phone number
// character (whitespaces are allowed before) or '+' is a next character
number.Inc();
if ( !(number.Peek() == '+') )
{
number.Mark(mark);
number.SkipSpace();
charac = number.Peek();
if ( !( IsValidPhoneNumberChar(charac) || charac == '+'
|| charac == '(' ) || charac == '-' || charac == '.' || charac == '/')
{
number.Inc();
continue;
}
else
{
number.UnGetToMark(mark);
number.UnGet();
number.Mark(startMark);
}
}
else
{
number.UnGet();
number.Mark(startMark);
number.Inc();
}
bracketsOpen++;
brackets++;
}
else
number.Mark(startMark);
if ( number.Peek().IsDigit() ) // If the character was a number
numberCount++;
else if ( bracketsOpen > 0 )
{
number.Inc();
TChar next = number.Peek();
TInt bracketsOpen2 = bracketsOpen;
while( (IsValidPhoneNumberChar( next ) || next.IsSpace()
|| next == '(' || next == ')' || next == 'p' || next == '+'
|| next == 'w' ) && !(number.Eos()) && number.Offset() < end)
{
if ( next == '(' )
bracketsOpen2++;
else if ( next == ')' )
bracketsOpen2--;
if ( bracketsOpen2 == 0 )
break;
number.Inc();
next = number.Peek();
}
number.UnGetToMark(startMark);
if ( bracketsOpen2 != 0 )
{
number.Inc();
continue;
}
}
number.Inc();
while ( number.Peek() == '(' && !(number.Eos()) && bracketsOpen > 0 )
{
number.Inc();
bracketsOpen++;
}
if ( number.Peek() == '+' && bracketsOpen > 0 )
number.Inc();
// a Valid first character has been found. Let's go forward as long as valid characters are found.
charac = number.Peek();
while( (IsValidPhoneNumberChar( charac ) || charac.IsSpace()
|| charac == '(' || charac == ')' || charac == 'p'
|| charac == 'w' ) && !(number.Eos()) && number.Offset() < end
&& charac != KCharLinefeed && charac != KCharFormfeed
&& charac != KCharCarriageReturn
&& charac != KCharLineSeparator
&& charac != KCharParagraphSeparator )
{
if ( number.Peek() == '(' )
{ // '(' can't be the last character in phone number
number.Mark(mark);
number.Inc();
TChar spaceJump = number.Peek();
while ( !number.Eos() && spaceJump.IsSpace() && spaceJump != KCharLinefeed
&& spaceJump != KCharFormfeed && spaceJump != KCharCarriageReturn
&& charac != KCharLineSeparator && spaceJump != KCharParagraphSeparator)
{
number.Inc();
spaceJump = number.Peek();
}
if ( !(IsValidPhoneNumberChar(number.Peek())) && number.Peek() != ')'
&& number.Peek() != '(' )
{
number.UnGetToMark(mark);
break;
}
TChar next = number.Peek();
TInt bracketsOpen2 = bracketsOpen + 1;
while( (IsValidPhoneNumberChar( next ) || next.IsSpace()
|| next == '(' || next == ')' || next == 'p'
|| next == 'w' ) && !(number.Eos()) && number.Offset() < end)
{
if ( next == '(' )
bracketsOpen2++;
else if ( next == ')' )
bracketsOpen2--;
if ( bracketsOpen2 == 0 )
break;
number.Inc();
next = number.Peek();
}
number.UnGetToMark(mark);
if ( bracketsOpen2 != 0 )
break;
bracketsOpen++;
brackets++;
}
else if ( number.Peek() == ')' )
{
if ( bracketsOpen <= 0 ) // there has to be equal number of brackets
break;
bracketsOpen--;
number.Mark(mark);
number.Inc();
if ( number.Peek() == '.' ) // '.' is not allowed after ')'
break;
number.UnGetToMark(mark);
}
else if ( number.Peek() == '-' || number.Peek() == 'w'
|| number.Peek() == 'p' || number.Peek() == '.' || number.Peek() == '/')
{ // Hyphen mark and 'p' & 'w' chars must be followed by a number
TChar last = number.Peek();
number.Mark(mark);
number.Inc();
TChar spaceJump = number.Peek();
while ( !number.Eos() && spaceJump.IsSpace() && spaceJump != KCharLinefeed
&& spaceJump != KCharFormfeed && spaceJump != KCharCarriageReturn
&& charac != KCharLineSeparator && spaceJump != KCharParagraphSeparator )
{
number.Inc();
spaceJump = number.Peek();
}
if ( !(number.Peek().IsDigit()) )
{
if (last == '.' && number.Peek() == ')' && bracketsOpen > 0 )
continue;
else
{
number.UnGetToMark(mark);
break;
}
}
number.UnGetToMark(mark);
}
else if ( number.Peek().IsDigit() )
numberCount++;
number.Inc();
charac = number.Peek();
}
// Get rid of whitespaces from the end
number.UnGet();
while( number.Peek().IsSpace() && !(number.Eos()))
number.UnGet();
number.Inc();
// ------------------------------------
number.Mark(endMark);
// If they exist, remove brackets from the beginning and the end
number.Mark(mark); // Let's mark where to continue the search
TBool endBrackets = ETrue;
do
{
number.UnGet();
if ( number.Peek() == ')' )
{
number.UnGetToMark(startMark);
if ( number.Peek() == '(' )
{
// If there's more than one pair of brackets -> don't strip them.
if ( brackets > 1 )
break;
number.Inc();
number.Mark(startMark);
number.UnGetToMark(endMark);
number.UnGet();
number.Mark(endMark);
// Get rid of whitespaces and periods from the end and from the beginning
number.UnGet();
while ( (number.Peek().IsSpace() || number.Peek() == '.')
&& number.Offset() > number.MarkedOffset(startMark) )
{ // from the end
number.UnGet();
}
number.Inc();
number.Mark(endMark);
number.UnGetToMark(startMark);
while ( (number.Peek().IsSpace() || number.Peek() == '.')
&& number.Offset() < number.MarkedOffset(endMark) )
{ // from the beginning
number.Inc();
}
number.Mark(startMark);
number.UnGetToMark(endMark);
// ----
}
else
endBrackets = EFalse;
}
else
endBrackets = EFalse;
}
while ( endBrackets );
number.UnGetToMark(mark);
// ----------------
if ( numberCount <= KFindItemMaxNumbers && numberCount >= iMinNumbers )
{
TPtrC tokenPtr = number.MarkedToken(startMark);
TInt tokensEnd = tokenPtr.Length();
TInt numbers = 0;
TInt partialNumber = 0;
TBool wasValidPhoneNumber = ETrue;
TInt i = 0;
for ( ; i < tokensEnd; i++ )
{
if ( tokenPtr[i] == '.' )
partialNumber = 0;
else if ( ((TChar)tokenPtr[i]).IsDigit() )
{
numbers++;
partialNumber++;
}
if ( ( partialNumber == 1 || partialNumber == 2 ) && i + 1 < tokensEnd )
{
if ( tokenPtr[i + 1] == '.' )
wasValidPhoneNumber = EFalse;
}
}
if (!wasValidPhoneNumber && numbers > 6)
wasValidPhoneNumber = ETrue;
if (wasValidPhoneNumber)
{
__ASSERT_DEBUG( number.MarkedOffset(startMark) + number.MarkedOffset(endMark)
- number.MarkedOffset(startMark) <= aText.Length(),
Panic(ETulPanicDescriptorLength) );
AddItemL( number.MarkedOffset(startMark),
number.MarkedOffset(endMark) - number.MarkedOffset(startMark),
EFindItemSearchPhoneNumberBin );
}
}
}
return (iFoundItems->Count() > 0);
}
/**
Search algorithm for searching generic URIs
@param aText Text that will be parsed
@return ETrue if any generic URI is found else returns EFalse
@leave KErrNone, if successful; otherwise one of the other system-wide error codes.
*/
TBool CTulAddressStringTokenizer::SearchGenericUriL( const TDesC& aText )
{
// Detect generic URI within the token
const TDesC& schemeStartArray = KURISchemeStartCharacters;
const TDesC& schemeBodyArray = KURISchemeBodyCharacters;
const TDesC& schemeTerminatorArray = KURISchemeTerminator;
const TDesC& URIArray = KURICharacters;
TBool wasValidUri = EFalse;
TLex text = aText;
while ( !text.Eos() )
{
// Discard characters until URI scheme terminator is found
while( !(text.Eos()) && schemeTerminatorArray.Locate(text.Peek()) == KErrNotFound )
text.Inc();
// if at end of the text, no legit URI found
if ( !text.Eos() )
{
// Store the schema end offset (+1 to skip ':')
TInt schemeEndOffset = text.Offset() + 1;
// Scheme must be at least 1 character long at the beginning of the text to be valid
if ( text.Offset() > 0 )
{
// Un-get last scheme character to begin examination
text.UnGet();
// Rewind until beginning of the URI
while ( text.Offset() > 0 && schemeBodyArray.Locate(text.Peek().GetLowerCase()) != KErrNotFound )
text.UnGet();
// Now text pointer is at first character of the URI
// Do go back through the scheme until a legal beginning character for URI
// is found or back to the (schemeEndOffset - 1) i.e. URI scheme terminator
while ( schemeStartArray.Locate(text.Peek().GetLowerCase()) == KErrNotFound && (text.Offset() + 1) < schemeEndOffset )
text.Inc();
// check if terminated because a valid start character was found when
// scheme terminator was reached.
if ( schemeStartArray.Locate(text.Peek().GetLowerCase()) != KErrNotFound )
{
// First character is a valid URI char, so the scheme is valid ->
// marks the beginning of the array
text.Mark();
// fast forward to the end of the scheme
while( text.Offset() < schemeEndOffset )
text.Inc();
// Get characters until end of schema
while( !(text.Eos()) && URIArray.Locate( text.Peek().GetLowerCase() ) != KErrNotFound )
text.Inc();
// remove certain punctuation from end of the URI, as it is likely
// to be part of the surrounding text.
text.UnGet();
//special processing for bracket
//only remove the end bracket if there is no open bracket in the uri
//not counting bracket pairs for efficiency
if (text.Peek()!=')' || text.MarkedToken().Locate(TChar('('))!=-1)
text.Inc();
text.UnGet();
if ( text.Peek() != '.' && text.Peek() != '?' && text.Peek() != ',')
text.Inc();
// URI cannot contain only scheme, so check that pointer was increased
// by at least one character
if ( schemeEndOffset != text.Offset() )
{
// Append found text to item array (it is now known to be
// syntactically valid URI as it contains characters after the scheme)
AddItemL( text.MarkedOffset(), text.Offset() - text.MarkedOffset(), EFindItemSearchScheme );
wasValidUri = ETrue;
}
}
else // First character of scheme is not legit, fast forward to end of the
// scheme anyway to continue search
{
while( text.Offset() < schemeEndOffset )
text.Inc();
}
}
else
text.Inc();
}
}
return wasValidUri;
}