/* private */ function Tokenizer()
{
+ global $wgLang;
+
$this->mPos=0;
$this->mTokenQueue=array();
+ $this->linkPrefixExtension = $wgLang->linkPrefixExtension();
}
# factory function
// proceeds character by character through the text, looking for characters needing
// special attention. Those are currently: I, R, ', [, ], newline
//
- // TODO: prefixed links for Arabic wikipedia not implemented yet
- // handling of French blanks not yet implemented
+ // TODO: handling of French blanks not yet implemented
function nextToken()
{
$fname = "Tokenizer::nextToken";
if ( count( $this->mQueuedToken ) != 0 ) {
// still one token from the last round around. Return that one first.
$token = array_shift( $this->mQueuedToken );
- } else if ( $this->mPos > $this->mTextLength )
- { // If no text is left, return "false".
+ } else if ( $this->mPos > $this->mTextLength ) {
+ // If no text is left, return "false".
$token = false;
} else {
$this->mPos += 3;
break 2; // switch + while
} else if ( $this->continues("[") ) {
- $queueToken["type"] = "[[";
+ $queueToken["type"] = "[[";
$queueToken["text"] = "";
+ // Check for a "prefixed link", e.g. Al[[Khazar]]
+ // Mostly for arabic wikipedia
+ if ( $this->linkPrefixExtension ) {
+ while ( $this->linkPrefixExtension
+ && ($len = strlen( $token["text"] ) ) > 0
+ && !ctype_space( $token["text"][$len-1] ) )
+ {
+ //prepend the character to the link's open tag
+ $queueToken["text"] = $token["text"][$len-1] . $queueToken["text"];
+ //remove character from the end of the text token
+ $token["text"] = substr( $token["text"], 0, -1);
+ }
+ }
$this->mQueuedToken[] = $queueToken;
$this->mPos += 2;
break 2; // switch + while
&& $this->mText[$this->mPos+1] == "'" )
{
$queueToken["type"] .= "'";
+ $queueToken["pos"] = $this->mPos;
$this->mPos ++;
}
}
break 2;
}
+ break;
+ case "!": // French spacing rules have a space before exclamation
+ case "?": // and question marks. Those have to become
+ case ":": // And colons, Hashar says ...
+ if ( $this->preceeded( " " ) )
+ {
+ // strip blank from Token
+ $token["text"] = substr( $token["text"], 0, -1 );
+ $queueToken["type"] = "blank";
+ $queueToken["text"] = " {$ch}";
+ $this->mQueuedToken[] = $queueToken;
+ $this->mPos ++;
+ break 2; // switch + while
+ }
+ break;
+ case "0": // A space between two numbers is used to ease reading
+ case "1": // of big numbers, e.g. 1 000 000. Those spaces need
+ case "2": // to be unbreakable
+ case "3":
+ case "4":
+ case "5":
+ case "6":
+ case "7":
+ case "8":
+ case "9":
+ if ( ($this->mTextLength >= $this->mPos +2)
+ && ($this->mText[$this->mPos+1] == " ")
+ && ctype_digit( $this->mText[$this->mPos+2] ) )
+ {
+ $queueToken["type"] = "blank";
+ $queueToken["text"] = $ch . " ";
+ $this->mQueuedToken[] = $queueToken;
+ $this->mPos += 2;
+ break 2; // switch + while
+ }
+ break;
+ case "\302": // first byte of UTF-8 Character Guillemet-left
+ if ( $this->continues( "\253 ") ) // second byte and a blank
+ {
+ $queueToken["type"] = "blank";
+ $queueToken["text"] = "\302\253 ";
+ $this->mQueuedToken[] = $queueToken;
+ $this->mPos += 3;
+ break 2; // switch + while
+ }
+ break;
+ case "\273": //last byte of UTF-8 Character Guillemet-right
+ if ( $this->preceeded( " \302" ) )
+ {
+ $queueToken["type"] = "blank";
+ $queueToken["text"] = " \302\273";
+ $token["text"] = substr( $token["text"], 0, -2 );
+ $this->mQueuedToken[] = $queueToken;
+ $this->mPos ++;
+ break 2; // switch + while
+ }
+ break;
+ case "&": //extensions like <timeline>, since HTML stripping has already been done,
+ //those look like <timeline>
+ if ( $this->continues( "lt;timeline>" ) )
+ {
+ $queueToken["type"] = "<timeline>";
+ $queueToken["text"] = "<timeline>";
+ $this->mQueuedToken[] = $queueToken;
+ $this->mPos += 16;
+ break 2; // switch + while
+ }
+ break;
+
} /* switch */
$token["text"].=$ch;
$this->mPos ++;
// function continues
// checks whether the mText continues with $cont from mPos+1
- function continues( $cont )
+ /* private */ function continues( $cont )
{
// If string is not long enough to contain $cont, return false
if ( $this->mTextLength < $this->mPos + strlen( $cont ) )
}
return true;
}
-
+
+ // function preceeded
+ // checks whether the mText is preceeded by $prec at position mPos
+ /* private */ function preceeded( $prec )
+ {
+ $len = strlen( $prec );
+ // if $prec is longer than the text up to mPos, return false
+ if ( $this->mPos < $len )
+ return false;
+ return ( 0 == strcmp( $prec, substr($this->mText, $this->mPos-$len, $len) ) );
+ }
+
+ function readAllUntil( $border )
+ {
+ $n = strpos( $this->mText, $border, $this->mPos );
+ if ( $n === false )
+ return "";
+ $ret = substr( $this->mText, $this->mPos, $n - $this->mPos );
+ $this->mPos = $n + strlen( $border ) + 1;
+ return $ret;
+ }
+
}