Statistics on sgmltag.py
------------------------

Unverified.

Program size: 141/522 code lines,  27.0%.

================================================================
Syntax errors
----------------------------------------------------------------
S1  
----------------------------------------------------------------


================================================================
Bugs that would be caught in a more strongly-typed language
----------------------------------------------------------------
T1  In SGMLTag.__init__(), this line:
        if  self.attrList:
    should be:
        if  attrList:
----------------------------------------------------------------

================================================================
Logic bugs
----------------------------------------------------------------
B1  SGMLTag.__str__():  Failed to produce opening "<"
----------------------------------------------------------------
B2  Tag_Attr_Scan_Value():  Wrong rule for unquoted strings.  It
    is looking for a string starting with giStartCset, but that
    includes only letters.  Loose HTML allows numbers to be
    unquoted as well, e.g., "<td colspan=8>".  So these lines:
        pastFirst  =  scan.any ( giStartCset )
        if  pastFirst: ...
    should be:
        pastFirst  =  scan.any ( giCset )
        if pastFirst is not None: ...
    Also, there is no `else' case for this test:
            pastOpen  =  scan.match ( '"' )
            if  pastOpen:
    But there should be an else case to issue an error if the
    value is neither alphanumeric nor a double-quote.
----------------------------------------------------------------
B3  In SGML_Tag_Scan_Comment(), this line crashes because `gi'
    is undefined:
        return SGMLTag ( gi, 0, None, string.join ( L, "" ) )
    It should be:
        return SGMLTag ( SGML_COMMENT_GI, 0, None, string.join ( L, "" ) )
----------------------------------------------------------------
B4  In SGMLTAG.__str__(), the initial "--" is missing from the
    reconstituted string.  This is because the .gi member is
    only "!", not "!--".  Change:
        if  self.gi == SGML_COMMENT_GI:
            return "<%s%s%s" % ( self.gi, self.text, SGML_COMMENT_TAIL )
    to:
        if  self.gi == SGML_COMMENT_GI:
            return "<%s%s%s" % ( SGML_COMMENT_HEAD,
                                 self.text, SGML_COMMENT_TAIL )
----------------------------------------------------------------
B5  In SGML_Tag_Scan_Comment(), the scanning logic does not move
    past the closing "-->", which causes it to be treated as
    ordinary text.  With this bug and B4, a comment of the form:
        <!-- text -->
    will be reconstituted as:
        <! text -->-->
    Change prime 2.2 from:
        #-- 2.2 --
        # [ if endPos is None ->
        #     scan  :=   scan advanced to the start of the next line
        #     L     +:=  (remainder of the current line) + "\n"
        #   else ->
        #     scan  :=   scan advanced up to endPos
        #     L     +:=  text from scan up to endPos
        #     done  :=   1 ]
        if  endPos is not None:
            L.append ( scan.tab ( endPos ) )
            done  =  1
        else:
            ...
    to:
        #-- 2.2 --
        # [ if endPos is None ->
        #     scan  :=   scan advanced to the start of the next line
        #     L     +:=  (remainder of the current line) + "\n"
        #   else ->
        #     scan  :=   scan advanced up to endPos, plus the length
        #                of SGML_COMMENT_TAIL
        #     L     +:=  text from scan up to endPos
        #     done  :=   1 ]
        if  endPos is not None:
            L.append ( scan.tab ( endPos ) )
            scan.move ( len ( SGML_COMMENT_TAIL ) )
            done  =  1
        else:
            ...
----------------------------------------------------------------
B6  In Tag_Attr_Scan_Value, if the value starts with a character
    that is not a double-quote and not in giCset, the error
    message is issued correctly ("Tag attribute values must be
    alphanumeric or enclosed in double-quotes"), but then it
    falls through to prime 2 and then prime 3, where this line
        return ( value, isQuoted )
    fails because `value' has never been set.  The fix is to
    add the return statement in this else clause:
        else:
            scan.error ( "Tag attribute values must be alphanumeric "
                         "or enclosed in double-quotes ('\"')." )
            return (None, 0)
----------------------------------------------------------------