Following is the changes:
* add ``int eline,echar,ecolumn`` and ``int open`` in class SGMLatom
int eline,echar,ecolumn; indicate the end position of current SGMLatom,
src[char..echar-1] got the source of current SGMLatom
int open; indicate whether the current tag has a matched end tag
* add ``function name_formater`` ``function argname_formater`` in SGML which can
be lower_case or upper_case etc.
the current existing htmls in internet is case-insensitive, so if we want to
work good with them, SGML should be case-insensitive
* recognize xml-compatible tag such as <p/>, set open=0 for these tags
When parse large file, the returned tree of original SGML maybe too deep,
cause stack overflow, recognizing xml-compatible tag can help a bit.
|