> Hi,
>
> I've worked out a code that can parse XML match equity tables w/o
using
> libxml. My code uses glib's g_markup_parser instead.
>
> However, I'm a bit in doubt if I should commit this code to the cvs.
>
> Pros: libxml2 becomes redundant and we can have a fully working
> installation independent of libxml2. Less code. Another use of oldList
> linked list removed from the code. (I dream that all oldList structures
> will be removed from the code one day).
>
> Cons: libxml2 works. It has a good error checking and the system can
> cope with strange files which comes with values first and info last,
or
> any strange order. The error checking is probably better in the existing
> system than with my g_markup based code.
>
> Should I commit my code? Is the independence from libxml2 worth the
> possible troubles we may get with ill-ordered xml met-tables. (Such
an
> ill-ordered met table will probably never occur, but who knows?)