Re: lynx-dev lynx and tables, looking for more information...

[Klaus Weide]

On Wed, 23 Feb 2000, Duncan Simpson wrote:

> > > I have some code that does text mode table support (in word2x) and think 
> > > it 
> > > takes a function to render something with a given width into a buffer. 
> > > Lines 
> > > of text seperated by newlines would be appropiate.
> > 
> > This is not the real problem, though.  The real problems are the need to
> > parse the whole table to know the widths of the columns and the need to 
> > cope with tables that cannot sensibly be displayed in 80 columns.
> > 
> 
> FYI the table layout code I have can cope with width problems already. All it 
> needs is input for each column width appropiate line breaks inserted to stuff 
> it into an appropiate width, which it calculates automatically. 

One big problem I see with integrating this into the Lynx code (at least
with my understanding of where it would be done): If your algorithm is to be
fed the "text" as it is now fed from HTML.c functions to GridText.c functions,
then it would have to be extended to not only take text characters, but also
"events" like "start (or end) an anchor here" and "start an input field here".
IOW, deal with HText_beginAnchor(), HText_beginInput(), etc. in addition
to HText_appendCharacter() etc.

> Given the 
> rendering engine the rest is well covered by the code. Incidently it does use 
> two passes through the table contents, but surely this should not be a huge 
> problem.

It would need some buffering up of all the table contents, without losing any
of the (renderable) contents' markup information.

  Klaus