[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: lynx-dev more TRST support
From: |
Klaus Weide |
Subject: |
Re: lynx-dev more TRST support |
Date: |
Tue, 23 Nov 1999 08:40:36 -0600 (CST) |
On Tue, 23 Nov 1999, Leonid Pauzner wrote:
> Yet another item: less strong meaning of "simple table"
>
> In general, "simple table" is one that fit into the screen width.
> So a larger width allow us to render more tables properly, though this
> is a rather expensive way since we could not split cell into several
> lines within TRST model (of cause, multiline cells via <br> or other
> block constrains like <p> will not fit the model also).
Horizontal scrolling might help.
> Looking the real tables I found out there are many cases when the first
> (header) row is too wide while the others rows more compact.
> Is it possible (and useful) to render such tables in a mixed mode, e.g.
> one (or two) worse rows rendered in LTMS mode while the rest rows
> rendered in TRST mode (ignoring the)?
Well that is the kind of smaller change that should be possible. It
is not completely straightforward though - the TRStable.c stuff is
itself very on-pass. In particular, once a <TD> or </TD> (or TH) has
been recorded at a certain column position, that will set the minimum
x value for that column immediately. There is currently no way to
forget all the effects of a line "later", e.g. when it turns out on
the 3rd or 4th TD in a row that the cells are too long.
Well that is changeable (add another level of "buffering"), so it's
not a big problem. I don't know what would be useful criteria for
"which (or how many) rows can be ignored?" (all?). But something
could be found.
Klaus