lynx-dev table missing fix

[Ilya Zakharevich]

About "disabling nesting tables" - IMO just remove all this code.  it
is a complete junk anyway...

I have no idea how it started; my conjecture is that Thomas just
applied one of my patches with #ifdef's and if's around it.  Since a
lot of code was added in these years, the functionality with the
"disabled" branch may work due to coincidence only.

Thomas never explained what should be the *semantic* of this switch.

There is an *official* way to change the logic when one gives up and
stops processing the table-formatting information.  And you do not
need a forest of if's surrounded by ifdef's to do this - just change
one variable (or a macro), NO_AGGRESSIVE_NEWROW.  One can easily
design other situations when to give up.

Hope this helps,
Ilya

P.S.  I also noticed the situation when some lines in a table "leak to the
      left" (indent disappears).  Will try to investigate it.

Aha, a simple guess from the visual appearence: first several lines
fit the width; so they are typeset with 0 indent (in the expectation
that they will be a part of a table cell anyway).  Then too-long a
line appears which wraps - now the table logic (wrongly) assumes that
it cannot fit this into a table, and gives up; now the rest is set up
as if it were not in a table.

2 bugs: first of all, the cells should have been indented again.
Second, there is no actual need to give up.

A food for thought what is the best compromise between several easy to
implement, but visually inconvenient behaviours...

; To UNSUBSCRIBE: Send "unsubscribe lynx-dev" to address@hidden