More table borders

[Jeremias Maerki <de...@greenmail.ch>]

I've just updated [1] with all break possibilities for the third example
[2] I mentioned in an earlier post [3].

[1] http://people.apache.org/~jeremias/fop/KnuthBoxesForTablesWithBorders.pdf
[2] http://people.apache.org/~jeremias/fop/table-border-special1.fo
[3] http://mail-archives.apache.org/mod_mbox/xmlgraphics-fop-dev/200504.mbox/%3c20050410171533.C5A5.DEV.JEREMIAS@greenmail.ch%3e

I'm now going to refine Simon's work [4] on borders a little bit further
to handle different border widths for individual cells. Hopefully, I can
finally start implementing the missing pieces after that.

[4] http://wiki.apache.org/xmlgraphics-fop/TableLayout/KnuthElementsForTables/RowBorder

I'm a bit concerned about the increasing complexity of this. It takes
considerable time just to understand and play through all these examples
and after that implementing it so that the code is still readable. Of
course, we documented our findings pretty well, although a newbie will
surely be totally lost. I just wonder if this is still maintainable. And
to make it even worse, I've already documented a known problem [5] (at
the bottom). I just hope it won't bite us later on.

[5] http://wiki.apache.org/xmlgraphics-fop/TableLayout/KnuthElementsForTables

Jeremias Maerki

Re: More table borders

[Jeremias Maerki <de...@greenmail.ch>]

My latest example is now fully explored and seems to work. It took me
almost all day, but at least I got some valuable insight for the
extended algorithm I need to write as the next step. I hope we now have
everything together and I don't get any surprises.

http://wiki.apache.org/xmlgraphics-fop/TableLayout/KnuthElementsForTables/RowBorder2

The nasty problem now is that I have to write/research the FOP session
for ApacheCon EU until Sunday which kills much of the drive I won back
today by running this example. Last Friday the message was sent that the
session papers are due 2005-05-01 and I only saw that message yesterday.
Kills my whole week's plan. And all that while not really feeling well.
Oh well.... :-(

BTW I hope you guys don't mind me keeping away from the break conditions
thread. Not enough brain capacity right now.

On 26.04.2005 11:05:05 Jeremias Maerki wrote:
> I'm now going to refine Simon's work [4] on borders a little bit further
> to handle different border widths for individual cells. Hopefully, I can
> finally start implementing the missing pieces after that.
> 
> [4] http://wiki.apache.org/xmlgraphics-fop/TableLayout/KnuthElementsForTables/RowBorder



Jeremias Maerki

Re: More table borders

[Jeremias Maerki <de...@greenmail.ch>]

On 26.04.2005 18:44:38 Luca Furini wrote:
> > I've already documented a known problem [5] (at
> > the bottom). I just hope it won't bite us later on.
> >
> > [5] http://wiki.apache.org/xmlgraphics-fop/TableLayout/KnuthElementsForTables
> 
> I don't think it is a dangerous problem: it can be recognized and properly
> handled, as you wrote, discarding the wrong elements after the first inner
> break and collecting the right ones (and this mechanism is needed anyway).

But I think that's going to be a project on its own. Discarding objects
means there is overhead which needs to be minimized. And doing that well
will not be so easy. I've done some investigation there and I don't want
to think about it right now.

> It is such an unlikely situation that I don't think we should worry about
> it: I really can't think of a real example of table with a monstrous row
> that must be broken more than once!! I think it is already quite unlikely
> to allow a page break in the middle of a row, which would force the reader
> to go back and forth several times, in order to read each cell from the
> beginning to the end! :-)

What about a blind table for indenting content? Isn't that a frequent
work-around for FOP 0.20.5? Fortunately, people will be able to use
start-indent with the new code.

Jeremias Maerki