xmlgraphics-fop-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Keiron Liddle" <kei...@aftexsw.com>
Subject Re: How layout works
Date Mon, 28 Apr 2003 23:39:41 GMT
> Keiron Liddle wrote:
> > I suppose you could say that it hasn't really been documented because it 
> > is quite difficult to get a good picture on a layout that will handle 
> > all that it needs to.
> 
> I don't think the problems are cause only by the complexity of the
> layout algorithms itself, although there are a few really unpleasant
> things to consider. The old code was quite powerful and could have
> been pushed much further (including proper inline areas, side floats
> and keeps, using backtracking), but a lot of people happily hacked
> around it. I believe the problems with the new code have also quite
> a bit to do with how the algorithms are implemented. I tried hard,
> after all I implemented a layout manager, but I still cant wrap my
> mind around the zillions of small hassles the code provides. I don't
> know *why* BreakPoss and Positions are different classes: Will there
> ever be a break possiblity whiuch is not a position, or vice versa?

That is why I had a good look at how to make it better and decided that it can all 
be done with a Position. The BreakPoss is not needed and was actually getting in 
the way for some things.

So a page starts at a Position and ends at a Position. The keeps etc can be found 
directly from this position, the distance down the page can be found from the 
position. So that the Position for a bottom level layout manager is the Position that 
is used at all levels.

Hopefully it will become clearer once I actually get a chance to do something.

> I can't cope with the Iterators. I have difficulties to decide when
> an inheritance provides an is-a relation or is mainly for code reuse,
> and why there are methods with the same name but apparently radically
> different purpose on some LayoutManagers and others. All these stuff
> may be there for some reason, but unless I'm told the reasons I won't
> be able to work efficiently on the layout core, always in fear I'll
> break something really important I just didn't see.
> 
> J.Pietschmann


---------------------------------------------------------------------
To unsubscribe, e-mail: fop-dev-unsubscribe@xml.apache.org
For additional commands, email: fop-dev-help@xml.apache.org


Mime
View raw message