Re: [docbook-apps] two-column problem

[Bob Stayton <bobs@sco.com>]

On Thu, Jun 19, 2003 at 10:48:45PM +0500, David Tolpin wrote:
> > 
> > Hi Bob,
> > 
> > I had this problem, too. It is a bug in the XSL:FO specification. Only 
> > elemets that are direct childs of fo:flow can span columns.
> > This was discussed here some time ago:
> > 
> > http://sources.redhat.com/ml/docbook-apps/2003-q2/msg00066.html
> > 
> > however, you could try Antenna House XSL Formatter 
> > http://www.antennahouse.com/ which (If I remember correctly) does not 
> > follow this specification bug.
> > 
> 
> Hi,
> 
> while it is said in XSL FO Recommendation, I hardly view it as a bug.
> It is required (and it is done thoughtfully) that blocks with span="all"
> be direct children of a flow. Otherwise, spanning multiple columns
> breaks XSL FO areas model.
> 
> The bug is in XSL FO stylesheets, and, in my opinion, it should be
> fixed. If Antenna House demonstrates a non-compliant behaviour,
> it can be used in the corresponding customization level; but the
> base stylesheets should not rely on it.

Looking at the DocBook XSL stylesheets, this could be
tough to implement.  The stylesheets often nest fo:blocks
as nested DocBook elements are processed.  If a column
span is requested in some nested element (like a table
in a section in a chapter), then all pending blocks would
have to be closed out, a new block started that is a direct
child of the fo:flow, and then start over with the blocks
that follow the table.

Or the stylesheet architecture would have to change
to a much flatter sequence of blocks that are direct
children of the fo:flow.

Or am I not understanding this problem correctly?

Bob Stayton                                 400 Encinal Street
Publications Architect                      Santa Cruz, CA  95060
Technical Publications                      voice: (831) 427-7796
The SCO Group                               fax:   (831) 429-1887
                                            email: bobs@sco.com