Vincent, hello. Naturally disappointed but not at all surprised this would fall in the "hard" category or it certainly would have been taken care of by now I'm sure.
I would like to report that I'm having good luck with inter-mixing one- and two-column layouts, again under the provisio that the single column with matches the two column. For local reasons, I'm not using the "span = 'all'" trick but it did work well even if the text required more space than was available on the first page. I'm sorry to report that the combination of useful progress being made here and the suggested level of difficulty precludes me getting a green light to tackle the problem right now. Cheers, rjs On 11/22/2010 12:40 PM, Vincent Hennebert wrote: > Hi Rob, > > On 18/11/10 15:22, Rob Sargent wrote: > >> Yes, thanks for the confirmation Vincent. How much interest do you feel >> there would be for such support? Level of difficulty? Pointers? >> > You’re hitting the infamous “changing IPD problem”: FOP is not able to > handle pages of different widths. And changing the column count implies > a change of the width according to which text must be broken into lines. > > FOP 1.0 contains a hack that provides limited support for different page > widths, but this doesn’t include the change of column counts. > > There is an interest in fully supporting this. The level of difficult is > hard. Search for “Changing IPD” in archives of the fop-dev mailing list > for some pointers. This also involves understanding Dr Donald E. Knuth’ > approach to typography based on dynamic programming. Mainly, his article > “Breaking Paragraphs Into Lines”. There is some information about the > Knuth algorithm on the FOP wiki: > http://wiki.apache.org/xmlgraphics-fop/DeveloperPages > > > >> Interestingly, it is almost implemented. One does see two columns on >> the second page in my example, it's just that they're the width as >> defined in the opening one-column layout. >> >> Btw, I've mananged to get a mix by having the opening 1-column in a >> 'block span=all' trick posted elsewhere, then a one column of width = >> width of one-of-two-columns. Seems to be working... >> > If you are able to determine, prior to rendering, the amount of content > that will fit on the first page, then this is a good workaround. > > > >> On 11/18/2010 05:04 AM, Vincent Hennebert wrote: >> >>> Hi Rob, >>> >>> While in theory it’s possible to use page masters with different column >>> counts, in practice it’s not supported by FOP yet. >>> >>> Sorry, >>> Vincent >>> >>> >>> On 14/11/10 19:32, Rob Sargent wrote: >>> >>> >>>> Using fop-0.95. Can move forward if needed. >>>> >>>> On 11/14/2010 12:31 PM, Rob Sargent wrote: >>>> >>>> >>>>> Is it possible to change the column count within a single flow with each >>>>> new simple-page master? >>>>> >>>>> >>>>> <fo:page-sequence-master master-name="document-sequence"> >>>>> <!-- one column region-body --> >>>>> <fo:single-page-master-reference master-reference="chapter-open" /> >>>>> <!-- two column region-body --> >>>>> <fo:single-page-master-reference master-reference="all-text-page-2" >>>>> /> >>>>> <!-- one column region-body --> >>>>> <fo:single-page-master-reference >>>>> master-reference="three-side-page-3" /> >>>>> <!-- two column region-body --> >>>>> <fo:single-page-master-reference master-reference="table-page-4" /> >>>>> <!-- two column region-body --> >>>>> <fo:single-page-master-reference master-reference="table-page-5" /> >>>>> <!-- one column region-body --> >>>>> <fo:single-page-master-reference >>>>> master-reference="three-side-page-6" /> >>>>> </fo:page-sequence-master master-name="document-sequence"> >>>>> > > Vincent > > --------------------------------------------------------------------- > To unsubscribe, e-mail: [email protected] > For additional commands, e-mail: [email protected] > > --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
