Re: Rules in Headers and Footers with Jade

[Oisin McGuinness <oisin@xxxxxxxx>]

W. Eliot Kimber said:

> If I use a rule that is displayed in the footer, it works, except that the
> resulting footer is flush with the bottom of the page, which I suspect is a
> side effect of the fact that a displayed rule is actually not a valid flow
> object within a footer (which only allows inline flow objects).

(This is long; summary is: Modifying Jade's header footer treatment in RTF backend
may help you...)

For a current SGML to RTF conversion project, we had many documents (letters)
which had rules specified as belonging in the footer. (No problem in the SGML
application, which is ArborText's "The Publisher"; it uses the infinitely flexible TeX
language as a backend.) I was agreeably surprised to find that the DSSSL code:

(element FMT-BASERULE
        (make rule
                length:                 (display-size)
                display-alignment:      'center
                line-thickness:          0.5pt
                orientation:            'horizontal))

followed by a normal (1 line paragraph) displayed fine in the footers, giving a rule
extending across the page, with a footer. Most time this printed fine, but we had
some problems where it mysteriously moved part way off the page.

>From the point of view of Word's dialogs (Word 7.0a for Win 95, running on NT), and from the point of view of some
Word Basic macros which were acting on the converted documents to replace footer
contents with EPS based logos, the headers and footers generated by Jade looked strange:

a)  in the File -> Page Setup -> Margins dialog, the From Edge settings for Header and Footer
        show up as 0.0", no matter what settings I gave to the header-margin and footer-margin
        characteristics in the style sheet.
b)  in the File -> Page Setup -> Layout dialog, the Headers and Footers checkbox for "Different
        Odd and Even" always showed checked, with Different First Page sometimes checked and not.
        This did not always match the page model for our SGML documents, which (in Publisher language)
        were sometimes Left/Right (Publisher book based), and sometimes First/Succ.
       This confused the Word Basic code: when asked to replace footers on 2nd pages with a logo
        in a document, the 3rd page footers would not get replaced. (Our letter style is 1st page has
       top logo, 2nd and subsequent pages have bottom logos.)
c)  The logo inserted in the footers would appear "cut off" when viewed in Word, since it was
        taller than the 1 line footer it was replacing. (By the way, it appears that Word Basic does not
        let you get at all the features of a Word document which one can get at in the RTF!)

I've dealt with these problems by modifying RtfFOTBuilder.cxx so that:

a) The RTF \headery and \footery variables which were being set to 0 in 
    RtfFOTBuilder::startSimplePageSequence are now set to the headerMargin and footerMargin values.
    In the RtfFOTBuilder::outputHeaderFooter I now omit setting space before (\sb in RTF), and
    space after (\sa in RTF).  Thus the philosophy changes from generating a paragraph with space
    before and after, to generating a header/footer line exactly positioned.

b) In RtfFOTBuilder::endAllSimplePageSequenceHeaderFooter I've added code to check whether
    the document has Left/Right mode, and based on that, maybe don't set \facinpg in ~RtfFOTBuilder
    when it outputs the document settings.  The kind of headers and footers put out has been changed
    also, and my Word developers are happy now. (If you like reading confusing specifications, see the 
    discussion on page 29 of the RTF 1.5 spec for header/footer treatment.) The settings visible in
   the File -> Page Setup -> Layout  now match exactly what we expect the page model to be.

c) Here Jade was setting \sl (line spacing) explicitly in headers and footers. I've turned that off, which
    according to the RTF spec will make Word calculate sizes from what is actually present. And it does
   so, solving my last problem. (That is, replacing footers from macros with objects of different sizes
  results in the right thing.)

Getting back to the rule in footer issue, I believe that your rules may work even better if you change Jade....

Cautions: 1) Our documents just generate 1 simple page sequence. I haven't tested the consequences
                  of these changes when there are several page sequences generated, giving rise to
                   several "sections" in RTF.
                  2) No tests have been done with other RTF readers besides Word 7.0a.
                  3) The Jade version used was jade 1.0.1; there is a small difference in the jadetest release
                       where extra spaces are inserted in the RTF for headers and footers between the pieces.
                  3) The code was finished and tested just yesterday; I haven't gotten James Clark's feedback
                      yet; presumably there were some good reasons for the way the Jade header footer code
                     is the way it is.
    

Diffs to RtfFOTBuilder.cxx are available on request.


Oisin McGuinness

Sumitomo Bank Capital Markets
277 Park Avenue
New York NY 10172
USA

(212)-224-4913, email: oisin@xxxxxxxx


 DSSSList info and archive:  http://www.mulberrytech.com/dsssl/dssslist