[Groff] Re: Available for small pdfmark project help

[Keith Marshall]

On Monday 08 June 2009 01:10:45 address@hidden wrote:
> Here is what I would like to do:
>
> groff -mpdfmark mymacrofile mytitle myfile mybiblio > myfile.ps
>
> I do not us ms macros, nor any other macro package.  I use a
> custom package.

Ok.  No problem with that.  Just take care to complete the .pdfsync 
actions, at the end of each formatting pass; (use an end macro).

> I would like the TOC to appear after the output from mytitle,

Ok.

> though I understand this is appended to the file at the end.

It is, if you use the standard TOC generation macros in ms; since you 
are using your own custom macro package, it's placed wherever you 
have provided for it.

> So I strip N .ps pages of TOC off the end of the postscript file,
> and insert it after mytitle output, before myfile output begins.

Yuk!  That's precisely the sort of collation nightmare you want to 
avoid; (it's more or less what pdfroff does, but it isn't nice).

> Seems to me I could also insert between mytitle and myfile a
> mytoc file which contains all the necessary info to have a
> linkable toc, thus no striping, no inserting required,

Yes, that's the strategy to strive for; consider:

  <input name="mymaster">
  .so mymacrofile
  .so mytitle
  .so mytoc
  .open toc mytoc
  .so myfile
  .so mybiblio
  .close toc
  </input>

  <input name="myfile">
  ...
  .write toc markup to create TOC entry
  ...
  .write toc markup for another TOC entry
  ...
  </input>

  $ touch mytoc
  $ groff -mpdfmark mymaster > myfile.ps

> perhaps a multiple pass to gather the necessary data.

At least three passes are required; (page numbers will change between 
first and second passes, so a correctly generated mytoc will not be 
available until the third pass begins, at the earliest).

> I see pdfroff uses awk to find .pdfhref D -N lines and saves
> this to a file.

Uhm; not quite!  The pdfmark macros emit some cryptically formatted 
information on standard error.  pdfroff captures this, and filters 
out particular patterns.  It uses one class of such patterns as a 
basis from which it *creates* ".pdfhref D" markup; this represents 
the exported reference dictionary, derived from ".pdfhref M -X" 
markup in the original document source.

> I see pdfroff uses awk again to find .pdfhref Z $2 $3 $4 and
> appends this to a file. 

Again, not quite.  Once more, it *creates* the ".pdfhref Z" markup, 
on the basis of grohtml patterns filtered from standard error; these 
represent the page co-ordinates of any ".pdfhref L" markup in the 
source.

> Question remains, how do I create this auxiliary file you speak
> of that contains the info necessary to create functional toc
> links ?!?

You can use groff's open, write and close requests, as indicated in 
the outline above.

> And what ought this file look like ?!?

Plain text, containing troff markup; likely no more than something 
along the lines of:

  .pdfhref L -D <tag1> -- "Introduction\0\0\a\t7"
  .pdfhref L -D <tag2> -- "Another Heading\0\0\a\t9"
  .pdfhref L -D <tag3> -- ...

> I prefer to stay within a perl environment to do this if possible.

Sorry.  Can't help you there.  What's wrong with staying within a 
*groff* environment?  I've attached a couple of toy document samples 
which illustrate the principles.  They are based on ms macros, but 
actually use only SH to create the section headings, (and LP for the 
following paragraphs, but this plays no part in the TOC generation);
you will need to adapt them, to work with your own macro set.

-- 

Regards,
Keith.

toc-example-1.ms.zip
Description: Zip archive

toc-example-2.ms.zip
Description: Zip archive