Re: Producing postscript for custom paper size (9x12" manuscript paper)?

[John Hawkinson]

Han-Wen Nienhuys <address@hidden> wrote on Tue,  4 Apr 2006
at 01:03:32 +0200 in <address@hidden>:

> >     Is there some correct way for me to extend (write-preamble)
> >     without editing framework-ps.scm? Is there some sort of
> >     hook I can use in GUILE to modify the existing function
> >     definition to add my extra stuff?
> 
> No,  I don't think so.

OK. I guess it is just my emacs mentality that everything written in
an extension language ought to be extendable. I'm not sure this makes
as much sense for LilyPond as other things.

> We should add support for arbitrary paper sizes, but dealing with
> actual printer specifics is best left outside of lilypond.

Err, I'm not sure I follow.

If LilyPond is generating PostScript output, it needs to specify the
size of the rendered page in the PostScript. Otherwise the PostScript
is incomplete and cannot be properly processed by printers and by the
PDF generator both (this is why the LilyPond-produced PDF of a 9x12
page runs off the edge when viewed in Acrobat Reader). They can only
guess a paper size, either by their own default, or based on the
%%DocumentPaperSizes: DSC comment, which is just an advisory comment
(and, actually, DocumentPaperSizes is deprecated; instead LilyPond
should be generating a %%DocumentMedia comment, which can include
actual page dimensions) and won't actually help with a printer (though
it might help with Acrobat Distiller or Ghostscript).

I don't think there were any printer-specifics in the code in my
email.  It should work for any compliant Level 2 or Level 3 Postscript
printer.  (It *is* true that printers can be finicky about this
media-selection stuff, and there is space for printers to be variant
in their PostScript implementations.)

> FWIW, I usually print through Acrobat reader if I need any strange
> paper fitting done.

This doesn't seem like a good solution if you want to use paper size
that is not a linear multiple of an existing page size; part of the
point of using 9x12 paper is to make use of the extra dimension, so if
you were forced to treat it as 9x11.5", it is not very satisfying. It
also means that LilyPond's ideas of sizing aren't carried through to
the final output, which seems wrong, but less critical.

Is there a reason you don't want to output the /PageSize info?
I assumed that this in output-ps.scm:

     11 ;;;; TODO:
     12 ;;;;   * %% Papersize in (header ...)

was talking about this.

Thanks!

address@hidden
  John Hawkinson