Re: terminal only for two weeks

On 2025-01-16, Salvador Mirzo wrote:

 > I suspect I imagine wrong how things actually work.  I thought
 > perhaps there would be a command line such as ``lpr --pages 7-14''.

As has already been pointed in this thread, CUPS, a fairly
common choice for a printer spooler in GNU/Linux systems,
provides lp(1) command that does have just such an option.

 > Now I believe a program like evince generates a PostScript of
 > the pages you asked it to and then sends this complete PostScript
 > document of the pages you requested to a pipe or file on disk
 > that lpr sends to the printer.

AIUI, traditional lpd(8) / lpr(1) do require the file to be
preprocessed in such a way before it is submitted for printing,
but even then, they do /not/ require for the file to be
PostScript: it's possible to setup the respective filters to
accept other formats, such as PDF.

 > So, if qpdf doesn't do the same, I'm out of luck in terms of
 > printing with lpr.  But I think I can find a program that takes
 > page ranges and transformations like scaling and produces a
 > PostScript document that I can send to lpr, so I can use qpdfview
 > and use the command line to print stuff out.

I'm not too familiar with qpdf(1) (and I don't think I've ever
used qpdfview [*]), but it does have a --pages option.  E. g.:

$ qpdf  --empty --pages in.pdf 5-8 -- out.pdf
$ qpdf  in.pdf --pages . 5-8 -- out.pdf

(The second variant preserves the input document metadata,
which isn't probably of much use for printing anyway.)

... A somewhat little-known fact is that once uncompressed, PDF
is largely a text file (perhaps unsurprising, given it comes
from the same company that created PostScript), though employing
byte offsets rather unrestrictedly.

qpdf(1) has a --qdf option that undoes compressesion and annotates
the file in such a way that the companion fix-qdf program can
fix the byte offsets, at least in certain cases, thus allowing the
PDF file to be edited with a text editor.  (Though probably using
a library, such as PDF::API2 for Perl, would be more practical
than trying to, say, adapt sed(1) for automated edits in this case.)

    [*] Given a choice, I tend to prefer HTML.  If the document I'm
interested in is only available in a PDF version, I tend to
use pdftotext(1).  If that fails to produce a legible version,
I resort to Zathura, preferring it mostly for its UI.