Path: utzoo!utgpu!news-server.csri.toronto.edu!rutgers!apple!olivea!tymix!tardis!udwarf.tymnet.com!carl
From: carl@udwarf.tymnet.com (Carl Baltrunas & Cherie Marinelli 0.1.9)
Newsgroups: comp.lang.perl
Subject: Re: Why is this so slow?
Message-ID: <0B030501.O4CUJJ@udwarf.tymnet.com>
Date: 8 Dec 90 05:27:19 GMT
Reply-To: carl%udwarf@tardis.tymnet.com
Organization: Catalyst Art
Lines: 54
X-Mailer: uAccess - Mac Release: 0.1.9


In article <1990Dec6.194412.9132@cunixf.cc.columbia.edu>, ben@cunixf.cc.columbia.edu (Ben Fried) writes:
> 
> I've written a filter in perl to do printer page counting for DEC ln03
> printers.  Right now it seems to be working fine, but it runs abysmally
> slowly.  The vast majority of time is taken up in the following loop:
>
> [perl script deleted] 
> 
> All that really matters is that I be able to count the
> number of pages on the input stream, and be able to break out of the
> loop after either (a) EOF on stdin, or (b) $pageno == $max_pages,
> whichever comes first.
>                                  ...more stuff deleted...
> Ben
> --
> Benjamin Fried
> ben@cunixf.cc.columbia.edu	rutgers!columbia!ben

I know that Ben didn't mention this, and so it wasn't in Randal's or Larry's
responses, so I thought that I'd bring it up.  (I've had to write filters on
non-unix machines to create microfiche-able output, so I've had to do this).

What about line length?  (Page width or whatever you want to call it) ..and
what about line wrap on long lines.  Some print filters do these things for
you, and this changes the the characteristics of the output from straight \n
and \f counting.

So, can you tell perl that while reading input lines to break lines at the point
you want to make it wrap, or do you have to use something like the following
modifications to randal's response:

$lines_per_page = 66; $max_pages = 100; # you forgot this
while (<>) {
    $lineno += ( (length($_) + $line_width - 1) / $line_width );
    $pageno += ( $lineno + $lines_per_page ) / $lines_per_page,   \
               $lineno = 0   if  /\f/;
    $pageno++, $lineno -= $lines_per_page if ($lineno >= $lines_per_page);
    die "Aborted at page $pageno\n" if $pageno >= $max_pages;
    print;
}

And one thing I left out... what to do about \t's (TABS) in text, since they
push the line width quite a bit depending on position.  (I won't mention all
those [which font?] [what about kerning?] word processing things.  They must
require handling each byte. :-)

PS. I guess an array of flags and point sizes could be built first based on the
    font used and a state machine that uses the byte value to lookup the width
    of each character and flags to add kerning offsets and can we break the line
    here... would make a pretty fast way of processing it.  Randal, can this be
    done easily in perl?  How about a 1 liner?

-Carl


 Carl A Baltrunas - Catalyst Art, Cherie Marinelli - Bijoux
 {sumex, apple}!oliveb!tymix!tardis!udwarf!{carl or cherie}
 {carl or cherie}%udwarf@tardis.tymnet.com
 {carl or cherie}@udwarf.tymnet.com