open a pipe from system..

Anonymous Monk has asked for the wisdom of the Perl Monks concerning the following question:

Hi all and a nice Sunday :-)

There is a non-Perl-program pdftohtml to convert pdf to html (and I want to convert that hereafter to txt).
I can do that like this:

	system "pdftohtml x.pdf, x.html";
	system "myPerlPrg.pl x.html x.txt";

not very elegant :-(
I would like to do that without using that temporary html-file x.html.

- Either by getting the pipe into the invoking-prg (? possible, how?):

	..
	system "pdftohtml x.pdf |";
	# now read this pipe | (???)

Or to to pipe the output of pdftohtml directly to another Perl-prg:

	system "pdftohtml x.pdf | myPrg.pl"

but how do I know about and open this pipe within myPrg.pl? I don't find it in @ARGV, correct?

Thanks in advance,
carl

Comment on open a pipe from system..

Replies are listed 'Best First'.
Re: open a pipe from system.. by Tanktalus (Canon) on May 01, 2005 at 15:26 UTC
You have the terminology right - you want to open the filehandle: `open my $fh, 'pdftohtml x.pdf \|' or die "Can't run pdftohtml: $!";` [download] However, that assumes that pdftohtml will spit its output to stdout rather than a file - I don't know this. The other alternative is for myPerlPrg.pl to fake this by having it take the x.pdf file, create the x.html file, and read it in a single execution. The command line would look like: `myPerlPrg.pl x.pdf x.txt` [download] You could create a temporary filename, pass that to pdftohtml, and then use that to read the HTML to convert to text. Hope that helps,	[reply] [d/l] [select]
Re^2: open a pipe from system.. by Anonymous Monk on May 01, 2005 at 15:36 UTC
ehem? I can run pdftohtml without using either system or exec? It means that `open my $fh, 'pdftohtml x.pdf \|' or die "Can't run pdftohtml: $!";` [download] 1st runs pdftohtml? 2nd creates a file handle (in var $fh)? 3rd I can read the output of pdftohtml from this $fh? Thanks a lot, going to try it :-) Carl	[reply] [d/l]
Re^3: open a pipe from system.. by Anonymous Monk on May 01, 2005 at 15:57 UTC
thanks a lot! This works perfectly: `open my $fh, "pdftohtml -noframes -stdout $pdf \|" or die "Can't run +pdftohtml: $!"; while (<$fh>) { &func($_); } close($fh);` [download] option -stdout is needed (in case s.o.else is searching) to catch the output. have a nice Subday Carl	[reply] [d/l]