From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from apollo.le.ac.uk ([143.210.16.125]) by hawkwind.utcs.toronto.edu with SMTP id <28918>; Tue, 13 Jun 2000 17:34:49 -0400 Received: from happy.star.le.ac.uk ([143.210.36.58]) by apollo.le.ac.uk with smtp (Exim 3.13 #2) id 131W4C-0005ih-00 for rc@hawkwind.utcs.toronto.edu; Mon, 12 Jun 2000 16:23:56 +0100 Received: (qmail 16629 invoked from network); 12 Jun 2000 15:24:18 -0000 Received: from ltpcg.star.le.ac.uk (tjg@143.210.36.203) by happy.star.le.ac.uk with SMTP; 12 Jun 2000 15:24:18 -0000 Subject: Re: environment again To: carlos@texne.com Cc: rc@hawkwind.utcs.toronto.edu In-Reply-To: Date: Mon, 12 Jun 2000 11:23:55 -0400 From: Tim Goodwin Message-ID: <4AcAAKEARTmN4wYA@ltsun0.star.le.ac.uk> > { Col1='data1'; Col2='data2'; Col3='some-big-data' } > [...] > sed 's/__SPECIAL_TAG__/'^$Col3'/' page_template.html What's the problem with simply doing this? Is Col3 so enormous that you're running into environment variable limits? Or are you worried about performance? If it's performance, "profile, don't speculate!" It would be very interesting to see some test results showing just how much it costs to have some enormous variables in the environment. Finally, if I've understood correctly what you're trying to achieve, there is a workaround, although it's rather ugly. eval 'Col3=() sed s/__SPECIAL_TAG__/'^$Col3^'/ page_template.html' Tim.