I've completed an alpha version of marc-search.cgi 4.0. It's much
more efficient and includes some often requested features. To wit,
Works with archives split over multiple directories.
Works on web sites that are tilde dependent. Make sure to set
$usersubdir to match your server settings in the configurable variables
section during installation!
Checks for a configuration file (.marc-search.cgf -- note leading dot!)
in the directory from which the search was initiated and reads the
archive name from it. Otherwise the directory name is used.
There's more to say, but for that you can read the pod documentation
I've spent so much time writing :-).
You can try it out at
You can also substitute ~friedman for the indiv/friedman part if you
want to try the tilde thing out.
BUT READ THE CAVEAT BELOW FIRST!!!
Since I didn't have any multiple directory archives of my own to play
with, I improvised and copied a bunch of unrelated ones (the five digit
numbers are course codes from my university) into a directory structure
organized by years and months. Suffice it that the message dates
therefore have no relationship whatsoever to the months I divided the
archive up into, since my test archive is really several small archives
which started around the same time (in 97). This isn't that important,
but I wanted to warn you to expect this discrepancy on >my test site<.
If somebody wants to gzip up a years worth of messages broken out into
months, I'd be grateful to have something more "realistic" to play around
More importantly, anyone interested in testing the script out on their
system should drop me a note -- friedman(_at_)uci(_dot_)edu(_dot_)
Many thanks to everyone who sent in helpful suggestions. I tried to
incorporate as many of them as possible and am pretty pleased with the
results thus far.
Eric D. Friedman