Perlfect Solutions
 

[Perlfect-search] mini Roadmap

Daniel Naber dnaber@mini.gt.owl.de
Sun, 13 Aug 2000 17:04:43 +0200
Hi,

there are some patches on my patches page 
(http://mini.gt.owl.de/~dnaber/perlfect/) that have not made it into 3.09. 
I wonder which patches should go into the next version:


indexer-cgi.diff and index_form.html
If you do not have ssh/telnet access to your server you can use this patch 
to start indexer.pl as a CGI. You need to set a password in conf.pl to 
make things more secure. Then you can start the indexing process using
index_form.html. Warning: Only use this if absolutely necessary.

--> This is only for people without telnet/ssh access... if used, it's not 
very secure. I doubt if it makes sense to add it?


multi-boost.diff
Adds an option to tune ranking. This will increase the ranking of a 
document if there's more than one matching term in that document.

--> this is a very short patch, it's optional and you can play with it 
(now need to call indexer.pl if you modify the option). I think this 
should go in.


joker.diff
Enables the use of the '*' as a joker/wildcard at the end of a search 
term. This is slower than search without jokers, and it does not work if 
+/- is used for that term.

--> This is just a hack, it's not efficient at all. It just iterates 
trough the terms. This should not go in, IMHO.


pdf.diff
This allows you to index PDF documents. You need pdftotext, which is part 
of xpdf. As a side effect, untitled documents aren't called "Untitled" 
anymore, but their filename is used as a link. Warning: This patch now 
includes a security check, but if you don't trust the people whose files 
you want to index, please do not use this option. It is neccessary to 
invoke a shell to index PDFs, and this is a potential security problem. 

--> Useful for big web pages that very often have at least some PDF files. 
It's quite short, as it uses an external program. I think this should go 
in.

Opinions?

Regards
 Daniel