Perlfect Solutions
 

[Perlfect-search] can I use the same index data files for UNIX and WINDOWS?

Dimitri Stamatis ds@prs.net
Wed, 27 Feb 2002 09:40:47 -0800
>> Can I use the same indexed binary data files for UNIX and WINDOWS?
>
>That should be possible in theory, but it never works in real life :-(
>

Thank you, Norm and Daniel, for your responses. Maybe I'm totally off on this, but could the problem be:
Windows represents carriage returns with 2 characters, whereas Unix represents them with only 1 character?

Maybe this is what's throwing off the search results when uploading the Perlfect-generated binary files between OS's.

Daniel wrote:
>DB_File is supposed to be byte order independent, according to the 
>documentation. The fact that it does not work in the real world is 
>probably be caused by the Berkeley DB backend, its different versions are 
>not compatible - so even if you have two Linux systems, they still need 
t>he same version of Berkeley DB so you can copy the files.

Maybe the problem is ActivePerl's implementation of DB_File? Since ActivePerl is ported to run on Windows, it implements things a little differently "under the hood" which may be screwing things up.

I'm going to try a simple test: I'm going to index two simple files on Windows and Unix and compare them. Maybe this will reveal something.

>All you need is Perl. Then just install Perlfect Search and index your 
>files. You don't need Apache, not even your mouse for that (besides, you 
>don't have to do anything to install Apache, Perl, etc: your Linux 
>distribution will do that for you).

Apparently Apache comes pre-installed on Linux. But, I think I still need to download and configure Apache's Tomcat server in order to run the portion of my site that is JSP/servlet driven (unless that comes pre-installed as well).

Thanks for all the advice!
Dimitri
____________________
Dimitri Stamatis
Senior Web Developer
Classical Archives, LLC

tel: 650.330.8050
email: dimitri@prs.net