First a quick comment that LXR is a great project. I've been using LXR for at least a couple of years and it has been my secret weapon. I didn't try to keep it a secret but sometimes folks just don't listen ;) I just installed LXR on our internal network so hopefully now they will see for themselves.
As I was setting up LXR I ran into a few things that I wanted to mention.
1) In Common.pm, (function: printhttp) the code that generates the Last-Modified header was not working right with Mozilla 1.7.11 on windows 2000. I'd get blank pages and the apache log showed a couple of errors (no headers sent and error on lines where expandtemplate was called) No problem with 1.7.4 on Linux or with IE 6.1 on windows. I didn't try to troubleshoot any further, just commented out this code for a quick and dirty workaround.
2) In Common.pm (function: clean_path), the regex for removing 'dodgy' characters was not compatible with our source tree since we have directories which contain a dash ('-') in the name.
3) In genxref I added a -i option to the invocation of glimpseindex. This makes the .glimpse_include file take precedence over .glimpse_exclude. I put '*' in .glimpse_exclude and *.c , *.asm , *.h and *.java in .glimpse_include so that only files with these extensions would be indexed. I thought that others might find this useful too.
1) I miss the file search feature from the previous version of LXR.
2) When doing an identifier search, when the symbol shows up repeatedly in a given file it would be nice to display the file name followed by a compact list of line numbers rather than repeating the file name on a separate line for each file number.
I may take a stab at these changes myself and if I do I'll see if anyone is interested. But my Perl experience is pretty much limited to referring to the camel book occasionally when I need to, so we'll see.
I tried to keep this brief, please know if there are any details I can add.
Thanks again for such a useful timesaver!