Note that it doesn't work like your typical Java/C++ static analysis tools, but instead injects itself into the running perl process, like Slime and other emacs interfaces for dynamic languages.

Sepia could do class browsing (see Sepia::methods and sepia-tree.el in the distribution), but I don't write enough heavily OO code to have felt the need to implement it. If you want to try writing it yourself, I would be happy to help.

Exuberant ctags should be able to output a TAGS file which emacs can grok enough to let you jump from calls to definition. Also possibly of interest might be ECB which understands Perl as one of the languages and gives you a tree of declarations and what not to jump through code.

(Slight disclaimer: I've been Switched™ to OS X for about a year now so my (x)emacs-fu is rusty, but this was the stuff I used when I was still an active xemacsian).

From the ECB FAQ, the Perl support sounds luke warm at best, and Semantic doesn't list it as one of the supported languages.

Did you actually use ECB with Perl? And if so, how? (imenu? etags?) And how well did it work? Was it difficult to set up?

/J

I want to say imenu, but again this was over a year ago and I don't see anything conclusive in my vestigial .xemacs files. Likewise I vaguely recall having to do a little work to get things going, but on the order of an hour or two not days. As for how well it worked, for the most part it found subs and packages pretty well (again, foggy memory here). It worked well enough I had a permanent keybinding set aside for it, at least. :)

The cake is a lie.
The cake is a lie.
The cake is a lie.