DXR needs Python/CGI to run :-)
My favorite difference between the Woboq Code Browser and DXR is to have a pure static website generated ahead of time without requiring dynamic code or a database on the HTTP server.
Very nice, I've been looking for something like this for my own open source project, have even been thinking about developing one myself (also using clang/llvm, to get metrics for visualization, call graphs, usages, data dependencies etc).
Current similar systems, as far as I know, are either very limited, or slow, or almost impossible to get to work for other projects than what they're based on.
Nice work. I hope the system itself will also be open source some day :)
The PHP project uses OpenGrok for its codebase (although curiously at an address suggesting it uses LXR instead). http://lxr.php.net/ is always the go-to place to read the soure code now, not the Git repo. It's especially useful when tearing through the Zend engine.
Is there an open-source equivalent of this anywhere? I have seen DXR/LXR but I think they use an "approximate" parser. So it will get confused by functions in different scopes with the same name, and so forth.
This actually uses clang which can do the analysis exactly.
(DXR also uses Clang as its parsing engine)