source code analysis tools

Try LXR

Theo

Normally an IDE, but some editors support Ctags:

Programmer's Notepad:

Programming features:

Ctags:

Quote: "... Ctags is a program that generates an index (or tag) file of names found in source and header files of various programming languages. Depending on the language, functions, variables, class members, macros and so on may be indexed. These tags allow definitions to be quickly and easily located by a text editor or other utility. ..."

Glenn

Try

Anton Erasmus

Ctags is good (vi represent, yo!).

For a "pretty" output, possibly running the source through doxygen

The output is improved if the source has comment tags formatted for doxygen but even without those it can be useful.

I've used "Crystal FLOW" and liked it but it's commercial and requires annual maintenance fees for updates/fixes. Also to move to a new machine. It was nice but not *that* nice, so my last maintenanced version sits on the older notebook.

--
Rich Webb     Norfolk, VA

When confronted with an unfamiliar code base I use Doxygen

to get a better understanding of it.

Doxygen (when used in combination with Graphviz) can generate dependency-, inheritance-, call- and include graphs, can tell you which functions refer to a particular function or variable and can generate hyperlinked source browser. Doxygen ain't perfect, it may get confused by complex macro's or templates, but it is free and has done a wonderful job for me so far.

I use cscope (usalling from within emacs).

--
Grant

Another vote for Doxygen here. You can use the wizard to quickly set it up to generate "documentation" for all code (headers and source), including the source in the "documentation" and including all the graphs. You end up with a large set of html documents with all the identifiers being clickable to trace usage and dependencies.

If you are using doxygen for this sort of thing (rather than using doxygen for proper documentation), make sure you only generate html - avoid pdf, rtf, chm formats or you can end up trying to generate books of thousands of pages!

The other tool that comes to mind is Eclipse - with modern versions, its indexer is pretty good, and can help with such cross-referencing.

You've had several suggestions which work on the source code. But such approaches won't always work in the presence of extensive preprocessor use. An IDE which uses information generated during compilation may produce better results.

On one particularly large project, I resorted to using "nm" to dump the symbol tables from the object files; these were processed with "sed" then imported into an SQL database.

Try

if you are using Windows.

Oliver

--
Oliver Betz, Munich
despammed.com is broken, use Reply-To: