If you are looking for free (as in “free speech”) C/C++ source code, csourcesearch.net could be interesting for you. Apparently, Robert Schultz spent one year using various tools to enter millions of lines of C/C++ code from free software in a database. Most of the code come from software under GPL (more than 40% ; see all the licences parsed and all the packages used). And results are color-coded for the pleasure of your eyes.
The server looks a bit overwhelmed this night. But it could be interesting to compare the Linux kernel swear count with a similar search on this database. More seriously, this database can become an invaluable tool when people or companies are pretending some free software project stole code from their closed-source projects (although I don’t think csourcesearch archive old versions of software packages).