In
our Mozilla Reverse Engineering Project, we are trying to extract architectural
information from the publicly available Mozilla
(Netscape) source code. Here's what we have done so far:
- The C source
files of Mozilla were parsed using our C parser cparse. The 4-tuple
RSF output for each file was stored in a separate file.
- All the RSF
files were fed into sortrsf, to eliminate duplicate tuples and
sort the tuples for faster processing.
- The htmlrsf
utility was used to create a HTML version of the Mozilla source code
with hypertext cross references and a 3-tuple RSF file for use with
rigiedit.
- In ongoing
work, the graph editor rigiedit is used to identify and extract
subsystems of Mozilla.
Preliminary
results:
- The original
graph has about 100,000 nodes (datatypes, variables, constants,
function, prototypes).
Downloadable
files:
Click here
to browse the Mozilla source code online.
The RSF and
HTML files available here were produced from the Mozilla version released
on October 8, 1998. The Netscape
Public License applies to all files provided here for download and/or
display. |