Intriguing suggestion! That's a different way to implement literacy. Tavis's suggestion works on GNU/Linux. What would be the best way to incorporate Tavis's suggestion into the screen build process?
Best, Krish On Thu, Feb 1, 2024 at 2:31 PM Tavis Ormandy <tav...@gmail.com> wrote: > > On 2024-01-19, Krish Jain wrote: > > If you folks like the basic idea, we could create a tidy git branch with > > literacy "baked in," thus eliminating the need for the top-level > > literatize script. One unconventional aspect of the process is that a > > tarball must be created that is then embedded into the executable > > immediately after the executable is linked. If you have any guidelines or > > suggestions for how best to do this, please advise. First, please kick the > > tires & let us know how it goes. > > > > I didn't read the whole paper, but it occurs to me that if this is > something you want, it is easy to do without any source modifications. > > For example, you could just do something like: > > $ objcopy --add-section=.srccode=screen.tar.gz /usr/bin/screen /usr/bin/screen > > Then to extract the source code: > > $ objcopy --dump-section=.srccode=screen.tar.gz /usr/bin/screen > > The advantage of this method is that you can also set section flags to > make it non-load (so effectively zero impact on load performance or > rss), and it's easy to strip out (--remove-section) if you want to save space. > > Just a thought! > > Tavis. > > -- > _o) $ lynx lock.cmpxchg8b.com > /\\ _o) _o) $ finger tav...@sdf.org > _\_V _( ) _( ) @taviso
