Tracepoints and not stopping the target

[Corey Minyard]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

I've been working on a debugging system that links into an application
and provides debugging services so you can connect a user interface, set
tracepoints, dump data, etc.  It is designed to run in a live soft
real-time system with minimal impact to the application.

Now the time has come to hook a source-level debugger onto it.  I have
designed it with GDB in mind, the tracepoint language is the bytecode
language used by GDB.  gdb seems to have one big hangup with this,
though.  It expects to be able to stop the application and won't read
the tracepoint data until the application is stopped.  This is obviously
not an option for me.  In the bytecode documentation it talks about this
being a future option, but it doesn't yet seem to be implemented.

Does anyone have an idea how hard this will be to add?  Would it be
enough to pretend to stop the application but not really stop it?  Has
anyone worked on this before?

Thanks,

- -Corey
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.0.6 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org

iD8DBQE+SUDNIXnXXONXERcRAqYkAJ44TEF7gbO+s7czJhtpBPqoiiUQ6ACfccex
K6I0bOX6lUFyQvcrA509L60=
=7EPl
-----END PGP SIGNATURE-----