[Bug translator/11353] elide side-effect-free probes

[fche at redhat dot com]

------- Additional Comments From fche at redhat dot com  2010-03-19 16:40 -------

> So the probes should be removed in semantic_pass_opt4 (where the warning is
> currently thrown)? Also, does 'removing them entirely' entail removing them from
> systemtap_session.probes? I seem to have somehow convinced myself that it does :/

Yes, but unfortunately that's not enough.  By this time in pass 2, derived_probes
are also enrolled in various derived_probe_groups.  So ones about to be elided
would also have to be 'un-enrolled'.  This could be done with an api extension
for derived_probe and derived_probe_group, and its implementation in a handful
of concrete classes.  This is sounding more and more involved....


-- 


http://sourceware.org/bugzilla/show_bug.cgi?id=11353

------- You are receiving this mail because: -------
You are the assignee for the bug, or are watching the assignee.