meeting 2006-10-17

meeting 2006-10-17

[Andrew Cagney]

Agenda (from last meeting with slight edits):

Testing: 
-action: FC6 (possibly kernel?) failures (cmoller)
-action: install tree problems, issue #210203 (cmoller, cagney)
-action: run headless  (ldimaggi, cagney)
-action: code coverage [stan, cagney]


Source View:
-action: thread independence investigation (mcvet)
-action: look at demangling (stan)


Command Line Utilities:
-action: early cut of fcore (pmuldoon)


Core: backend/middle end
-Tim is pursuing libdw/libdwfl
-Face to face meeting complications
-Mark is on vacation next week, orientation in 3 weeks
-Mark observer to temporarily suspend the task
-Mark working on blocking observer


Java-GNOME:
https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=209346
http://sourceware.org/bugzilla/show_bug.cgi?id=3116
http://bugzilla.gnome.org/show_bug.cgi?id=353298


New ports:
- PPC 64 and libunwind - IBM
- ubuntu - ??


Documentation and demoflow:
- general discussion for now


Update
-Goals: continue, step, breakpoint
-Have rpm available for when FC6 unfreezes as part of updates

minutes 2006-10-17

[Stan Cox]

Testing:
Chris looking at stopped process problem (210693) along with Roland
RHTS automatic tests can be run manually
Code Coverage (Stan,Andrew)

-action: simpler test for 210693 (Oct-20, cmoller)
-action: try out RHTS tests (Oct-20, cmoller)
-action: code coverage (Oct-20, scox)

Source View:
cdt parser bug fixing (Rick)
-perhaps actually a libgcj bug?
libiberty demangler and libiberty licensing

-action: demangler plan (Oct-20, stan)

Command Line Utilities:
first cut of fcore complete
gelf always uses 64 bit header
-should be able to get 32 bit corefile on a 64 bit machine
-currently elf headers look the same as gcore
Instead of panicing, detach process and let core continue to run.
-change attach,continue,stop,detach TO  attach,detach
-add abandoning state like detach but 'worse'

-action: consider how to load core file (Oct-20, pmuldoon)
-action: Process detach (Oct-27, mcvet)


Core: backend/middle end
stepping is committed via a blocking observer
Tim looking at high level breakpoints

-action: Pick this up while Mark on vacation (Oct-27, npremji)


Java-GNOME:
Need these resolved:
https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=209346
http://sourceware.org/bugzilla/show_bug.cgi?id=3116
http://bugzilla.gnome.org/show_bug.cgi?id=353298

-action: Resolve java-gnome problems (?, ?)

New ports:
IBM PPC libunwind status?
Ubuntu build problems


Documentation and demo workflow:
Nice to have: show workflows for debugging and monitoring
Documentation will use yelp, the gnome help system
-to what extent does it use gnome?

Re: minutes 2006-10-17

[Wu Zhou]

Stan Cox wrote:

> New ports:
> IBM PPC libunwind status?

Corey is head down in some other emergency recently.  Hopefully he can return back to libunwind soon.