This is the mail archive of the frysk@sources.redhat.com mailing list for the frysk project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Frysk weekly meeting minutes 20060711


Guadec report: attendance was good. Fair number of people in the
talk. Done demo but not slides, because only 5 minutes session.

syscall tracing work: wrote ftrace application. Chokes on syscalls
need to find better way to figure out which syscall are legal. Is
there a database somwhere? Seems like there isn't. Kernel audit system
probably has something like that. Sami to find out.

Libunwind build/linking problem with gcj: Rick/Andrew: two identically
named symbols in 2 different libraries, libunwind and
libgcj. libunwind.so or libunwind-<target>.so? Maybe try to use the
target one instead of the generic one. There are different ways to
build it. Workaround is there, but doesn't solve the symbol
clash. This is fixed in gcc-4.2, but that's irrelevant to us for
now. Only way is to not build that stuff in. Fix: workaround is in.
Proper fix seems to be simple makefile change. Needs to be in by
tomorrow. Rick. 

Libunwind status: Adam: checked out the lower half of the
bindings. Callbacks need to be pointed at ptrace. Enabler for raw
addresses backtraces.

Multiple invocations of frysk: Mike: problem on amd64. Still being
worked on. right now it cannot bring up the window of the currently
running frysk on amd64. Can we get this fixed by tomorrow on x64-64.

Frysk on Frysk: cannot debug same instance of frysk (obviously) but
should be able to debug another instance of frysk. We need to enable
that.

There is a bug about frysk debugging itself (same pid): ignore that
bug, we don't want to do that. Please close that bug. Phil needs to
cleanup the process list though, we don't want to have Frysk listed or
can we grey it out so that it cannot be selected? Try to get this done
by tomorrow.

Event viewer: Nurdin: accessibility: being added but having trouble
finding a testing program. Using sniff from dogtail, but it seems to
get itself into a locked state. Talk to Len.

utrace: Chris fulltime on it. Writing tests for utrace. 

back burner item: investigate why the different platforms of
java-gnome behave differently.


register window: nothing to do for now. Tim had a few 32/64 bits bugs.
End of day tomorrow for amd64.

memory window: should work on 64-bit. Tim checked something in. (I
didn't catch the details sorry). Mostly a port of the register window,
including bugs :-)

64-bit: Tim: source comes up fine, reg window (working through that
now) memory window not yet. Need to run through the tests.

breakpoints: in progress, tasks are understood. no showstoppers.

console/terminal workflow: people are goin to discuss this offline.

process list window problems:

- will show also defunct processes, because they are caught in between
  states. Len filed a bug.  

- "cannot find name": on Phil's buglist. Same as defunct process
  problem: cannot get a name of executable.

- frysk listed twice: Filter it out

- define sorting: need to sort them alphabetically. 

Phil says they should all be done by tomorrow. Yea!

gcov: on its way. Results should be available next week. Stan has this
on backburner.

dogtail on ppc64: Len will give it a spin tomorrow just to see is it
works at all. IBM will do more serious testing.





Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]