This is the mail archive of the 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]

Re: minutes 2006-09-26


Thanks for the minutes. All of us in China will begin the vacation starting from 10/1... will get back at around 10/8.


we are also interested in fcore and other little utilities, we are very happy to know if there are any progress on that.

Stan Cox wrote:
Utrace (Chris <moller>)
 -ptrace on utrace behavior appears to be different than original utrace
 -If a kill fails then the next ptrace operation is ignored (#203385?)
 -Thread/Process status when a process clones and then later kills the
  clone (#?)
 -Action:  Get the C version of the test for 203385 written and
-Roland is out of action for a while.
-frysk 3205 (bugzilla 206475) Ptrace requests may be masked by failed
 signal delivery.
 -Test case written and a fix is being tested
-frysk 3231  (bugzilla 207674) PTRACE_DETACH doesn't deliver signals
 under utrace.
 -ptrace behavior changed; investigate.
-Action: run frysk tests with patched kernel

Backtrace (Mike <mcvet>)
-Missing frames problem solved
-Sometimes additional frames are displayed
-rt/tests/TestStackBacktrace innermost frame line number is off by 1
-Action: Debug additional frames and line number problem

Fstack (Nurdin <npremji>)
-Changing from jargs to getopt in all callers
-Fstack ready other than adding getopt support
-Action: Work out multi-threaded fstack (+pmuldoon)
-Action: Investigate the gdb pstack bug (+cagney)
 (Thread is in a state that the caller is expecting an event and doesn't
  receive it.)

64 bit (Tim <timoore>)
-32 on 64 bit is done
-fixed multi architecture syscall handling

Breakpoint (Mark <mjw>)
-Working on supporting source line breakpoints
-Blocking observer (was called stepping observer) may help
 -New topic for next week
 -When added to a task then the task becomes blocked
 -Purpose is to stop a task
 -When the observer is removed then the task continues
 -Challenge is when to add and remove observer so single stepping is
-Mark/Tim meeting
-Action: Stepping observer design?

Disassemble (Mike <mcvet>)
-Adding multiple tasks into source windows handling
-Blocking observer would help here too
-Multiple threads in a source window is a candidate for a workflow
-Andrew posting url about multiple threading for insight
-Action: Transition topic to handling multi threaded displays

Testing & Dogtail (Phil <pmuldoon>)
Dogtail is okay for testing except results are not always reliable.
-Future desirable is to also use frame buffer and junit testing
-Getting dogtail to run headless on brew
 -bugfix in frame buffer required
 -Bug cause possibilities are: headless behavior,
  gnome memory smashing, corba (was it corba or something else?) bug
-The memory smashing bug has been isolated

Source (Stan <stan>, Rick <rmoseley>)
-Meeting to discuss symbol/gui interaction
-Symbol handling refactored to make gui interaction easier
-Looking at prior frames support with libunwind.
-Action: prior frame support

Coverage (Stan, Andrew)
-Investigate infrastructure changes needed to support gcov
-No change

Syscall (Phil)
-Yao's caching syscall ideas
 -Syscall observers are more complicated than other observers from the
  core perspective

Core file (Phil)
-Fcore ... create corefile from running process and load a corefile
-How should offline corefiles be handled

RPM (Stepan <kasal>)
-Spinning RPMs and working with Phil with dogtail testing
-FC6 push on Friday or Monday
-Cannot run dogtail on FC5 because dogtail is not available to brew
-How to handle dogtail testing on FC6
 -make it a postinstall test due to time to run tests
-How to handle errata: respin or backport?
 -Hesitant to maintain a branch as is done for gcc toolchains
-Give people advance warning on when branch will be cut

Regards - Wu Zhou

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