On 03/05/13 15:04:52, Pedro Alves wrote: > On 05/03/2013 12:38 PM, Abid, Hafiz wrote: > > > Hi Pedro, > > Thanks for letting me know. There is a testcase in > status-stop.exp(test_tstart_tstart) to check this case but it is not > working as intended. I will work on a patch. > > Thanks. > > -- > Pedro Alves > Hi Pedro, I have attached a patch. It calls 'stop_tracing' from the QTinit handler. I was wondering that setting 'tracing' to 0 may be enough. But stop_tracing seems to do a lot of other things which may be required. So I thought it is better to go with it. Another possibility is to let QTBuffer:size handler return ok if the new size is equal to current size irrespective of the value of 'tracing'. As I mentioned earlier, a testcase was already present for this scenario but it was missing the error message that came after we press 'y'. I have handled it but would like to know if there is a better way to do it. I saw no regression on native-gdbserver run. Regards, Abid gdb/testsuite/ChangeLog: 2013-05-03 Hafiz Abid Qadeer * status-stop.exp (test_tstart_tstart): Check for error returned by the second 'tstart' command. gdb/gdbserver/ChangeLog: 2013-05-03 Hafiz Abid Qadeer * tracepoint.c (cmd_qtinit): Call 'stop_tracing'.