[PATCH v2] Indicate batch mode failures by exiting with nonzero status

Tom Tromey tom@tromey.com
Fri Aug 31 14:33:00 GMT 2018


>>>>> "Gary" == Gary Benson <gbenson@redhat.com> writes:

Gary> Yes please! Is it failing on *every* platform? I looked in gdb.log.xz
Gary> but there's no output from GDB, only the PASS/FAIL lines.  So I don't
Gary> really know what to do, I can't reproduce :/

It fails for me too.  I ran with -v and saw this in the .log:

    RESULT = 11271 exp6 0 0
    PASS: gdb.base/batch-exit-status.exp: -batch: [lindex $result 2] == 0
    PASS: gdb.base/batch-exit-status.exp: -batch: $actual_status == $expect_status
    Quitting /home/tromey/gdb/build/gdb/testsuite/../../gdb/gdb -nw -nx -data-directory /home/tromey/gdb/build/gdb/testsuite/../data-directory 
    doing kill, pid is 11271
    Spawning /home/tromey/gdb/build/gdb/testsuite/../../gdb/gdb -nw -nx -data-directory /home/tromey/gdb/build/gdb/testsuite/../data-directory -batch -jslkflsdjlkfjlksdjf
    RESULT = 11271 exp6 0 0
    PASS: gdb.base/batch-exit-status.exp: -batch -jslkflsdjlkfjlksdjf: [lindex $result 2] == 0
    FAIL: gdb.base/batch-exit-status.exp: -batch -jslkflsdjlkfjlksdjf: $actual_status == $expect_status
    Quitting /home/tromey/gdb/build/gdb/testsuite/../../gdb/gdb -nw -nx -data-directory /home/tromey/gdb/build/gdb/testsuite/../data-directory 
    doing kill, pid is 11271


Note how the PID doesn't change.

I suspect that what is happening is that gdb_spawn_id isn't being reset,
and so we're hitting this code in default_gdb_spawn:

    if [info exists gdb_spawn_id] {
	return 0
    }

Something has to unset this after each invocation.
Maybe calling clear_gdb_spawn_id, but I don't know if that's best or if
there should be something else.

Tom



More information about the Gdb-patches mailing list