Re: [PATCH 2/2] Test GDB connection to GDBserver with no symbol files

[Pedro Alves <palves@redhat.com>]

On 02/24/2016 02:26 PM, Luis Machado wrote:
> This test exercises the scenario where we attempt to connect GDB to GDBserver
> in standard remote mode, query the symbol file path, attempt to open said
> symbol file on GDB's end and fail, causing the connection to drop abruptly.
> 
> I wish i could make all the other PASSes go away and only keep the last test,
> as that is the most important one.  

You can pass an empty string as test message to gdb_test, etc. to suppress it,
but I don't like doing that, because then if one of those setup steps fails,
it fails silently.  I think that if we want to make that work nicely, it
should be based on scoping the setup steps with something similar to
with_test_prefix:

setup_tests_scope {
   gdb_test_no_output "set sysroot" "clear sysroot"
   ...
}

gdb_test ... "real test"

Tests run within setup_tests_scope scope would pass silently, but fail
loudly.

> 
> diff --git a/gdb/testsuite/gdb.server/attach-with-no-symbol.exp b/gdb/testsuite/gdb.server/attach-with-no-symbol.exp
> new file mode 100644
> index 0000000..a6d0d85
> --- /dev/null
> +++ b/gdb/testsuite/gdb.server/attach-with-no-symbol.exp

I think this should be "connect", not "attach".

> @@ -0,0 +1,68 @@
> +# This testcase is part of GDB, the GNU debugger.
> +
> +# Copyright 2016 Free Software Foundation, Inc.
> +
> +# This program is free software; you can redistribute it and/or modify
> +# it under the terms of the GNU General Public License as published by
> +# the Free Software Foundation; either version 3 of the License, or
> +# (at your option) any later version.
> +#
> +# This program is distributed in the hope that it will be useful,
> +# but WITHOUT ANY WARRANTY; without even the implied warranty of
> +# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
> +# GNU General Public License for more details.
> +#
> +# You should have received a copy of the GNU General Public License
> +# along with this program.  If not, see <http://www.gnu.org/licenses/>.
> +
> +# Test GDB's ability to properly connect to GDBserver with no pre-opened
> +# symbol file.  If GDB is buggy, it will drop the connection when
> +# it attempts to open the symbol file indicated by GDBserver and fails.
> +#
> +# This test is only meaningful for standard remote connections.
> +
> +load_lib gdbserver-support.exp
> +
> +standard_testfile
> +
> +if { [skip_gdbserver_tests] } {
> +    return 0
> +}
> +
> +if { [prepare_for_testing $testfile.exp $testfile $srcfile debug] } {
> +    return -1
> +}
> +
> +# Make sure we're disconnected, in case we're testing with an
> +# extended-remote board, therefore already connected.
> +gdb_test "disconnect" ".*"
> +
> +# Discard any symbol files that we have opened.
> +gdb_test "file" ".*" "Discard symbol table" \
> +    {Discard symbol table from `.*'\? \(y or n\) } "y"
> +
> +set target_exec [gdbserver_download_current_prog]
> +
> +# Start GDBserver.
> +set res [gdbserver_start "" $target_exec]
> +
> +set gdbserver_protocol [lindex $res 0]
> +set gdbserver_gdbport [lindex $res 1]
> +
> +# Prevent access to the symbol file.
> +if {[file exists $binfile]} {
> +    system "chmod 000 $binfile"
> +}

Seems like this won't work with remote-host testing?

> +
> +# Set sysroot to something non-target and possibly also invalid so that GDB is
> +# unable to open the symbol file.
> +gdb_test_no_output "set sysroot" "Disable downloading of symbol files"

It's more usual for test messages to start with lower case.

> +
> +# Connect to GDBserver.
> +gdb_target_cmd $gdbserver_protocol $gdbserver_gdbport
> +
> +# Check if GDB succeeded connecting to GDBserver by attempting to detach from
> +# the process.
> +gdb_test "detach" \
> +	 ".*Detaching from program: , process.*Ending remote debugging.*" \
> +	 "Connection to GDBserver succeeded"
> 

Likewise.

How about moving the body of the test to a procedure, and then
also testing deleting the file instead of making it unreadable?

Wouldn't the chmod make gdb also fail to load the file with the default
target: sysroot, since gdbserver would fail to read it?  I think
that should also be a test axis.

Thanks,
Pedro Alves