Re: [PATCH v5 2/2] gdb: do autoload before notifying Python side in new_objfile event

[Simon Marchi <simon.marchi@polymtl.ca>]

>> +# Make the -gdb.py script available to gdb, it is automatically loaded 
> by
>> +# gdb if it is put in the same directory as the library.
>> +set remote_python_autoload_file \
>> +    [gdb_remote_download host $python_autoload_file]
>> +
>> +gdb_test_no_output \
>> +    "set auto-load safe-path ${remote_python_autoload_file}" \
>> +    "set auto-load safe-path"
>> +
>> +# Load the Python file that defines a handler for the new_objfile event,
>> +# which will generate the output to check later
>> +# (prints information on available pretty-printers for objfile).
> 
> In v5, the handler no longer generates output (for the good case), so maybe this comment should be changed to something like this:
> 
> # Load the Python file that defines a handler for the new_objfile event,
> # which will set a global variable if the pretty-printer is available.


Shortened it to just:

# Load the Python file that defines a handler for the new_objfile event.

I think that's enough details.

>> +set remote_python_event_handler_file\
>> +    [gdb_remote_download host $python_event_handler_file]
>> +gdb_test_no_output "source ${remote_python_event_handler_file}" "load python file"
>> +
>> +gdb_load ${binfile}
>> +
>> +gdb_test_no_output "set print pretty on"
>> +
>> +# Check that the handler prints output when test library is loaded
>> +# and that the pretty-printer from the auto-loaded Python file has been
>> +# registered.
> 
> Same as for the comment above (no more output generated by the handler). I guess this comment can just be dropped, since there's a separate one for the check below now.

Indeed, did that.

Thanks,

Simon