Re: [PATCH] Add __repr__() implementation to a few Python types

[Bruno Larsen <blarsen@redhat.com>]

On 18/01/2023 19:02, Andrew Burgess via Gdb-patches wrote:
>> --- a/gdb/testsuite/gdb.python/py-arch.exp
>> +++ b/gdb/testsuite/gdb.python/py-arch.exp
>> @@ -29,6 +29,8 @@ if ![runto_main] {
>>   # Test python/15461.  Invalid architectures should not trigger an
>>   # internal GDB assert.
>>   gdb_py_test_silent_cmd "python empty = gdb.Architecture()" "get empty arch" 0
>> +gdb_test "python print(repr (empty))" "<gdb\\.Architecture \\(invalid\\)>" \
>> +    "Test empty achitecture __repr__ does not trigger an assert"
> I guess I was surprised that so many of the new tests included an
> explicit call to repr, given the premise of the change was that simply
> 'print(empty)' would now print something useful.
>
> I guess maybe it doesn't hurt to_also_  include some explicit repr
> calls, but I was expecting most tests to just be printing the object
> directly.
>
> Is there some reasoning here that I'm missing?
>
When you make a print(empty) call, python internally calls __str__ 
instead or __repr__. The latter is only called when the former is not 
available or when it is being printed by the interpreter. Since 
py-type.c already includes a __str__ implementation, the test would 
still pass if you just had print(empty) without Matheus's changes.

-- 
Cheers,
Bruno