Re: [PATCH][gdb/testsuite] Don't abort testrun for invalid command in test-case

[Pedro Alves <palves@redhat.com>]

On 6/17/20 3:14 PM, Tom de Vries wrote:
> On 6/16/20 2:47 PM, Pedro Alves wrote:
>> On 6/11/20 11:55 PM, Tom de Vries wrote:
>>
>>> Hmm, what is the distinction between gdb_init and default_gdb_init?
>>>
>>> All the other uses in gdb.exp of pattern foo/default_foo have an
>>> implementation:
>>> ...
>>> proc foo {} {
>>>   [default_foo]
>>> }
>>> ...
>>> but gdb_init is much more than that. Why is that different?
>>>
>>
>> I don't know.  I guess it shouldn't.  I guess people (including me) added to
>> gdb_init over time without realizing they were breaking the pattern.  Maybe nobody
>> notices it because whatever is overriding gdb_init renames the original one and
>> then calls it.
> 
> Hmm, thanks for the explanation.  I feel we need to improve this
> situation somehow, but I'm not sure yet how.

Move stuff in gdb_init to default_gdb_init, and add a comment to these
functions to not add stuff in there?  That seems like the obvious.
I'm not sure why we have that pattern in the first place though, given
that you can rename/override functions in tcl anyhow.  This all predates me.

Thanks,
Pedro Alves