Re: [PATCH v2 3/4] gdb/progspace: add solib_ops pointer in program_space

[Simon Marchi <simon.marchi@efficios.com>]

On 6/11/25 2:14 PM, Guinevere Larsen wrote:
>> @@ -1273,9 +1277,9 @@ solib_name_from_address (struct program_space *pspace, CORE_ADDR address)
>>   bool
>>   solib_keep_data_in_core (CORE_ADDR vaddr, unsigned long size)
>>   {
>> -  const solib_ops *ops = gdbarch_so_ops (current_inferior ()->arch ());
>> +  const solib_ops *ops = current_program_space->solib_ops ();
>>   -  if (ops->keep_data_in_core)
>> +  if (ops != nullptr && ops->keep_data_in_core != nullptr)
>
> Can this function be called before the program space is fully setup?
> 
> Otherwise I think we should either assume that it was set correctly,
> or assert it. Better to be loud and easy to find the bug than
> accidentally adding a regression in a hard-to-spot place.
> 
> This goes to all other similar places where you check for ops !=
> nullptr.

I don't really know for sure for all the methods.
program_space::m_solib_ops is nullptr before you run the program, and I
don't know off-hand which methods are reachable with an inferior that's
not running yet.  I'll try to see if I can figure it out.

Simon