Re: [PATCH v3 2/5] Introduce frame_info_ptr smart pointer class

[Bruno Larsen via Gdb-patches <gdb-patches@sourceware.org>]

Hi Pedro,

On 24/08/2022 17:20, Pedro Alves wrote:
> Hi Bruno,
>
> On 2022-08-24 3:24 p.m., Bruno Larsen wrote:
>> Hi Pedro,
>>
>> Thanks for the review, and sorry for the delayed response.
>>
>> On 25/07/2022 19:52, Pedro Alves wrote:
>>> On 2022-07-25 6:06 p.m., Bruno Larsen via Gdb-patches wrote:
>>>
>>>> +
>>>> +private:
>>>> +
>>>> +  /* The underlying pointer.  */
>>>> +  frame_info *m_ptr = nullptr;
>>>> +
>>>> +  /* All frame_info_ptr objects are kept on a circular doubly-linked
>>>> +     list.  This keeps their construction and destruction costs
>>>> +     reasonably small.  To make the implementation a little simpler,
>>>> +     we guarantee that there is always at least one object on the list
>>>> +     -- this "root".  */
>>> This comment is stale -- this is no longer a full frame_info object.
>> I'm not sure why you mention it being a full frame_info object.
> I meant frame_info_ptr instead of frame_info, as that's what it was in the
> original version:
>
>   +  /* All frame_info_ptr objects are kept on a circular doubly-linked
>   +     list.  This keeps their construction and destruction costs
>   +     reasonably small.  To make the implementation a little simpler,
>   +     we guarantee that there is always at least one object on the list
>   +     -- this "root".  */
>   +  static frame_info_ptr root;
>
>> root was never a usable frame_info_ptr, it was just used as a pointer to the frame_info_ptr list from the outside, and to make intrusive list operations easy as we don't have to check for an empty list. I did reword the comment to:
>>
>> /* All frame_info_ptr objects are kept on an intrusive list.
>>    This keeps their construction and destruction costs
>>    reasonably small.  To make the implementation a little simpler,
>>    we guarantee that there is always at least one object on the list
>>    - this "root".  It is only used to simplify intrusive_list operations.  */
>>
>> to hopefully explain things better.
> Not really.  The comment was there because the linked list implementation used one
> element type as "root" (aka "head") object, a full frame_info_ptr object, same type as
> all other elements, even if only its m_next/m_prev pointers were used.  This part:
>
>    "To make the implementation a little simpler, we guarantee that there is always
>     at least one object on the list this "root"."
>
> ... with the conversion to intrusive list, that comment no longer makes sense.  We don't
> store a "root" object of element type any more.
>
> In fact, I think you should rename the "root" to frame_list, or some such, like:
>
>   - static intrusive_list<frame_info_ptr> root;
>   + static intrusive_list<frame_info_ptr> frame_list;
>
> ... and don't talk about "root" any longer; however the head node is implemented, it's
> an intrusive_list implementation detail.

Ah, I see what you mean. I guess I still had the old implementation in 
mind when I wrote the comment, and agree with you that renaming the 
variable would avoid a lot of confusion. I'll rethink this bit and send 
a v3 later this week, thanks for the explanation!

-- 
Cheers,
Bruno