Re: [HACKERS] pg_shmem_allocations view

Robert Haas <robertmhaas@gmail.com> writes:
> On Wed, Dec 18, 2019 at 10:59 AM Alvaro Herrera
> <alvherre@2ndquadrant.com> wrote:
>> Can we please stop splitting this error message in two?
>> 
>> +                errmsg("materialize mode required, but it is not " \
>> +                       "allowed in this context")));
>> 
>> (What's with the newline escape there anyway?)

> That message is like that everywhere in the tree, including the
> escape, except for a couple of instances in contrib which deviate. If
> you want to go change them all, feel free, and I'll adjust this to
> match the then-prevailing style.

I agree with Alvaro that that is *not* project style, particularly not
the newline escape.  Like Robert, I'm not quite fussed enough to go
change it, but +1 if Alvaro wants to.

> It seems to me that you could plausibly define this view to show
> either (a) the amount of space that the caller actually tried to
> allocate or (b) the amount of space that the allocator decided to
> allocate, after padding, and it's not obvious that (b) is a better
> definition than (a).

> That having been said, you're correct that the padding space is
> currently reported as <anonymous>, and that does seem wrong.

It seems like it'd be worth subdividing "<anonymous>" into the actual
anonymous allocations and the allocator overhead (which is both
padding and whatever the shmem allocator itself eats).  Maybe call
the latter "<overhead>".  After which, I'd be tempted to call the
free space "<free>" rather than giving it a null name.

            regards, tom lane