> On Wed, Dec 30, 2020 at 09:01:37PM +0100, Dmitry Dolgov wrote:
> > make check fails
>
> Yeah, apparently I forgot to enable asserts back after the last
> benchmarking discussion, and missed some of those. Will fix.
>
> > 2. The index position was ignored.
> >
> > postgres=# update foo set a['a'][10] = '20';
> > UPDATE 1
> > postgres=# select * from foo;
> > ┌─────────────┐
> > │ a │
> > ╞═════════════╡
> > │ {"a": [20]} │
> > └─────────────┘
> > (1 row)
>
> I just realized I haven't included "filling the gaps" part, that's why
> it works as before. Can add this too.
>
> > 1. quietly ignored update
> >
> > postgres=# update foo set a['a'][10] = '20';
> > UPDATE 1
> > postgres=# select * from foo;
> > ┌────┐
> > │ a │
> > ╞════╡
> > │ {} │
> > └────┘
> > (1 row)
>
> This belongs to the original jsonb_set implementation. Although if we
> started to change it anyway with "filling the gaps", maybe it's fine to
> add one more flag to tune its behaviour in this case as well. I can
> check how complicated that could be.
Here is what I had in mind. Assert issue in main patch is fixed (nothing
serious, it was just the rawscalar check for an empty jsonb created
during assignment), and the second patch contains all the bits with
"filling the gaps" including your suggestion about creating the whole
path if it's not present. The latter (creating the chain of empty
objects) I haven't tested that much, but if there are any issues or
concerns I guess it will not prevent the main patch from going forward.