Re: BUG #18692: Segmentation fault when extending a varchar column with a gist index with custom signal length

On Fri, Nov 8, 2024 at 4:05 AM Tom Lane <tgl@sss.pgh.pa.us> wrote:
> Alexander Korotkov <aekorotkov@gmail.com> writes:
> > On Thu, Nov 7, 2024 at 8:47 AM Tender Wang <tndrwang@gmail.com> wrote:
> >> Based on Tom's analysis, I  provide a POC patch. I'm not sure if it is right
> >> to use DEFAULT_COLLATION_OID in the patch.
>
> > Thank you.  But I'm not sure about DEFAULT_COLLATION_OID.
>
> I don't quite trust that either.  But since we only care about
> equality, wouldn't it be OK to use C_COLLATION_OID?
>
> > Therefore, we can compare two text[] just with datumIsEqual().
> > Attached patch implements this.
>
> I think this is nonsense.  What about toasted datums, or even
> just short-header ones?  The one coming from an on-disk tuple
> is pretty likely to be short-header for plausible sizes of
> the options, but the one we just constructed in memory will
> not be.

You are correct.  I quickly skim trough the sources and didn't find a
function which compares detoasted contents of Datums excluding
headers.  So, yes, comparison using C-collation seems the most
reasonable option.

------
Regards,
Alexander Korotkov
Supabase