On 10/29/21, 2:14 AM, "Michael Paquier" <michael@paquier.xyz> wrote:
> By the way, while looking at this area of the code (particularly
> tsvector and gist), I was under the impression that it is safe to
> assume that we don't particularly need to rebuild the index when
> changing its attoptions this way, but I have also the feeling to be
> rather naive here.
This is my concern as well. I remember seeing some weird errors in
the thread that led to fdd8857, and I was able to reproduce it with
this new patch. If you apply the following patch for the hstore test,
it will begin failing.
--
diff --git a/contrib/hstore/expected/hstore.out b/contrib/hstore/expected/hstore.out
index 64a3272b9c..a4c78a40c8 100644
--- a/contrib/hstore/expected/hstore.out
+++ b/contrib/hstore/expected/hstore.out
@@ -1352,6 +1352,7 @@ create index hidx on testhstore using gist(h gist_hstore_ops(siglen=2025));
ERROR: value 2025 out of bounds for option "siglen"
DETAIL: Valid values are between "1" and "2024".
create index hidx on testhstore using gist(h gist_hstore_ops(siglen=2024));
+alter index hidx alter column 1 set (siglen=100);
set enable_seqscan=off;
select count(*) from testhstore where h @> 'wait=>NULL';
count
diff --git a/contrib/hstore/sql/hstore.sql b/contrib/hstore/sql/hstore.sql
index a59db66b0a..5e2a1ec99f 100644
--- a/contrib/hstore/sql/hstore.sql
+++ b/contrib/hstore/sql/hstore.sql
@@ -308,6 +308,7 @@ drop index hidx;
create index hidx on testhstore using gist(h gist_hstore_ops(siglen=0));
create index hidx on testhstore using gist(h gist_hstore_ops(siglen=2025));
create index hidx on testhstore using gist(h gist_hstore_ops(siglen=2024));
+alter index hidx alter column 1 set (siglen=100);
set enable_seqscan=off;
select count(*) from testhstore where h @> 'wait=>NULL';
--
However, if you add a REINDEX right after the ALTER INDEX command, it
will succeed.
Nathan