sql: fix expandIndexName to return correct table mutable

[chengxiong-ruan]

Fixes #81211
A regression was introduced with #80769 that the table mutable
got created from a uncommitted descriptor. This pr tries to fix
that.

Release note: None.

[cockroach-teamcity]

This change is 

[ajwerner]

Heh, @ecwall hit me up about this too. I had already typed something, curious to see how it makes you feel:

--- a/pkg/sql/resolver.go
+++ b/pkg/sql/resolver.go
@@ -698,8 +698,24 @@ func expandIndexName(
        // Memoize the table name that was found. tn is a reference to the table name
        // stored in index.Table.
        *tn = tableName
-       tblMutable := tbl.NewBuilder().BuildExistingMutable().(*tabledesc.Mutable)
-       return &tableName, tblMutable, nil
+
+       // We don't have a handle to the mutable descriptor, only the immutable. We
+       // don't know whether this is a newly created table or an existing table,
+       // so we can't build the mutable from the immutable. Instead, now that we
+       // know the right name, go resolve the corresponding mutable descriptor.
+       _, obj, err := sc.Accessor().GetObjectDesc(
+               ctx, txn, tn.Catalog(), tn.Schema(), string(tn.ObjectName), tree.ObjectLookupFlags{
+                       CommonLookupFlags: tree.CommonLookupFlags{
+                               Required:       true,
+                               RequireMutable: true,
+                       },
+                       DesiredObjectKind: tree.TableObject,
+               },
+       )
+       if err != nil {
+               return nil, nil, err
+       }
+       return &tableName, obj.(*tabledesc.Mutable), nil
 }
 

[chengxiong-ruan]

@ajwerner I'm gonna leave it as it is since we both don't have strong feelings.

Reviewable status: complete! 0 of 0 LGTMs obtained (waiting on @ajwerner)

[ajwerner]

Reviewed 2 of 2 files at r1, all commit messages.
Reviewable status: complete! 1 of 0 LGTMs obtained (waiting on @chengxiong-ruan)

[ajwerner]

does some form of this need a backport to release-22.1?

Reviewable status: complete! 1 of 0 LGTMs obtained (waiting on @chengxiong-ruan)

[chengxiong-ruan]

good question. backport is not needed, because it was doing the right thing, and my refactoring did it wrong and it was not backported

[chengxiong-ruan]

TFTR! @ajwerner
bors r+

[craig]

Build failed (retrying...):

• GitHub CI (Cockroach)

[craig]

Build succeeded:

• GitHub CI (Cockroach)