Fix allow_experimental_projection_optimization with enable_global_with_statement

[azat]

Changelog category (leave one):

• Bug Fix (user-visible misbehaviour in official stable or prestable release)

Changelog entry (a user-readable short description of the changes that goes to CHANGELOG.md):
Fix allow_experimental_projection_optimization with enable_global_with_statement (before it may lead to Stack size too large error in case of multiple expressions in WITH clause, and also it executes scalar subqueries again and again, so not it will be more optimal).

allow_experimental_projection_optimization requires one more
InterpreterSelectQuery, which with enable_global_with_statement will
apply ApplyWithAliasVisitor if the query is not subquery.

But this should not be done for queries from
MergeTreeData::getQueryProcessingStage()/getQueryProcessingStageWithAggregateProjections()
since this will duplicate WITH statements over and over.

This will also fix scalar.xml perf tests, that leads to the following
error now:

scalar.query0.prewarm0: DB::Exception: Stack size too large.

And since it has very long query in the log, this leads to the following
perf test error:

_csv.Error: field larger than field limit (131072)

Cc: @amosbird
Refs: #34456

[azat]

Performance Comparison (actions) [4/4] — Errors while building the report.

The error is for the server without this patch (upstream):

$ fgrep scalar.query0.prewarm0 left-server-log.log  | tail -n2 | head -c150; echo
2022.02.16 20:46:17.632470 [ 333 ] {scalar.query0.prewarm0} <Error> TCPHandler: Code: 306. DB::Exception: Stack size too large. Stack address: 0x7fdc2
$ fgrep scalar.query0.prewarm0 right-server-log.log  | tail -n2
2022.02.16 20:46:18.496707 [ 336 ] {scalar.query0.prewarm0} <Information> executeQuery: Read 10000013 rows, 114.47 MiB in 0.852315803 sec., 11732755 rows/sec., 134.31 MiB/sec.
2022.02.16 20:46:18.497345 [ 336 ] {scalar.query0.prewarm0} <Debug> MemoryTracker: Peak memory usage (for query): 23.93 MiB.