Fix crash when rewriting IN to EXISTS in some case

[davenger]

Fixes #88569
The crash was caused by 1st argument of original IN function being reset to nullptr here:

ClickHouse/src/Analyzer/Resolve/resolveFunction.cpp

Lines 390 to 395 in da741fb

	auto & copy_of_in_first_parameter = function_in_arguments_nodes[0];
	auto subquery_projection = std::make_shared<IdentifierNode>(Identifier{unique_column_name});
	equals_function_node_ptr->getArguments().getNodes() = {
	std::move(copy_of_in_first_parameter), /// x

Even though the function gets replaced in the query tree, in some case resolveFunction might be called again on a copy of the original FunctionNode pointer, e.g. in case when IN is in IF condition:

ClickHouse/src/Analyzer/Resolve/resolveFunction.cpp

Line 250 in da741fb

resolveExpressionNode(if_function_condition, scope, false /*allow_lambda_expression*/, false /*allow_table_expression*/);

Changelog category (leave one):

• Not for changelog (changelog entry is not required)

Changelog entry (a user-readable short description of the changes that goes into CHANGELOG.md):

...

Documentation entry for user-facing changes

• Documentation is written (mandatory for new features)

[clickhouse-gh]

Workflow [PR], commit [db280c6]

Summary: ❌

job_name	test_name	status	info	comment
Integration tests (amd_binary, 2/5)		failure
	test_merge_tree_s3/test.py::test_merge_canceled_by_s3_errors[node-broken_s3_always_multi_part]	FAIL	cidb	#87560
Performance Comparison (amd_release, master_head, 3/3)		failure
	Configure	failure

[scanhex12]

LGTM, but please change label to "critical bugfix"

[davenger]

but please change label to "critical bugfix"

AFAIK even though this is a crash, as long as this is a bug in new functionality that hasn't been included in any release yet it's not considered a "bugfix" from changelog standpoint.

[clickgapai]

Hi @davenger @scanhex12 — while reviewing this PR I found the following:

• Bug (P1): Shared-ptr copy (not clone) of IN first argument at line 616 causes NOT_FOUND_COLUMN_IN_BLOCK for multi-column tuple IN inside IF — Shared-ptr copy (not clone) of IN first argument at line 616 causes NOT_FOUND_COLUMN_IN_BLOCK for multi-column tuple IN inside IF #102672

Happy to discuss — close anything that's wrong or already addressed.