Unified: Extract defer statements

Author: paldepind

unified/extractor/ast_types.yml

@@ -64,6 +64,7 @@ supertypes:
     - for_each_stmt
     - while_stmt
     - do_while_stmt
+    - defer_stmt
     - labeled_stmt
   callable:
     - top_level
@@ -346,6 +347,9 @@ named:
     body?: block
     condition: expr
 
+  defer_stmt:
+    body: expr
+
   # `do { body } catch pattern { ... } catch ...`. Swift uses `do`/`catch`
   # for error handling; for languages with `try`/`catch`, this is the same shape.
   try_expr:

unified/extractor/src/languages/swift/swift.rs

@@ -932,6 +932,8 @@ fn translation_rules() -> Vec<Rule<SwiftContext>> {
             =>
             (do_while_stmt condition: {cond} body: {body})
         ),
+        // A `defer` statement
+        rule!((deferStmt body: @body) => (defer_stmt body: {body})),
         // A labeled statement (`outer: for … { }`). swift-syntax stores the
         // label and colon as separate tokens, so the label token is already the
         // bare name (no trailing `:` to strip).
@@ -1394,7 +1396,6 @@ fn translation_rules() -> Vec<Rule<SwiftContext>> {
         rule!((compositionType) => (unsupported_node)),
         rule!((consumeExpr) => (unsupported_node)),
         rule!((copyExpr) => (unsupported_node)),
-        rule!((deferStmt) => (unsupported_node)),
         rule!((discardStmt) => (unsupported_node)),
         rule!((fallThroughStmt) => (unsupported_node)),
         rule!((ifConfigDecl) => (unsupported_node)),

unified/extractor/tests/corpus/swift/control-flow/defer-statement.output

@@ -84,7 +84,15 @@ top_level source="⟨body⟩"
           body:
             block source="{\n  ⟨stmt⟩\n  ⟨stmt⟩\n}"
               stmt:
-                unsupported_node "defer { print(\"cleanup\") }" source="defer { print(\"cleanup\") }"
+                defer_stmt source="defer ⟨body⟩"
+                  body:
+                    block source="{ ⟨stmt⟩ }"
+                      stmt:
+                        call_expr source="⟨callee⟩(⟨argument⟩)"
+                          callee: identifier "print" source="print"
+                          argument:
+                            argument source="⟨value⟩"
+                              value: string_literal "\"cleanup\"" source="\"cleanup\""
                 call_expr source="⟨callee⟩(⟨argument⟩)"
                   callee: identifier "print" source="print"
                   argument:

unified/ql/lib/codeql/unified/internal/Ast.qll

@@ -528,6 +528,18 @@ module Unified {
     }
   }
 
+  /** A class representing `defer_stmt` nodes. */
+  class DeferStmt extends @unified_defer_stmt, F::Stmt {
+    /** Gets the name of the primary QL class for this element. */
+    final override string getAPrimaryQlClass() { result = "DeferStmt" }
+
+    /** Gets the node corresponding to the field `body`. */
+    final F::Expr getBody() { unified_defer_stmt_def(this, result) }
+
+    /** Gets a field or child node of this node. */
+    final override F::AstNode getAFieldOrChild() { unified_defer_stmt_def(this, result) }
+  }
+
   /** A class representing `destructor_declaration` nodes. */
   class DestructorDeclaration extends @unified_destructor_declaration, F::Callable, F::Member,
     F::Stmt
@@ -1619,6 +1631,8 @@ module Unified {
       or
       result = node.(ContinueExpr).getLabelNameNode() and i = -1 and name = "getLabelNameNode"
       or
+      result = node.(DeferStmt).getBody() and i = -1 and name = "getBody"
+      or
       result = node.(DestructorDeclaration).getBody() and i = -1 and name = "getBody"
       or
       result = node.(DestructorDeclaration).getModifier(i) and name = "getModifier"
@@ -1874,6 +1888,8 @@ module UnifiedFinal {
 
   final class ContinueExpr = F::ContinueExpr;
 
+  final class DeferStmt = F::DeferStmt;
+
   final class DestructorDeclaration = F::DestructorDeclaration;
 
   final class DoWhileStmt = F::DoWhileStmt;

unified/ql/lib/unified.dbscheme

@@ -400,6 +400,11 @@ unified_continue_expr_def(
   unique int id: @unified_continue_expr
 );
 
+unified_defer_stmt_def(
+  unique int id: @unified_defer_stmt,
+  int body: @unified_expr ref
+);
+
 #keyset[unified_destructor_declaration, index]
 unified_destructor_declaration_modifier(
   int unified_destructor_declaration: @unified_destructor_declaration ref,
@@ -746,7 +751,7 @@ unified_return_expr_def(
   unique int id: @unified_return_expr
 );
 
-@unified_stmt = @unified_accessor_declaration | @unified_class_like_declaration | @unified_constructor_declaration | @unified_destructor_declaration | @unified_do_while_stmt | @unified_expr | @unified_for_each_stmt | @unified_function_declaration | @unified_guard_if_stmt | @unified_import_declaration | @unified_labeled_stmt | @unified_operator_syntax_declaration | @unified_type_alias_declaration | @unified_variable_declaration | @unified_while_stmt
+@unified_stmt = @unified_accessor_declaration | @unified_class_like_declaration | @unified_constructor_declaration | @unified_defer_stmt | @unified_destructor_declaration | @unified_do_while_stmt | @unified_expr | @unified_for_each_stmt | @unified_function_declaration | @unified_guard_if_stmt | @unified_import_declaration | @unified_labeled_stmt | @unified_operator_syntax_declaration | @unified_type_alias_declaration | @unified_variable_declaration | @unified_while_stmt
 
 #keyset[unified_string_interpolation_expr, index]
 unified_string_interpolation_expr_element(
@@ -1001,7 +1006,7 @@ unified_trivia_tokeninfo(
   string value: string ref
 );
 
-@unified_ast_node = @unified_accessor_declaration | @unified_argument | @unified_array_literal | @unified_associated_type_declaration | @unified_base_type | @unified_binary_expr | @unified_block | @unified_bound_type_constraint | @unified_break_expr | @unified_bulk_importing_pattern | @unified_call_expr | @unified_catch_clause | @unified_class_like_declaration | @unified_conditional_pattern | @unified_constructor_declaration | @unified_continue_expr | @unified_destructor_declaration | @unified_do_while_stmt | @unified_equality_type_constraint | @unified_expr_pattern | @unified_for_each_stmt | @unified_function_declaration | @unified_function_expr | @unified_generic_type_expr | @unified_guard_if_stmt | @unified_if_expr | @unified_import_declaration | @unified_initializer_declaration | @unified_key_value_pair | @unified_labeled_stmt | @unified_map_literal | @unified_member_access_expr | @unified_named_pattern | @unified_operator_syntax_declaration | @unified_or_pattern | @unified_parameter | @unified_pattern_guard_expr | @unified_return_expr | @unified_string_interpolation_expr | @unified_switch_case | @unified_switch_expr | @unified_throw_expr | @unified_token | @unified_top_level | @unified_trivia_token | @unified_try_expr | @unified_tuple_expr | @unified_type_alias_declaration | @unified_type_cast_expr | @unified_type_parameter | @unified_type_test_expr | @unified_unary_expr | @unified_unresolved_operator_sequence | @unified_variable_declaration | @unified_while_stmt
+@unified_ast_node = @unified_accessor_declaration | @unified_argument | @unified_array_literal | @unified_associated_type_declaration | @unified_base_type | @unified_binary_expr | @unified_block | @unified_bound_type_constraint | @unified_break_expr | @unified_bulk_importing_pattern | @unified_call_expr | @unified_catch_clause | @unified_class_like_declaration | @unified_conditional_pattern | @unified_constructor_declaration | @unified_continue_expr | @unified_defer_stmt | @unified_destructor_declaration | @unified_do_while_stmt | @unified_equality_type_constraint | @unified_expr_pattern | @unified_for_each_stmt | @unified_function_declaration | @unified_function_expr | @unified_generic_type_expr | @unified_guard_if_stmt | @unified_if_expr | @unified_import_declaration | @unified_initializer_declaration | @unified_key_value_pair | @unified_labeled_stmt | @unified_map_literal | @unified_member_access_expr | @unified_named_pattern | @unified_operator_syntax_declaration | @unified_or_pattern | @unified_parameter | @unified_pattern_guard_expr | @unified_return_expr | @unified_string_interpolation_expr | @unified_switch_case | @unified_switch_expr | @unified_throw_expr | @unified_token | @unified_top_level | @unified_trivia_token | @unified_try_expr | @unified_tuple_expr | @unified_type_alias_declaration | @unified_type_cast_expr | @unified_type_parameter | @unified_type_test_expr | @unified_unary_expr | @unified_unresolved_operator_sequence | @unified_variable_declaration | @unified_while_stmt
 
 unified_ast_node_location(
   unique int node: @unified_ast_node ref,

unified/ql/test/library-tests/controlflow/basicblock-slices.expected

@@ -129,9 +129,14 @@
 | 152 | cfg.swift:152:7:152:18 | Block | 'Block -V false -^ ReturnExpr' |
 | 168 | cfg.swift:168:1:184:1 | FunctionDeclaration | 'FunctionDeclaration' |
 | 168 | cfg.swift:168:16:168:16 | x | 'x -> Block' |
-| 170 | cfg.swift:170:3:172:3 |  | '' |
-| 174 | cfg.swift:174:3:176:3 |  | '' |
-| 178 | cfg.swift:178:3:183:3 |  | '' |
+| 170 | cfg.swift:170:3:172:3 | DeferStmt | 'DeferStmt -V Block' |
+| 171 | cfg.swift:171:5:171:9 | print | 'print -> Argument -V "4" -^ print(...)' |
+| 174 | cfg.swift:174:3:176:3 | DeferStmt | 'DeferStmt -V Block' |
+| 175 | cfg.swift:175:5:175:9 | print | 'print -> Argument -V "3" -^ print(...)' |
+| 178 | cfg.swift:178:3:183:3 | DeferStmt | 'DeferStmt -V Block' |
+| 179 | cfg.swift:179:5:179:9 | print | 'print -> Argument -V "1" -^ print(...)' |
+| 180 | cfg.swift:180:6:182:5 | DeferStmt | 'DeferStmt -V Block' |
+| 181 | cfg.swift:181:7:181:11 | print | 'print -> Argument -V "2" -^ print(...)' |
 | 186 | cfg.swift:186:1:198:1 | FunctionDeclaration | 'FunctionDeclaration' |
 | 186 | cfg.swift:186:9:186:9 | x | 'x -> Block' |
 | 187 | cfg.swift:187:3:197:3 | IfExpr | 'IfExpr -V x -> 2 -^ ... > ...' |