feat: augment C with semgrep constructs

lang/semgrep-grammars/src/semgrep-c/grammar.js

@@ -10,22 +10,73 @@ module.exports = grammar(base_grammar, {
   name: 'c',
 
   conflicts: ($, previous) => previous.concat([
+    // This conflict arises from the case of
+    // 'if' parenthesized_expression semgrep_ellipsis_metavar • '('  …
+    // we don't know if we should reduce the whole if or just the metavar
+    // not a very realistic case, so I don't care to fix this here
+    [$.if_statement, $._expression]
   ]),
 
   /*
      Support for semgrep ellipsis ('...') and metavariables ('$FOO'),
      if they're not already part of the base grammar.
   */
   rules: {
-  /*
     semgrep_ellipsis: $ => '...',
+    semgrep_deep_expression: $ => seq('<...', $._expression, '...>'),
+    semgrep_typed_metavar: $ => seq('(', $.type_descriptor, $.semgrep_metavariable, ')'),
+    semgrep_metavariable: $ => /\$[A-Z_][A-Z_0-9]*/,
+    semgrep_ellipsis_metavar: $ => /\$\.\.\.[a-zA-Z_][a-zA-Z_0-9]*/,
+
+    // Alternate "entry point". Allows parsing a standalone expression.
+    semgrep_expression: ($) => seq("__SEMGREP_EXPRESSION", $._expression),
 
     _expression: ($, previous) => {
       return choice(
         $.semgrep_ellipsis,
+        $.semgrep_deep_expression,
+        $.semgrep_typed_metavar,
+        $.semgrep_ellipsis_metavar,
         ...previous.members
-      );
-    }
-  */
+      )
+    },
+
+    _statement: ($, previous) => {
+      return choice(
+        ...previous.members,
+        // This needs to have a little more precedence, so that we can parse
+        // { ... <stmts> }
+        // properly, such that `...` is for the semgrep ellipsis.
+        prec(1, $.semgrep_ellipsis),
+        prec(1, $.semgrep_ellipsis_metavar),
+      )
+    },
+
+    _top_level_statement: ($, previous) => {
+      return choice(
+        ...previous.members,
+        prec(1, $.semgrep_ellipsis),
+        prec(1, $.semgrep_ellipsis_metavar)
+      )
+    },
+
+    _for_statement_body: ($, previous) => choice(
+      previous,
+      $.semgrep_ellipsis
+    ),
+
+    _field_identifier: ($, previous) => choice(
+      previous,
+      $.semgrep_ellipsis,
+      $.semgrep_ellipsis_metavar
+    ),
+
+    // don't need to edit parameter_declaration because variadic_parameter already exists
+
+    // Alternative entry point for pattern parsing
+    translation_unit: ($, previous) => choice(
+      previous,
+      $.semgrep_expression
+    )
   }
 });

lang/semgrep-grammars/src/semgrep-c/test/corpus/semgrep.txt

@@ -0,0 +1,301 @@
+======================================================================
+Deep expression
+======================================================================
+
+__SEMGREP_EXPRESSION <... foo($X) ...>
+
+---
+
+(translation_unit
+  (semgrep_expression
+    (semgrep_deep_expression
+      (call_expression
+        (identifier)
+        (argument_list
+          (identifier))))))
+
+======================================================================
+Typed metavariable
+======================================================================
+
+__SEMGREP_EXPRESSION ($T $X)
+
+---
+
+(translation_unit
+  (semgrep_expression
+    (semgrep_typed_metavar
+      (type_descriptor
+        (type_identifier))
+      (semgrep_metavariable))))
+
+======================================================================
+Sizeof ellipsis
+======================================================================
+
+__SEMGREP_EXPRESSION sizeof(...)
+
+---
+
+(translation_unit
+  (semgrep_expression
+    (sizeof_expression
+      (parenthesized_expression
+        (semgrep_ellipsis)))))
+
+======================================================================
+Ellipsis expression
+======================================================================
+
+__SEMGREP_EXPRESSION ...
+
+---
+
+(translation_unit
+  (semgrep_expression
+    (semgrep_ellipsis)))
+
+======================================================================
+Ellipsis argument
+======================================================================
+
+foo(..., 1, ..., 2, ...);
+
+---
+
+(translation_unit
+  (expression_statement
+    (call_expression
+      (identifier)
+      (argument_list
+        (semgrep_ellipsis)
+        (number_literal)
+        (semgrep_ellipsis)
+        (number_literal)
+        (semgrep_ellipsis)))))
+
+======================================================================
+Ellipsis statement
+======================================================================
+
+int x = 2;
+...
+int y = 3;
+
+---
+
+(translation_unit
+  (declaration
+    (primitive_type)
+    (init_declarator
+      (identifier)
+      (number_literal)))
+  (semgrep_ellipsis)
+  (declaration
+    (primitive_type)
+    (init_declarator
+      (identifier)
+      (number_literal))))
+
+======================================================================
+Ellipsis statement in function
+======================================================================
+
+int main() {
+  int x = 2;
+  ...
+  int y = 3;
+}
+
+---
+
+(translation_unit
+  (function_definition
+    (primitive_type)
+    (function_declarator
+      (identifier)
+      (parameter_list))
+    (compound_statement
+      (declaration
+        (primitive_type)
+        (init_declarator
+          (identifier)
+          (number_literal)))
+      (semgrep_ellipsis)
+      (declaration
+        (primitive_type)
+        (init_declarator
+          (identifier)
+          (number_literal))))))
+
+======================================================================
+Ellipsis in for loop
+======================================================================
+
+for (...) {}
+
+---
+
+(translation_unit
+  (for_statement
+    (semgrep_ellipsis)
+    (compound_statement)))
+
+======================================================================
+Metavariables
+======================================================================
+
+__SEMGREP_EXPRESSION $X
+
+---
+
+(translation_unit
+  (semgrep_expression
+    (identifier)))
+
+======================================================================
+Metavariables in function definition
+======================================================================
+
+$T $F($T2 $ARG) {
+  ...
+}
+
+---
+
+(translation_unit
+  (function_definition
+    (type_identifier)
+    (function_declarator
+      (identifier)
+      (parameter_list
+        (parameter_declaration
+          (type_identifier)
+          (identifier))))
+    (compound_statement
+      (semgrep_ellipsis))))
+
+======================================================================
+Ellipsis metavariable
+======================================================================
+
+$...X
+
+---
+
+(translation_unit
+  (semgrep_ellipsis_metavar))
+
+======================================================================
+Ellipsis metavariable as expression
+======================================================================
+
+__SEMGREP_EXPRESSION $...X
+
+---
+
+(translation_unit
+  (semgrep_expression
+    (semgrep_ellipsis_metavar)))
+
+======================================================================
+Ellipsis metavariable between statements
+======================================================================
+
+int x = 1;
+$...STMTS
+int y = 2;
+
+---
+
+(translation_unit
+  (declaration
+    (primitive_type)
+    (init_declarator
+      (identifier)
+      (number_literal)))
+  (semgrep_ellipsis_metavar)
+  (declaration
+    (primitive_type)
+    (init_declarator
+      (identifier)
+      (number_literal))))
+
+======================================================================
+Ellipsis metavariable between statements in function
+======================================================================
+
+int main(){
+  int x = 1;
+  $...STMTS
+  int y = 2;
+}
+
+---
+
+(translation_unit
+  (function_definition
+    (primitive_type)
+    (function_declarator
+      (identifier)
+      (parameter_list))
+    (compound_statement
+      (declaration
+        (primitive_type)
+        (init_declarator
+          (identifier)
+          (number_literal)))
+      (semgrep_ellipsis_metavar)
+      (declaration
+        (primitive_type)
+        (init_declarator
+          (identifier)
+          (number_literal))))))
+
+======================================================================
+Ellipsis in dot access
+======================================================================
+
+__SEMGREP_EXPRESSION a. ... .c
+
+---
+
+(translation_unit
+  (semgrep_expression
+    (field_expression
+      (field_expression
+        (identifier)
+        (semgrep_ellipsis))
+      (field_identifier))))
+
+======================================================================
+Ellipsis in dot access
+======================================================================
+
+__SEMGREP_EXPRESSION a. ... .c
+
+---
+
+(translation_unit
+  (semgrep_expression
+    (field_expression
+      (field_expression
+        (identifier)
+        (semgrep_ellipsis))
+      (field_identifier))))
+
+======================================================================
+Ellipsis metavariable in dot access
+======================================================================
+
+__SEMGREP_EXPRESSION a. $...STUFF .c
+
+---
+
+(translation_unit
+  (semgrep_expression
+    (field_expression
+      (field_expression
+        (identifier)
+        (semgrep_ellipsis_metavar))
+      (field_identifier))))

lang/semgrep-grammars/src/semgrep-cpp/grammar.js

@@ -93,10 +93,16 @@ module.exports = grammar(base_grammar, {
       prec(1,$.semgrep_ellipsis)
     ),
 
+    _for_statement_body: ($, previous) => choice(
+      previous,
+      $.semgrep_ellipsis
+    ),
+
     // For method chaining, like foo. ... .bar()
     _field_identifier: ($, previous) => choice(
       previous,
-      $.semgrep_ellipsis
+      $.semgrep_ellipsis,
+      $.semgrep_named_ellipsis
     ),
 
     // So we prefer to parse a unary left fold for