prism 1.8.0 → 1.8.2

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 2344922b08aa30076aab32c5a92e0d7f21a05f03bb85d089cc69e228a71e3b20
-  data.tar.gz: 7a44533dd2827ec9f6c31fc69c533c9f90b88520aef71219a348aa61fa460fbd
+  metadata.gz: 1c783657a881bf48ea7f582fd8460ce7230ffc142a67d6eab53f06c5878d2739
+  data.tar.gz: 761e009ead12f8984342f4aa44a8efb8e2c312b1b19527ba9e6399ba8a9990aa
 SHA512:
-  metadata.gz: c278e7b881b89f51150850df09062d9a9bd5e5076746e4b86ab199729a588f8aa346e12b6ba9b95dc0cf53592866bf216977bbb7da10cef6bd8a018681e9f45f
-  data.tar.gz: '009dee7695d1f7d58adb68d3829f97ecff445cc0eb8509526979f938932e553abdc2ba601049df67b8b29a75b7d542a9fb06d4c1984131d6f60e2ac9e3cf83c2'
+  metadata.gz: 247eb5ba72e15a0bf3136a45582218c232cb300f34b33304bbc63dcfa2ab8fbd728192aff1f63de3b6d2bf23a18afb1711460e28f8e5ce1cc478f5af6e37b205
+  data.tar.gz: f0ee65d2c2ff74a3bc9fe0709177467fafdaf0a79aa0c43e8e5f82bfc17c855d3546aa6442d45de55dcea4fe06e567879e60d8a23448a151a1e9d539ccfbd1fd

data/CHANGELOG.md

@@ -4,6 +4,25 @@ All notable changes to this project will be documented in this file.
 
 The format is based on [Keep a Changelog](http://keepachangelog.com/en/1.0.0/) and this project adheres to [Semantic Versioning](http://semver.org/spec/v2.0.0.html).
 
+## [Unreleased]
+
+## [1.8.2] - 2026-10-21
+
+### Changed
+
+- Fix build errors with C23 qualifier-preserving `memchr`.
+- Do not warn for ambiguous `/` when targeting Ruby 4.1.
+- Fix array-bounds warnings with newer GCC.
+
+## [1.8.1] - 2026-03-16
+
+### Changed
+
+- Fix `not` binding power in endless methods.
+- Correctly handle `and?` and similar on Ruby 4.0.
+- Fix error message for block/lambda with `...` argument.
+- Fix `in` handling.
+
 ## [1.8.0] - 2026-01-12
 
 ### Added
@@ -19,8 +38,6 @@ The format is based on [Keep a Changelog](http://keepachangelog.com/en/1.0.0/) a
 - Decouple ripper translator from ripper library.
 - Sync Prism::Translation::ParserCurrent with Ruby 4.0.
 
-## [Unreleased]
-
 ## [1.7.0] - 2025-12-18
 
 ### Added
@@ -731,7 +748,9 @@ The format is based on [Keep a Changelog](http://keepachangelog.com/en/1.0.0/) a
 
 - 🎉 Initial release! 🎉
 
-[unreleased]: https://github.com/ruby/prism/compare/v1.8.0...HEAD
+[unreleased]: https://github.com/ruby/prism/compare/v1.8.2...HEAD
+[1.8.2]: https://github.com/ruby/prism/compare/v1.8.1...v1.8.2
+[1.8.1]: https://github.com/ruby/prism/compare/v1.8.0...v1.8.1
 [1.8.0]: https://github.com/ruby/prism/compare/v1.7.0...v1.8.0
 [1.7.0]: https://github.com/ruby/prism/compare/v1.6.0...v1.7.0
 [1.6.0]: https://github.com/ruby/prism/compare/v1.5.2...v1.6.0

data/config.yml

@@ -17,6 +17,8 @@ errors:
   - ARGUMENT_FORWARDING_UNBOUND
   - ARGUMENT_NO_FORWARDING_AMPERSAND
   - ARGUMENT_NO_FORWARDING_ELLIPSES
+  - ARGUMENT_NO_FORWARDING_ELLIPSES_LAMBDA
+  - ARGUMENT_NO_FORWARDING_ELLIPSES_BLOCK
   - ARGUMENT_NO_FORWARDING_STAR
   - ARGUMENT_NO_FORWARDING_STAR_STAR
   - ARGUMENT_SPLAT_AFTER_ASSOC_SPLAT

data/ext/prism/extension.h

@@ -1,7 +1,7 @@
 #ifndef PRISM_EXT_NODE_H
 #define PRISM_EXT_NODE_H
 
-#define EXPECTED_PRISM_VERSION "1.8.0"
+#define EXPECTED_PRISM_VERSION "1.8.2"
 
 #include <ruby.h>
 #include <ruby/encoding.h>

data/include/prism/defines.h

@@ -257,4 +257,22 @@
     #define PRISM_FALLTHROUGH
 #endif
 
+/**
+ * Tell the compiler that the given expression can be assumed to be true. Unlike
+ * assert, this emits no runtime check — it only feeds the optimizer's value
+ * range analysis so it can prune impossible paths. Use it to communicate an
+ * invariant the caller guarantees but that the compiler cannot otherwise prove.
+ */
+#if defined(__clang__)
+    #define PRISM_ASSUME(expr_) __builtin_assume(expr_)
+#elif defined(_MSC_VER)
+    #define PRISM_ASSUME(expr_) __assume(expr_)
+#elif defined(__GNUC__) && (__GNUC__ >= 13)
+    #define PRISM_ASSUME(expr_) __attribute__((assume(expr_)))
+#elif defined(__GNUC__)
+    #define PRISM_ASSUME(expr_) ((expr_) ? (void) 0 : __builtin_unreachable())
+#else
+    #define PRISM_ASSUME(expr_) ((void) 0)
+#endif
+
 #endif

data/include/prism/diagnostic.h

@@ -48,6 +48,8 @@ typedef enum {
     PM_ERR_ARGUMENT_FORWARDING_UNBOUND,
     PM_ERR_ARGUMENT_NO_FORWARDING_AMPERSAND,
     PM_ERR_ARGUMENT_NO_FORWARDING_ELLIPSES,
+    PM_ERR_ARGUMENT_NO_FORWARDING_ELLIPSES_LAMBDA,
+    PM_ERR_ARGUMENT_NO_FORWARDING_ELLIPSES_BLOCK,
     PM_ERR_ARGUMENT_NO_FORWARDING_STAR,
     PM_ERR_ARGUMENT_NO_FORWARDING_STAR_STAR,
     PM_ERR_ARGUMENT_SPLAT_AFTER_ASSOC_SPLAT,

data/include/prism/util/pm_memchr.h

@@ -24,6 +24,6 @@
  * @return A pointer to the first occurrence of the character in the source
  *     string, or NULL if no such character exists.
  */
-void * pm_memchr(const void *source, int character, size_t number, bool encoding_changed, const pm_encoding_t *encoding);
+const void * pm_memchr(const void *source, int character, size_t number, bool encoding_changed, const pm_encoding_t *encoding);
 
 #endif

data/include/prism/version.h

@@ -19,11 +19,11 @@
 /**
  * The patch version of the Prism library as an int.
  */
-#define PRISM_VERSION_PATCH 0
+#define PRISM_VERSION_PATCH 2
 
 /**
  * The version of the Prism library as a constant string.
  */
-#define PRISM_VERSION "1.8.0"
+#define PRISM_VERSION "1.8.2"
 
 #endif

data/lib/prism/serialize.rb

@@ -25,7 +25,7 @@ module Prism
 
     # The patch version of prism that we are expecting to find in the serialized
     # strings.
-    PATCH_VERSION = 0
+    PATCH_VERSION = 2
 
     # Deserialize the dumped output from a request to parse or parse_file.
     #
@@ -354,6 +354,8 @@ module Prism
         :argument_forwarding_unbound,
         :argument_no_forwarding_ampersand,
         :argument_no_forwarding_ellipses,
+        :argument_no_forwarding_ellipses_lambda,
+        :argument_no_forwarding_ellipses_block,
         :argument_no_forwarding_star,
         :argument_no_forwarding_star_star,
         :argument_splat_after_assoc_splat,

data/prism.gemspec

@@ -2,7 +2,7 @@
 
 Gem::Specification.new do |spec|
   spec.name = "prism"
-  spec.version = "1.8.0"
+  spec.version = "1.8.2"
   spec.authors = ["Shopify"]
   spec.email = ["ruby@shopify.com"]
 

data/src/diagnostic.c

@@ -10,7 +10,7 @@
 
 #include "prism/diagnostic.h"
 
-#define PM_DIAGNOSTIC_ID_MAX 324
+#define PM_DIAGNOSTIC_ID_MAX 326
 
 /** This struct holds the data for each diagnostic. */
 typedef struct {
@@ -112,6 +112,8 @@ static const pm_diagnostic_data_t diagnostic_messages[PM_DIAGNOSTIC_ID_MAX] = {
     [PM_ERR_ARGUMENT_FORWARDING_UNBOUND]        = { "unexpected `...` in an non-parenthesized call", PM_ERROR_LEVEL_SYNTAX },
     [PM_ERR_ARGUMENT_NO_FORWARDING_AMPERSAND]   = { "unexpected `&`; no anonymous block parameter", PM_ERROR_LEVEL_SYNTAX },
     [PM_ERR_ARGUMENT_NO_FORWARDING_ELLIPSES]    = { "unexpected ... when the parent method is not forwarding", PM_ERROR_LEVEL_SYNTAX },
+    [PM_ERR_ARGUMENT_NO_FORWARDING_ELLIPSES_LAMBDA] = { "unexpected ... in lambda argument", PM_ERROR_LEVEL_SYNTAX },
+    [PM_ERR_ARGUMENT_NO_FORWARDING_ELLIPSES_BLOCK]  = { "unexpected ... in block argument", PM_ERROR_LEVEL_SYNTAX },
     [PM_ERR_ARGUMENT_NO_FORWARDING_STAR]        = { "unexpected `*`; no anonymous rest parameter", PM_ERROR_LEVEL_SYNTAX },
     [PM_ERR_ARGUMENT_NO_FORWARDING_STAR_STAR]   = { "unexpected `**`; no anonymous keyword rest parameter", PM_ERROR_LEVEL_SYNTAX },
     [PM_ERR_ARGUMENT_SPLAT_AFTER_ASSOC_SPLAT]   = { "unexpected `*` splat argument after a `**` keyword splat argument", PM_ERROR_LEVEL_SYNTAX },
@@ -442,6 +444,8 @@ pm_diagnostic_id_human(pm_diagnostic_id_t diag_id) {
         case PM_ERR_ARGUMENT_FORWARDING_UNBOUND: return "argument_forwarding_unbound";
         case PM_ERR_ARGUMENT_NO_FORWARDING_AMPERSAND: return "argument_no_forwarding_ampersand";
         case PM_ERR_ARGUMENT_NO_FORWARDING_ELLIPSES: return "argument_no_forwarding_ellipses";
+        case PM_ERR_ARGUMENT_NO_FORWARDING_ELLIPSES_LAMBDA: return "argument_no_forwarding_ellipses_lambda";
+        case PM_ERR_ARGUMENT_NO_FORWARDING_ELLIPSES_BLOCK: return "argument_no_forwarding_ellipses_block";
         case PM_ERR_ARGUMENT_NO_FORWARDING_STAR: return "argument_no_forwarding_star";
         case PM_ERR_ARGUMENT_NO_FORWARDING_STAR_STAR: return "argument_no_forwarding_star_star";
         case PM_ERR_ARGUMENT_SPLAT_AFTER_ASSOC_SPLAT: return "argument_splat_after_assoc_splat";
@@ -757,6 +761,7 @@ pm_diagnostic_id_human(pm_diagnostic_id_t diag_id) {
 static inline const char *
 pm_diagnostic_message(pm_diagnostic_id_t diag_id) {
     assert(diag_id < PM_DIAGNOSTIC_ID_MAX);
+    PRISM_ASSUME(diag_id < PM_DIAGNOSTIC_ID_MAX);
 
     const char *message = diagnostic_messages[diag_id].message;
     assert(message);
@@ -767,6 +772,7 @@ pm_diagnostic_message(pm_diagnostic_id_t diag_id) {
 static inline uint8_t
 pm_diagnostic_level(pm_diagnostic_id_t diag_id) {
     assert(diag_id < PM_DIAGNOSTIC_ID_MAX);
+    PRISM_ASSUME(diag_id < PM_DIAGNOSTIC_ID_MAX);
 
     return (uint8_t) diagnostic_messages[diag_id].level;
 }

data/src/prism.c

@@ -9979,8 +9979,21 @@ parser_lex(pm_parser_t *parser) {
                                 following && (
                                     (peek_at(parser, following) == '&' && peek_at(parser, following + 1) == '&') ||
                                     (peek_at(parser, following) == '|' && peek_at(parser, following + 1) == '|') ||
-                                    (peek_at(parser, following) == 'a' && peek_at(parser, following + 1) == 'n' && peek_at(parser, following + 2) == 'd' && !char_is_identifier(parser, following + 3, parser->end - (following + 3))) ||
-                                    (peek_at(parser, following) == 'o' && peek_at(parser, following + 1) == 'r' && !char_is_identifier(parser, following + 2, parser->end - (following + 2)))
+                                    (
+                                        peek_at(parser, following) == 'a' &&
+                                        peek_at(parser, following + 1) == 'n' &&
+                                        peek_at(parser, following + 2) == 'd' &&
+                                        peek_at(parser, next_content + 3) != '!' &&
+                                        peek_at(parser, next_content + 3) != '?' &&
+                                        !char_is_identifier(parser, following + 3, parser->end - (following + 3))
+                                    ) ||
+                                    (
+                                        peek_at(parser, following) == 'o' &&
+                                        peek_at(parser, following + 1) == 'r' &&
+                                        peek_at(parser, next_content + 2) != '!' &&
+                                        peek_at(parser, next_content + 2) != '?' &&
+                                        !char_is_identifier(parser, following + 2, parser->end - (following + 2))
+                                    )
                                 )
                             ) {
                                 if (!lexed_comment) parser_lex_ignored_newline(parser);
@@ -10051,6 +10064,8 @@ parser_lex(pm_parser_t *parser) {
                                 peek_at(parser, next_content) == 'a' &&
                                 peek_at(parser, next_content + 1) == 'n' &&
                                 peek_at(parser, next_content + 2) == 'd' &&
+                                peek_at(parser, next_content + 3) != '!' &&
+                                peek_at(parser, next_content + 3) != '?' &&
                                 !char_is_identifier(parser, next_content + 3, parser->end - (next_content + 3))
                             ) {
                                 if (!lexed_comment) parser_lex_ignored_newline(parser);
@@ -10067,6 +10082,8 @@ parser_lex(pm_parser_t *parser) {
                             if (
                                 peek_at(parser, next_content) == 'o' &&
                                 peek_at(parser, next_content + 1) == 'r' &&
+                                peek_at(parser, next_content + 2) != '!' &&
+                                peek_at(parser, next_content + 2) != '?' &&
                                 !char_is_identifier(parser, next_content + 2, parser->end - (next_content + 2))
                             ) {
                                 if (!lexed_comment) parser_lex_ignored_newline(parser);
@@ -10757,7 +10774,10 @@ parser_lex(pm_parser_t *parser) {
                     }
 
                     if (lex_state_spcarg_p(parser, space_seen)) {
-                        pm_parser_warn_token(parser, &parser->current, PM_WARN_AMBIGUOUS_SLASH);
+                        // https://bugs.ruby-lang.org/issues/21994
+                        if (parser->version <= PM_OPTIONS_VERSION_CRUBY_4_0) {
+                            pm_parser_warn_token(parser, &parser->current, PM_WARN_AMBIGUOUS_SLASH);
+                        }
                         lex_mode_push_regexp(parser, '\0', '/');
                         LEX(PM_TOKEN_REGEXP_BEGIN);
                     }
@@ -13859,6 +13879,7 @@ parse_parameters(
     bool allows_forwarding_parameters,
     bool accepts_blocks_in_defaults,
     bool in_block,
+    pm_diagnostic_id_t diag_id_forwarding,
     uint16_t depth
 ) {
     pm_do_loop_stack_push(parser, false);
@@ -13914,7 +13935,7 @@ parse_parameters(
             }
             case PM_TOKEN_UDOT_DOT_DOT: {
                 if (!allows_forwarding_parameters) {
-                    pm_parser_err_current(parser, PM_ERR_ARGUMENT_NO_FORWARDING_ELLIPSES);
+                    pm_parser_err_current(parser, diag_id_forwarding);
                 }
 
                 bool succeeded = update_parameter_state(parser, &parser->current, &order);
@@ -14594,6 +14615,7 @@ parse_block_parameters(
             false,
             accepts_blocks_in_defaults,
             true,
+            is_lambda_literal ? PM_ERR_ARGUMENT_NO_FORWARDING_ELLIPSES_LAMBDA : PM_ERR_ARGUMENT_NO_FORWARDING_ELLIPSES_BLOCK,
             (uint16_t) (depth + 1)
         );
         if (!is_lambda_literal) {
@@ -18836,7 +18858,17 @@ parse_expression_prefix(pm_parser_t *parser, pm_binding_power_t binding_power, b
                     if (match1(parser, PM_TOKEN_PARENTHESIS_RIGHT)) {
                         params = NULL;
                     } else {
-                        params = parse_parameters(parser, PM_BINDING_POWER_DEFINED, true, false, true, true, false, (uint16_t) (depth + 1));
+                        params = parse_parameters(
+                            parser,
+                            PM_BINDING_POWER_DEFINED,
+                            true,
+                            false,
+                            true,
+                            true,
+                            false,
+                            PM_ERR_ARGUMENT_NO_FORWARDING_ELLIPSES,
+                            (uint16_t) (depth + 1)
+                        );
                     }
 
                     lex_state_set(parser, PM_LEX_STATE_BEG);
@@ -18861,7 +18893,17 @@ parse_expression_prefix(pm_parser_t *parser, pm_binding_power_t binding_power, b
 
                     lparen = not_provided(parser);
                     rparen = not_provided(parser);
-                    params = parse_parameters(parser, PM_BINDING_POWER_DEFINED, false, false, true, true, false, (uint16_t) (depth + 1));
+                    params = parse_parameters(
+                        parser,
+                        PM_BINDING_POWER_DEFINED,
+                        false,
+                        false,
+                        true,
+                        true,
+                        false,
+                        PM_ERR_ARGUMENT_NO_FORWARDING_ELLIPSES,
+                        (uint16_t) (depth + 1)
+                    );
 
                     // Reject `def * = 1` and similar. We have to specifically check
                     // for them because they create ambiguity with optional arguments.
@@ -18917,7 +18959,7 @@ parse_expression_prefix(pm_parser_t *parser, pm_binding_power_t binding_power, b
                     context_push(parser, PM_CONTEXT_RESCUE_MODIFIER);
 
                     pm_token_t rescue_keyword = parser->previous;
-                    pm_node_t *value = parse_expression(parser, pm_binding_powers[PM_TOKEN_KEYWORD_RESCUE_MODIFIER].right, false, false, PM_ERR_RESCUE_MODIFIER_VALUE, (uint16_t) (depth + 1));
+                    pm_node_t *value = parse_expression(parser, PM_BINDING_POWER_MATCH + 1, false, false, PM_ERR_RESCUE_MODIFIER_VALUE, (uint16_t) (depth + 1));
                     context_pop(parser);
 
                     statement = UP(pm_rescue_modifier_node_create(parser, statement, &rescue_keyword, value));
@@ -20214,7 +20256,7 @@ parse_assignment_value(pm_parser_t *parser, pm_binding_power_t previous_binding_
         pm_token_t rescue = parser->current;
         parser_lex(parser);
 
-        pm_node_t *right = parse_expression(parser, pm_binding_powers[PM_TOKEN_KEYWORD_RESCUE_MODIFIER].right, false, false, PM_ERR_RESCUE_MODIFIER_VALUE, (uint16_t) (depth + 1));
+        pm_node_t *right = parse_expression(parser, PM_BINDING_POWER_MATCH + 1, false, false, PM_ERR_RESCUE_MODIFIER_VALUE, (uint16_t) (depth + 1));
         context_pop(parser);
 
         return UP(pm_rescue_modifier_node_create(parser, value, &rescue, right));
@@ -20320,7 +20362,7 @@ parse_assignment_values(pm_parser_t *parser, pm_binding_power_t previous_binding
             }
         }
 
-        pm_node_t *right = parse_expression(parser, pm_binding_powers[PM_TOKEN_KEYWORD_RESCUE_MODIFIER].right, accepts_command_call_inner, false, PM_ERR_RESCUE_MODIFIER_VALUE, (uint16_t) (depth + 1));
+        pm_node_t *right = parse_expression(parser, PM_BINDING_POWER_MATCH + 1, accepts_command_call_inner, false, PM_ERR_RESCUE_MODIFIER_VALUE, (uint16_t) (depth + 1));
         context_pop(parser);
 
         return UP(pm_rescue_modifier_node_create(parser, value, &rescue, right));
@@ -21639,12 +21681,6 @@ parse_expression(pm_parser_t *parser, pm_binding_power_t binding_power, bool acc
      ) {
         node = parse_expression_infix(parser, node, binding_power, current_binding_powers.right, accepts_command_call, (uint16_t) (depth + 1));
 
-        if (context_terminator(parser->current_context->context, &parser->current)) {
-            // If this token terminates the current context, then we need to
-            // stop parsing the expression, as it has become a statement.
-            return node;
-        }
-
         switch (PM_NODE_TYPE(node)) {
             case PM_MULTI_WRITE_NODE:
                 // Multi-write nodes are statements, and cannot be followed by
@@ -21757,6 +21793,17 @@ parse_expression(pm_parser_t *parser, pm_binding_power_t binding_power, bool acc
                     break;
             }
         }
+
+        if (context_terminator(parser->current_context->context, &parser->current)) {
+            pm_binding_powers_t next_binding_powers = pm_binding_powers[parser->current.type];
+            if (
+                !next_binding_powers.binary ||
+                binding_power > next_binding_powers.left ||
+                (PM_NODE_TYPE_P(node, PM_CALL_NODE) && pm_call_node_command_p((pm_call_node_t *) node))
+            ) {
+                return node;
+            }
+        }
     }
 
     return node;

data/src/util/pm_memchr.c

@@ -7,7 +7,7 @@
  * we need to search for a character in a buffer that could be the trailing byte
  * of a multibyte character.
  */
-void *
+const void *
 pm_memchr(const void *memory, int character, size_t number, bool encoding_changed, const pm_encoding_t *encoding) {
     if (encoding_changed && encoding->multibyte && character >= PRISM_MEMCHR_TRAILING_BYTE_MINIMUM) {
         const uint8_t *source = (const uint8_t *) memory;

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: prism
 version: !ruby/object:Gem::Version
-  version: 1.8.0
+  version: 1.8.2
 platform: ruby
 authors:
 - Shopify