[pull] master from ruby:master

.github/workflows/zjit-macos.yml

@@ -98,7 +98,7 @@ jobs:
           rustup install ${{ matrix.rust_version }} --profile minimal
           rustup default ${{ matrix.rust_version }}
 
-      - uses: taiki-e/install-action@b6b84cf49ebfe0176417bdce007c624f0db37f20 # v2.86.2
+      - uses: taiki-e/install-action@5b4d68e2e660441203ab128a23676f1e4faf1532 # v2.86.3
         with:
           tool: nextest@0.9
         if: ${{ matrix.test_task == 'zjit-check' }}

.github/workflows/zjit-ubuntu.yml

@@ -152,7 +152,7 @@ jobs:
           ruby-version: '3.1'
           bundler: none
 
-      - uses: taiki-e/install-action@b6b84cf49ebfe0176417bdce007c624f0db37f20 # v2.86.2
+      - uses: taiki-e/install-action@5b4d68e2e660441203ab128a23676f1e4faf1532 # v2.86.3
         with:
           tool: nextest@0.9
         if: ${{ matrix.test_task == 'zjit-check' }}

NEWS.md

@@ -179,7 +179,7 @@ They are still available on rubygems.org and can be installed with
   * 0.6.3 to [v0.6.4][pp-v0.6.4]
 * prism 1.9.0
   * 1.7.0 to [v1.8.0][prism-v1.8.0], [v1.8.1][prism-v1.8.1], [v1.9.0][prism-v1.9.0]
-* psych 5.4.0
+* psych 5.5.0
   * 5.3.1 to [v5.4.0][psych-v5.4.0]
 * resolv 0.7.1
   * 0.7.0 to [v0.7.1][resolv-v0.7.1]

compile.c

@@ -628,17 +628,28 @@ setup_branch(const rb_code_location_t *loc, const char *type, VALUE structure, V
 }
 
 static VALUE
-decl_branch_base(rb_iseq_t *iseq, VALUE key, const rb_code_location_t *loc, const char *type)
+decl_branch_base(rb_iseq_t *iseq, int node_id, const rb_code_location_t *loc, const char *type)
 {
     if (!branch_coverage_valid_p(iseq, loc->beg_pos.lineno)) return Qundef;
 
     /*
-     * if !structure[node]
-     *   structure[node] = [type, first_lineno, first_column, last_lineno, last_column, branches = {}]
+     * A branch base is keyed by [source_hash (in two halves), node_id, first_lineno],
+     * which identifies the branch node stably even across (re-)evals against
+     * the same path.
+     *
+     * if !structure[key]
+     *   structure[key] = [type, first_lineno, first_column, last_lineno, last_column, branches = {}]
      * else
-     *   branches = structure[node][5]
+     *   branches = structure[key][5]
      * end
      */
+    uint64_t source_hash = ISEQ_BODY(iseq)->source_hash;
+    VALUE key = rb_ary_new_from_args(4,
+        ULONG2NUM((unsigned long)(source_hash >> 32)),
+        ULONG2NUM((unsigned long)(source_hash & 0xffffffff)),
+        INT2FIX(node_id),
+        INT2FIX(loc->beg_pos.lineno));
+    rb_ary_freeze(key);
 
     VALUE structure = RARRAY_AREF(ISEQ_BRANCH_COVERAGE(iseq), 0);
     VALUE branch_base = rb_hash_aref(structure, key);
@@ -7082,7 +7093,7 @@ compile_if(rb_iseq_t *iseq, LINK_ANCHOR *const ret, const NODE *const node, int
     ADD_SEQ(ret, cond_seq);
 
     if (then_label->refcnt && else_label->refcnt) {
-        branches = decl_branch_base(iseq, PTR2NUM(node), nd_code_loc(node), type == NODE_IF ? "if" : "unless");
+        branches = decl_branch_base(iseq, nd_node_id(node), nd_code_loc(node), type == NODE_IF ? "if" : "unless");
     }
 
     if (then_label->refcnt) {
@@ -7162,7 +7173,7 @@ compile_case(rb_iseq_t *iseq, LINK_ANCHOR *const ret, const NODE *const orig_nod
 
     CHECK(COMPILE(head, "case base", RNODE_CASE(node)->nd_head));
 
-    branches = decl_branch_base(iseq, PTR2NUM(node), nd_code_loc(node), "case");
+    branches = decl_branch_base(iseq, nd_node_id(node), nd_code_loc(node), "case");
 
     node = RNODE_CASE(node)->nd_body;
     EXPECT_NODE("NODE_CASE", node, NODE_WHEN, COMPILE_NG);
@@ -7267,7 +7278,7 @@ compile_case2(rb_iseq_t *iseq, LINK_ANCHOR *const ret, const NODE *const orig_no
     VALUE branches = Qfalse;
     int branch_id = 0;
 
-    branches = decl_branch_base(iseq, PTR2NUM(orig_node), nd_code_loc(orig_node), "case");
+    branches = decl_branch_base(iseq, nd_node_id(orig_node), nd_code_loc(orig_node), "case");
 
     INIT_ANCHOR(body_seq);
     endlabel = NEW_LABEL(nd_line(node));
@@ -8266,7 +8277,7 @@ compile_case3(rb_iseq_t *iseq, LINK_ANCHOR *const ret, const NODE *const orig_no
     INIT_ANCHOR(body_seq);
     INIT_ANCHOR(cond_seq);
 
-    branches = decl_branch_base(iseq, PTR2NUM(node), nd_code_loc(node), "case");
+    branches = decl_branch_base(iseq, nd_node_id(node), nd_code_loc(node), "case");
 
     node = RNODE_CASE3(node)->nd_body;
     EXPECT_NODE("NODE_CASE3", node, NODE_IN, COMPILE_NG);
@@ -8470,7 +8481,7 @@ compile_loop(rb_iseq_t *iseq, LINK_ANCHOR *const ret, const NODE *const node, in
     if (tmp_label) ADD_LABEL(ret, tmp_label);
 
     ADD_LABEL(ret, redo_label);
-    branches = decl_branch_base(iseq, PTR2NUM(node), nd_code_loc(node), type == NODE_WHILE ? "while" : "until");
+    branches = decl_branch_base(iseq, nd_node_id(node), nd_code_loc(node), type == NODE_WHILE ? "while" : "until");
 
     const NODE *const coverage_node = RNODE_WHILE(node)->nd_body ? RNODE_WHILE(node)->nd_body : node;
     add_trace_branch_coverage(
@@ -9113,7 +9124,7 @@ qcall_branch_start(rb_iseq_t *iseq, LINK_ANCHOR *const recv, VALUE *branches, co
     LABEL *else_label = NEW_LABEL(nd_line(line_node));
     VALUE br = 0;
 
-    br = decl_branch_base(iseq, PTR2NUM(node), nd_code_loc(node), "&.");
+    br = decl_branch_base(iseq, nd_node_id(node), nd_code_loc(node), "&.");
     *branches = br;
     ADD_INSN(recv, line_node, dup);
     ADD_INSNL(recv, line_node, branchnil, else_label);

ext/psych/lib/psych/versions.rb

@@ -2,7 +2,7 @@
 
 module Psych
   # The version of Psych you are using
-  VERSION = '5.4.0'
+  VERSION = '5.5.0'
 
   if RUBY_ENGINE == 'jruby'
     DEFAULT_SNAKEYAML_VERSION = '2.10'.freeze

ext/psych/lib/psych/visitors/to_ruby.rb

@@ -445,9 +445,6 @@ def deduplicate key
         end
       end
 
-      def merge_key hash, key, val
-      end
-
       def revive klass, node
         s = register(node, klass.allocate)
         init_with(s, revive_hash({}, node, true), node)

ext/psych/lib/psych/visitors/yaml_tree.rb

@@ -496,9 +496,6 @@ def visit_hash_subclass o
         end
       end
 
-      def dump_list o
-      end
-
       def dump_exception o, msg
         tag = ['!ruby/exception', o.class.name].join ':'
 

ext/psych/psych_parser.c

@@ -52,19 +52,28 @@ static int io_reader(void * data, unsigned char *buf, size_t size, size_t *read)
     return 1;
 }
 
+/* The parser calls back into Ruby for every event, so a handler can call
+ * Psych::Parser#parse again on the same object.  parse() reinitialises the
+ * parser it is handed, which would pull the input out from under the loop
+ * still driving it, so keep a flag to reject a reentrant call. */
+typedef struct {
+    yaml_parser_t yaml_parser;
+    int parsing;
+} psych_parser_t;
+
 static void dealloc(void * ptr)
 {
-    yaml_parser_t * parser;
+    psych_parser_t * parser;
 
-    parser = (yaml_parser_t *)ptr;
-    yaml_parser_delete(parser);
+    parser = (psych_parser_t *)ptr;
+    yaml_parser_delete(&parser->yaml_parser);
     xfree(parser);
 }
 
 #if 0
 static size_t memsize(const void *ptr)
 {
-    const yaml_parser_t *parser = ptr;
+    const psych_parser_t *parser = ptr;
     /* TODO: calculate parser's size */
     return 0;
 }
@@ -81,10 +90,10 @@ static const rb_data_type_t psych_parser_type = {
 
 static VALUE allocate(VALUE klass)
 {
-    yaml_parser_t * parser;
-    VALUE obj = TypedData_Make_Struct(klass, yaml_parser_t, &psych_parser_type, parser);
+    psych_parser_t * parser;
+    VALUE obj = TypedData_Make_Struct(klass, psych_parser_t, &psych_parser_type, parser);
 
-    yaml_parser_initialize(parser);
+    yaml_parser_initialize(&parser->yaml_parser);
 
     return obj;
 }
@@ -257,18 +266,29 @@ static VALUE protected_event_location(VALUE pointer)
     return rb_funcall3(args[0], id_event_location, 4, args + 1);
 }
 
-static VALUE parse(VALUE self, VALUE handler, VALUE yaml, VALUE path)
+struct parse_args {
+    psych_parser_t * psych_parser;
+    VALUE self;
+    VALUE handler;
+    VALUE yaml;
+    VALUE path;
+};
+
+static VALUE parse_body(VALUE ptr)
 {
-    yaml_parser_t * parser;
+    struct parse_args * pargs = (struct parse_args *)ptr;
+    yaml_parser_t * parser = &pargs->psych_parser->yaml_parser;
+    VALUE self = pargs->self;
+    VALUE handler = pargs->handler;
+    VALUE yaml = pargs->yaml;
+    VALUE path = pargs->path;
     yaml_event_t event;
     int done = 0;
     int state = 0;
     int parser_encoding = YAML_ANY_ENCODING;
     int encoding = rb_utf8_encindex();
     rb_encoding * internal_enc = rb_default_internal_encoding();
 
-    TypedData_Get_Struct(self, yaml_parser_t, &psych_parser_type, parser);
-
     yaml_parser_delete(parser);
     yaml_parser_initialize(parser);
 
@@ -312,6 +332,10 @@ static VALUE parse(VALUE self, VALUE handler, VALUE yaml, VALUE path)
         event_args[3] = end_line;
         event_args[4] = end_column;
         rb_protect(protected_event_location, (VALUE)event_args, &state);
+        if (state) {
+            yaml_event_delete(&event);
+            rb_jump_tag(state);
+        }
 
         switch(event.type) {
             case YAML_STREAM_START_EVENT:
@@ -496,7 +520,11 @@ static VALUE parse(VALUE self, VALUE handler, VALUE yaml, VALUE path)
             rb_protect(protected_end_mapping, handler, &state);
             break;
           case YAML_NO_EVENT:
+            /* Once libyaml has produced the stream end, every later call
+             * succeeds with a zeroed event and YAML_STREAM_END_EVENT can no
+             * longer be reached.  Stop rather than loop forever. */
             rb_protect(protected_empty, handler, &state);
+            done = 1;
             break;
           case YAML_STREAM_END_EVENT:
             rb_protect(protected_end_stream, handler, &state);
@@ -510,6 +538,37 @@ static VALUE parse(VALUE self, VALUE handler, VALUE yaml, VALUE path)
     return self;
 }
 
+static VALUE parse_ensure(VALUE ptr)
+{
+    psych_parser_t * parser = (psych_parser_t *)ptr;
+
+    parser->parsing = 0;
+
+    return Qnil;
+}
+
+static VALUE parse(VALUE self, VALUE handler, VALUE yaml, VALUE path)
+{
+    psych_parser_t * parser;
+    struct parse_args pargs;
+
+    TypedData_Get_Struct(self, psych_parser_t, &psych_parser_type, parser);
+
+    if (parser->parsing) {
+        rb_raise(rb_const_get(mPsych, rb_intern("Exception")),
+                "parser is already parsing, it cannot be reused from a handler callback");
+    }
+    parser->parsing = 1;
+
+    pargs.psych_parser = parser;
+    pargs.self         = self;
+    pargs.handler      = handler;
+    pargs.yaml         = yaml;
+    pargs.path         = path;
+
+    return rb_ensure(parse_body, (VALUE)&pargs, parse_ensure, (VALUE)parser);
+}
+
 /*
  * call-seq:
  *    parser.mark # => #<Psych::Parser::Mark>
@@ -521,13 +580,13 @@ static VALUE mark(VALUE self)
 {
     VALUE mark_klass;
     VALUE args[3];
-    yaml_parser_t * parser;
+    psych_parser_t * parser;
 
-    TypedData_Get_Struct(self, yaml_parser_t, &psych_parser_type, parser);
+    TypedData_Get_Struct(self, psych_parser_t, &psych_parser_type, parser);
     mark_klass = rb_const_get_at(cPsychParser, rb_intern("Mark"));
-    args[0] = SIZET2NUM(parser->mark.index);
-    args[1] = SIZET2NUM(parser->mark.line);
-    args[2] = SIZET2NUM(parser->mark.column);
+    args[0] = SIZET2NUM(parser->yaml_parser.mark.index);
+    args[1] = SIZET2NUM(parser->yaml_parser.mark.line);
+    args[2] = SIZET2NUM(parser->yaml_parser.mark.column);
 
     return rb_class_new_instance(3, args, mark_klass);
 }