ox 2.14.16 → 2.14.18

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 2ba2060f84fb1cfb231c2dd8be4e8cf54e1533f7381f88ce45ab75ff18bb5ed5
-  data.tar.gz: c3a4ef0949a291c44a1502ce21bdecd6d1a594f6b2260f37dd042e0f87fe676d
+  metadata.gz: fd178b1fd415d3a97ef3b60e73896d52fd3eea1e9ef7a32dd0de316e00a21542
+  data.tar.gz: f9813bab8a071e6ba9efa07a92c228d3ea783a29503cfc763e72e73220899fe8
 SHA512:
-  metadata.gz: c63630cec28f306285cc3126dca0feb2f997da02394519b7a8649bfb05d00604f4b54778dc07434acb202dfa575ca1cfab35f1cb5f9904c4939a7ce694f5d802
-  data.tar.gz: 6f01842a610c01ae966bf897f79232cb17b46e6d3ea4c9b5c58996bd59ccd3f60a60626d4fdc76b4d5c78d1b79ceedf82ab7ea7cd2b3b9fa0802103a40ae9655
+  metadata.gz: 1a7cfc1bfbe0693f52f66f662d4f0f8eeabe20bc50ddefb854abde16e1f1e1c8236b64a2c1ebd51d6de18a697f588ff4d60f854e7b50b383db14ad6eed57ae92
+  data.tar.gz: ae9344f3607e94721eafda29f7255be6a36a0340dd4de3b03d249e616126ec340f5c9485f855ba342f33c951d7153c1b47e0b95aaada4e1936741f2b6755452c

data/CHANGELOG.md

@@ -2,6 +2,18 @@
 
 All changes to the Ox gem are documented here. Releases follow semantic versioning.
 
+## [2.14.18] - 2024-03-21
+
+### Fixed
+
+- UTF8 element names now load correctly thansk to @Uelb.
+
+## [2.14.17] - 2023-07-14
+
+### Fixed
+
+- The sax parser in html mode now allows unquoted attribute values with complaints.
+
 ## [2.14.16] - 2023-04-11
 
 ### Fixed

data/ext/ox/builder.c

@@ -337,21 +337,13 @@ static VALUE builder_new(int argc, VALUE *argv, VALUE self) {
 
         rb_check_type(*argv, T_HASH);
         if (Qnil != (v = rb_hash_lookup(*argv, ox_indent_sym))) {
-#ifdef RUBY_INTEGER_UNIFICATION
             if (rb_cInteger != rb_obj_class(v)) {
-#else
-            if (rb_cFixnum != rb_obj_class(v)) {
-#endif
                 rb_raise(ox_parse_error_class, ":indent must be a fixnum.\n");
             }
             indent = NUM2INT(v);
         }
         if (Qnil != (v = rb_hash_lookup(*argv, ox_size_sym))) {
-#ifdef RUBY_INTEGER_UNIFICATION
             if (rb_cInteger != rb_obj_class(v)) {
-#else
-            if (rb_cFixnum != rb_obj_class(v)) {
-#endif
                 rb_raise(ox_parse_error_class, ":size must be a fixnum.\n");
             }
             buf_size = NUM2LONG(v);
@@ -400,21 +392,13 @@ static VALUE builder_file(int argc, VALUE *argv, VALUE self) {
 
         rb_check_type(argv[1], T_HASH);
         if (Qnil != (v = rb_hash_lookup(argv[1], ox_indent_sym))) {
-#ifdef RUBY_INTEGER_UNIFICATION
             if (rb_cInteger != rb_obj_class(v)) {
-#else
-            if (rb_cFixnum != rb_obj_class(v)) {
-#endif
                 rb_raise(ox_parse_error_class, ":indent must be a fixnum.\n");
             }
             indent = NUM2INT(v);
         }
         if (Qnil != (v = rb_hash_lookup(argv[1], ox_size_sym))) {
-#ifdef RUBY_INTEGER_UNIFICATION
             if (rb_cInteger != rb_obj_class(v)) {
-#else
-            if (rb_cFixnum != rb_obj_class(v)) {
-#endif
                 rb_raise(ox_parse_error_class, ":size must be a fixnum.\n");
             }
             buf_size = NUM2LONG(v);
@@ -461,21 +445,13 @@ static VALUE builder_io(int argc, VALUE *argv, VALUE self) {
 
         rb_check_type(argv[1], T_HASH);
         if (Qnil != (v = rb_hash_lookup(argv[1], ox_indent_sym))) {
-#ifdef RUBY_INTEGER_UNIFICATION
             if (rb_cInteger != rb_obj_class(v)) {
-#else
-            if (rb_cFixnum != rb_obj_class(v)) {
-#endif
                 rb_raise(ox_parse_error_class, ":indent must be a fixnum.\n");
             }
             indent = NUM2INT(v);
         }
         if (Qnil != (v = rb_hash_lookup(argv[1], ox_size_sym))) {
-#ifdef RUBY_INTEGER_UNIFICATION
             if (rb_cInteger != rb_obj_class(v)) {
-#else
-            if (rb_cFixnum != rb_obj_class(v)) {
-#endif
                 rb_raise(ox_parse_error_class, ":size must be a fixnum.\n");
             }
             buf_size = NUM2LONG(v);
@@ -732,9 +708,7 @@ static VALUE builder_text(int argc, VALUE *argv, VALUE self) {
         strip_invalid_chars = Qfalse;
     }
 
-    if (T_STRING != rb_type(v)) {
-        v = rb_funcall(v, ox_to_s_id, 0);
-    }
+    v = rb_String(v);
     i_am_a_child(b, true);
     append_string(b, StringValuePtr(v), RSTRING_LEN(v), xml_element_chars, RTEST(strip_invalid_chars));
 
@@ -754,9 +728,7 @@ static VALUE builder_cdata(VALUE self, VALUE data) {
     const char    *end;
     int            len;
 
-    if (T_STRING != rb_type(v)) {
-        v = rb_funcall(v, ox_to_s_id, 0);
-    }
+    v   = rb_String(v);
     str = StringValuePtr(v);
     len = (int)RSTRING_LEN(v);
     s   = str;
@@ -796,9 +768,7 @@ static VALUE builder_raw(VALUE self, VALUE text) {
     const char    *end;
     int            len;
 
-    if (T_STRING != rb_type(v)) {
-        v = rb_funcall(v, ox_to_s_id, 0);
-    }
+    v   = rb_String(v);
     str = StringValuePtr(v);
     len = (int)RSTRING_LEN(v);
     s   = str;
@@ -857,11 +827,7 @@ static VALUE builder_get_indent(VALUE self) {
  * - +indent+ (Fixnum) indentaion level, negative values excludes terminating newline
  */
 static VALUE builder_set_indent(VALUE self, VALUE indent) {
-#ifdef RUBY_INTEGER_UNIFICATION
     if (rb_cInteger != rb_obj_class(indent)) {
-#else
-    if (rb_cFixnum != rb_obj_class(indent)) {
-#endif
         rb_raise(ox_parse_error_class, "indent must be a fixnum.\n");
     }
 

data/ext/ox/dump.c

@@ -159,12 +159,8 @@ static Type obj_class_code(VALUE obj) {
     case T_OBJECT: return (ox_document_clas == clas || ox_element_clas == clas) ? RawCode : ObjectCode;
     case T_REGEXP: return RegexpCode;
     case T_BIGNUM: return BignumCode;
-#ifdef T_COMPLEX
     case T_COMPLEX: return ComplexCode;
-#endif
-#ifdef T_RATIONAL
     case T_RATIONAL: return RationalCode;
-#endif
     case T_CLASS: return ClassCode;
     default: return 0;
     }
@@ -440,19 +436,13 @@ inline static void dump_num(Out out, VALUE obj) {
 }
 
 static void dump_time_thin(Out out, VALUE obj) {
-    char  buf[64];
-    char *b = buf + sizeof(buf) - 1;
-#if HAVE_RB_TIME_TIMESPEC
+    char            buf[64];
+    char           *b    = buf + sizeof(buf) - 1;
     struct timespec ts   = rb_time_timespec(obj);
     time_t          sec  = ts.tv_sec;
     long            nsec = ts.tv_nsec;
-#else
-    time_t sec  = NUM2LONG(rb_funcall2(obj, ox_tv_sec_id, 0, 0));
-    long   nsec = NUM2LONG(rb_funcall2(obj, ox_tv_nsec_id, 0, 0));
-    // long		nsec = NUM2LONG(rb_funcall2(obj, ox_tv_usec_id, 0, 0)) * 1000;
-#endif
-    char *dot = b - 10;
-    long  size;
+    char           *dot  = b - 10;
+    long            size;
 
     *b-- = '\0';
     for (; dot < b; b--, nsec /= 10) {
@@ -495,18 +485,12 @@ static void dump_date(Out out, VALUE obj) {
 }
 
 static void dump_time_xsd(Out out, VALUE obj) {
-    struct tm *tm;
-#if HAVE_RB_TIME_TIMESPEC
+    struct tm      *tm;
     struct timespec ts   = rb_time_timespec(obj);
     time_t          sec  = ts.tv_sec;
     long            nsec = ts.tv_nsec;
-#else
-    time_t sec  = NUM2LONG(rb_funcall2(obj, ox_tv_sec_id, 0, 0));
-    long   nsec = NUM2LONG(rb_funcall2(obj, ox_tv_nsec_id, 0, 0));
-    // long		nsec = NUM2LONG(rb_funcall2(obj, ox_tv_usec_id, 0, 0)) * 1000;
-#endif
-    int  tzhour, tzmin;
-    char tzsign = '+';
+    int             tzhour, tzmin;
+    char            tzsign = '+';
 
     if (out->end - out->cur <= 33) {
         grow(out, 33);
@@ -777,7 +761,7 @@ static void dump_obj(ID aid, VALUE obj, int depth, Out out) {
                 e.indent = -1;
                 out->w_end(out, &e);
             } else if (0 == strcmp("BigDecimal", classname)) {
-                volatile VALUE rs = rb_funcall(obj, ox_to_s_id, 0);
+                volatile VALUE rs = rb_String(obj);
 
                 e.type = BigDecimalCode;
                 out->w_start(out, &e);
@@ -821,13 +805,8 @@ static void dump_obj(ID aid, VALUE obj, int depth, Out out) {
         } else {
             char num_buf[16];
             int  d2 = depth + 1;
-#ifdef RUBY_INTEGER_UNIFICATION
             long i;
-            long cnt = NUM2LONG(rb_struct_size(obj));
-#else   // UNIFY_FIXNUM_AND_INTEGER
-            int i;
-            int cnt = (int)RSTRUCT_LEN(obj);
-#endif  // UNIFY_FIXNUM_AND_INTEGER
+            long cnt   = NUM2LONG(rb_struct_size(obj));
             e.type     = StructCode;
             e.clas.str = rb_class2name(clas);
             e.clas.len = strlen(e.clas.str);
@@ -866,7 +845,6 @@ static void dump_obj(ID aid, VALUE obj, int depth, Out out) {
             dump_gen_element(obj, depth + 1, out);
             out->w_end(out, &e);
         } else { /* Object */
-#if HAVE_RB_IVAR_FOREACH
             e.type   = (Qtrue == rb_obj_is_kind_of(obj, rb_eException)) ? ExceptionCode : ObjectCode;
             cnt      = (int)rb_ivar_count(obj);
             e.closed = (0 >= cnt);
@@ -879,29 +857,6 @@ static void dump_obj(ID aid, VALUE obj, int depth, Out out) {
                 out->depth = od;
                 out->w_end(out, &e);
             }
-#else
-            volatile VALUE vars = rb_obj_instance_variables(obj);
-            // volatile VALUE	   vars = rb_funcall2(obj, rb_intern("instance_variables"), 0, 0);
-
-            e.type   = (Qtrue == rb_obj_is_kind_of(obj, rb_eException)) ? ExceptionCode : ObjectCode;
-            cnt      = (int)RARRAY_LEN(vars);
-            e.closed = (0 >= cnt);
-            out->w_start(out, &e);
-            if (0 < cnt) {
-                const VALUE *np = RARRAY_PTR(vars);
-                ID           vid;
-                unsigned int od = out->depth;
-                int          i;
-
-                out->depth = depth + 1;
-                for (i = cnt; 0 < i; i--, np++) {
-                    vid = rb_to_id(*np);
-                    dump_var(vid, rb_ivar_get(obj, vid), out);
-                }
-                out->depth = od;
-                out->w_end(out, &e);
-            }
-#endif
         }
         break;
     }
@@ -940,7 +895,6 @@ static void dump_obj(ID aid, VALUE obj, int depth, Out out) {
         out->w_end(out, &e);
         break;
     }
-#ifdef T_COMPLEX
     case T_COMPLEX: e.type = ComplexCode; out->w_start(out, &e);
 #ifdef RCOMPLEX
         dump_obj(0, RCOMPLEX(obj)->real, depth + 1, out);
@@ -951,8 +905,6 @@ static void dump_obj(ID aid, VALUE obj, int depth, Out out) {
 #endif
         out->w_end(out, &e);
         break;
-#endif
-#ifdef T_RATIONAL
     case T_RATIONAL: e.type = RationalCode; out->w_start(out, &e);
 #ifdef RRATIONAL
         dump_obj(0, RRATIONAL(obj)->num, depth + 1, out);
@@ -963,7 +915,6 @@ static void dump_obj(ID aid, VALUE obj, int depth, Out out) {
 #endif
         out->w_end(out, &e);
         break;
-#endif
     case T_CLASS: {
         e.type     = ClassCode;
         e.clas.str = rb_class2name(obj);

data/ext/ox/extconf.rb

@@ -32,14 +32,9 @@ CONFIG['warnflags'].slice!(/ -Wsuggest-attribute=format/)
 CONFIG['warnflags'].slice!(/ -Wdeclaration-after-statement/)
 CONFIG['warnflags'].slice!(/ -Wmissing-noreturn/)
 
-have_func('rb_time_timespec')
-have_func('rb_struct_alloc_noinit')
-have_func('rb_obj_encoding')
-have_func('rb_ivar_foreach')
 have_func('rb_ext_ractor_safe', 'ruby.h')
 have_func('pthread_mutex_init')
 have_func('rb_enc_interned_str')
-have_func('rb_time_nano_new')
 have_func('index')
 
 have_header('ruby/st.h')

data/ext/ox/gen_load.c

@@ -73,9 +73,7 @@ static void create_doc(PInfo pi) {
 
     helper_stack_init(&pi->helpers);
     doc = rb_obj_alloc(ox_document_clas);
-#ifdef RB_GC_GUARD
     RB_GC_GUARD(doc);
-#endif
     nodes = rb_ary_new();
     rb_ivar_set(doc, ox_attributes_id, rb_hash_new());
     rb_ivar_set(doc, ox_nodes_id, nodes);
@@ -104,7 +102,7 @@ static void create_prolog_doc(PInfo pi, const char *target, Attr attrs) {
                 VALUE rstr = rb_str_new2(attrs->name);
 
                 rb_enc_associate(rstr, pi->options->rb_enc);
-                sym = rb_funcall(rstr, ox_to_sym_id, 0);
+                sym = rb_str_intern(rstr);
             } else {
                 sym = ID2SYM(rb_intern(attrs->name));
             }

data/ext/ox/hash_load.c

@@ -93,14 +93,26 @@ static void add_str(PInfo pi, VALUE s) {
 }
 
 static void add_text(PInfo pi, char *text, int closed) {
-    add_str(pi, rb_str_new2(text));
+    VALUE s = rb_str_new2(text);
+    if (0 != pi->options->rb_enc) {
+        rb_enc_associate(s, pi->options->rb_enc);
+    }
+    add_str(pi, s);
 }
 
 static void add_cdata(PInfo pi, const char *text, size_t len) {
-    add_str(pi, rb_str_new(text, len));
+    VALUE s = rb_str_new2(text);
+    if (0 != pi->options->rb_enc) {
+        rb_enc_associate(s, pi->options->rb_enc);
+    }
+    add_str(pi, s);
 }
 
 static void add_element(PInfo pi, const char *ename, Attr attrs, int hasChildren) {
+    VALUE s = rb_str_new2(ename);
+    if (0 != pi->options->rb_enc) {
+        rb_enc_associate(s, pi->options->rb_enc);
+    }
     if (helper_stack_empty(&pi->helpers)) {
         create_top(pi);
     }
@@ -111,12 +123,14 @@ static void add_element(PInfo pi, const char *ename, Attr attrs, int hasChildren
         volatile VALUE a;
 
         for (; 0 != attrs->name; attrs++) {
+            key = rb_str_new2(attrs->name);
+            if (0 != pi->options->rb_enc) {
+                rb_enc_associate(key, pi->options->rb_enc);
+            }
             if (Qnil != pi->options->attr_key_mod) {
-                key = rb_funcall(pi->options->attr_key_mod, ox_call_id, 1, rb_str_new2(attrs->name));
+                key = rb_funcall(pi->options->attr_key_mod, ox_call_id, 1, key);
             } else if (Yes == pi->options->sym_keys) {
-                key = rb_id2sym(rb_intern(attrs->name));
-            } else {
-                key = rb_str_new2(attrs->name);
+                key = rb_id2sym(rb_intern_str(key));
             }
             val = rb_str_new2(attrs->value);
             if (0 != pi->options->rb_enc) {
@@ -127,17 +141,21 @@ static void add_element(PInfo pi, const char *ename, Attr attrs, int hasChildren
         a = rb_ary_new();
         rb_ary_push(a, h);
         mark_value(pi, a);
-        helper_stack_push(&pi->helpers, rb_intern(ename), a, ArrayCode);
+        helper_stack_push(&pi->helpers, rb_intern_str(s), a, ArrayCode);
     } else {
-        helper_stack_push(&pi->helpers, rb_intern(ename), Qnil, NoCode);
+        helper_stack_push(&pi->helpers, rb_intern_str(s), Qnil, NoCode);
     }
 }
 
 static void add_element_no_attrs(PInfo pi, const char *ename, Attr attrs, int hasChildren) {
+    VALUE s = rb_str_new2(ename);
+    if (0 != pi->options->rb_enc) {
+        rb_enc_associate(s, pi->options->rb_enc);
+    }
     if (helper_stack_empty(&pi->helpers)) {
         create_top(pi);
     }
-    helper_stack_push(&pi->helpers, rb_intern(ename), Qnil, NoCode);
+    helper_stack_push(&pi->helpers, rb_intern_str(s), Qnil, NoCode);
 }
 
 static int umark_hash_cb(VALUE key, VALUE value, VALUE x) {
@@ -224,8 +242,22 @@ static void finish(PInfo pi) {
     xfree(pi->marked);
 }
 
+static void set_encoding_from_instruct(PInfo pi, Attr attrs) {
+    for (; 0 != attrs->name; attrs++) {
+        if (0 == strcmp("encoding", attrs->name)) {
+            pi->options->rb_enc = rb_enc_find(attrs->value);
+        }
+    }
+}
+
+static void instruct(PInfo pi, const char *target, Attr attrs, const char *content) {
+    if (0 == strcmp("xml", target)) {
+        set_encoding_from_instruct(pi, attrs);
+    }
+}
+
 struct _parseCallbacks _ox_hash_callbacks = {
-    NULL,
+    instruct,
     NULL,
     NULL,
     NULL,
@@ -238,7 +270,7 @@ struct _parseCallbacks _ox_hash_callbacks = {
 ParseCallbacks ox_hash_callbacks = &_ox_hash_callbacks;
 
 struct _parseCallbacks _ox_hash_cdata_callbacks = {
-    NULL,
+    instruct,
     NULL,
     NULL,
     add_cdata,
@@ -251,7 +283,7 @@ struct _parseCallbacks _ox_hash_cdata_callbacks = {
 ParseCallbacks ox_hash_cdata_callbacks = &_ox_hash_cdata_callbacks;
 
 struct _parseCallbacks _ox_hash_no_attrs_callbacks = {
-    NULL,
+    instruct,
     NULL,
     NULL,
     NULL,
@@ -264,7 +296,7 @@ struct _parseCallbacks _ox_hash_no_attrs_callbacks = {
 ParseCallbacks ox_hash_no_attrs_callbacks = &_ox_hash_no_attrs_callbacks;
 
 struct _parseCallbacks _ox_hash_no_attrs_cdata_callbacks = {
-    NULL,
+    instruct,
     NULL,
     NULL,
     add_cdata,

data/ext/ox/obj_load.c

@@ -106,11 +106,7 @@ inline static VALUE structname2obj(const char *name) {
         }
     }
     ost = rb_const_get(ox_struct_class, rb_intern(s));
-#if HAVE_RB_STRUCT_ALLOC_NOINIT
     return rb_struct_alloc_noinit(ost);
-#else
-    return rb_struct_new(ost);
-#endif
 }
 
 inline static VALUE parse_ulong(const char *s, PInfo pi) {
@@ -643,27 +639,21 @@ static void end_element(PInfo pi, const char *ename) {
                 rb_hash_aset(gh->obj, ph->obj, h->obj);
             } break;
             case ComplexCode:
-#ifdef T_COMPLEX
                 if (Qundef == ph->obj) {
                     ph->obj = h->obj;
                 } else {
                     ph->obj = rb_complex_new(ph->obj, h->obj);
                 }
-#else
-                set_error(&pi->err, "Complex Objects not implemented in Ruby 1.8.7", pi->str, pi->s);
-                return;
-#endif
                 break;
             case RationalCode: {
                 if (Qundef == h->obj || RUBY_T_FIXNUM != rb_type(h->obj)) {
                     set_error(&pi->err, "Invalid object format", pi->str, pi->s);
                     return;
                 }
-#ifdef T_RATIONAL
                 if (Qundef == ph->obj) {
                     ph->obj = h->obj;
                 } else {
-                    if (Qundef == ph->obj || RUBY_T_FIXNUM != rb_type(h->obj)) {
+                    if (Qundef == ph->obj || RUBY_T_FIXNUM != rb_type(ph->obj)) {
                         set_error(&pi->err, "Corrupt parse stack, container is wrong type", pi->str, pi->s);
                         return;
                     }
@@ -673,10 +663,6 @@ static void end_element(PInfo pi, const char *ename) {
                     ph->obj = rb_rational_new(ph->obj, h->obj);
 #endif
                 }
-#else
-                set_error(&pi->err, "Rational Objects not implemented in Ruby 1.8.7", pi->str, pi->s);
-                return;
-#endif
                 break;
             }
             default:
@@ -719,11 +705,7 @@ static VALUE parse_double_time(const char *text, VALUE clas) {
     for (; text - dot <= 9; text++) {
         v2 *= 10;
     }
-#if HAVE_RB_TIME_NANO_NEW
     return rb_time_nano_new(v, v2);
-#else
-    return rb_time_new(v, v2 / 1000);
-#endif
 }
 
 typedef struct _tp {
@@ -774,11 +756,7 @@ static VALUE parse_xsd_time(const char *text, VALUE clas) {
     tm.tm_hour = (int)cargs[3];
     tm.tm_min  = (int)cargs[4];
     tm.tm_sec  = (int)cargs[5];
-#if HAVE_RB_TIME_NANO_NEW
     return rb_time_nano_new(mktime(&tm), cargs[6]);
-#else
-    return rb_time_new(mktime(&tm), cargs[6] / 1000);
-#endif
 }
 
 // debug functions
@@ -816,7 +794,7 @@ static void debug_stack(PInfo pi, const char *comment) {
                 if (HashCode == h->type) {
                     VALUE v;
 
-                    v   = rb_funcall2(h->var, rb_intern("to_s"), 0, 0);
+                    v   = rb_String(h->var);
                     key = StringValuePtr(v);
                 } else if (ObjectCode == (h - 1)->type || ExceptionCode == (h - 1)->type ||
                            RangeCode == (h - 1)->type || StructCode == (h - 1)->type) {

data/ext/ox/ox.c

@@ -73,11 +73,6 @@ ID ox_start_element_id;
 ID ox_string_id;
 ID ox_text_id;
 ID ox_to_c_id;
-ID ox_to_s_id;
-ID ox_to_sym_id;
-ID ox_tv_nsec_id;
-ID ox_tv_sec_id;
-ID ox_tv_usec_id;
 ID ox_value_id;
 
 VALUE ox_encoding_sym;
@@ -687,17 +682,13 @@ static VALUE to_obj(VALUE self, VALUE ruby_xml) {
         xml = ALLOCA_N(char, len);
     }
     memcpy(xml, x, len);
-#ifdef RB_GC_GUARD
     rb_gc_disable();
-#endif
     obj = ox_parse(xml, len - 1, ox_obj_callbacks, 0, &options, &err);
     if (SMALL_XML < len) {
         xfree(xml);
     }
-#ifdef RB_GC_GUARD
     RB_GC_GUARD(obj);
     rb_gc_enable();
-#endif
     if (err_has(&err)) {
         ox_err_raise(&err);
     }
@@ -740,130 +731,123 @@ static VALUE to_gen(VALUE self, VALUE ruby_xml) {
     return obj;
 }
 
-static VALUE load(char *xml, size_t len, int argc, VALUE *argv, VALUE self, VALUE encoding, Err err) {
-    VALUE           obj;
-    struct _options options = ox_default_options;
-
-    if (1 == argc && rb_cHash == rb_obj_class(*argv)) {
-        VALUE h = *argv;
-        VALUE v;
-
-        if (Qnil != (v = rb_hash_lookup(h, mode_sym))) {
-            if (object_sym == v) {
-                options.mode = ObjMode;
-            } else if (optimized_sym == v) {
-                options.mode = ObjMode;
-            } else if (generic_sym == v) {
-                options.mode = GenMode;
-            } else if (limited_sym == v) {
-                options.mode = LimMode;
-            } else if (hash_sym == v) {
-                options.mode = HashMode;
-            } else if (hash_no_attrs_sym == v) {
-                options.mode = HashNoAttrMode;
-            } else {
-                rb_raise(ox_parse_error_class, ":mode must be :generic, :object, :limited, :hash, :hash_no_attrs.\n");
-            }
-        }
-        if (Qnil != (v = rb_hash_lookup(h, effort_sym))) {
-            if (auto_define_sym == v) {
-                options.effort = AutoEffort;
-            } else if (tolerant_sym == v) {
-                options.effort = TolerantEffort;
-            } else if (strict_sym == v) {
-                options.effort = StrictEffort;
-            } else {
-                rb_raise(ox_parse_error_class, ":effort must be :strict, :tolerant, or :auto_define.\n");
-            }
-        }
-        if (Qnil != (v = rb_hash_lookup(h, skip_sym))) {
-            if (skip_none_sym == v) {
-                options.skip = NoSkip;
-            } else if (skip_off_sym == v) {
-                options.skip = OffSkip;
-            } else if (skip_return_sym == v) {
-                options.skip = CrSkip;
-            } else if (skip_white_sym == v) {
-                options.skip = SpcSkip;
-            } else {
-                rb_raise(ox_parse_error_class, ":skip must be :skip_none, :skip_return, :skip_white, or :skip_off.\n");
-            }
-        }
-
-        if (Qnil != (v = rb_hash_lookup(h, trace_sym))) {
-            Check_Type(v, T_FIXNUM);
-            options.trace = FIX2INT(v);
-        }
-        if (Qnil != (v = rb_hash_lookup(h, symbolize_keys_sym))) {
-            options.sym_keys = (Qfalse == v) ? No : Yes;
+static int load_options_cb(VALUE k, VALUE v, VALUE opts) {
+    Options copts = (Options)opts;
+
+    if (mode_sym == k) {
+        if (object_sym == v) {
+            copts->mode = ObjMode;
+        } else if (optimized_sym == v) {
+            copts->mode = ObjMode;
+        } else if (generic_sym == v) {
+            copts->mode = GenMode;
+        } else if (limited_sym == v) {
+            copts->mode = LimMode;
+        } else if (hash_sym == v) {
+            copts->mode = HashMode;
+        } else if (hash_no_attrs_sym == v) {
+            copts->mode = HashNoAttrMode;
+        } else {
+            rb_raise(ox_parse_error_class, ":mode must be :generic, :object, :limited, :hash, :hash_no_attrs.\n");
         }
-        options.element_key_mod = rb_hash_lookup2(h, element_key_mod_sym, options.element_key_mod);
-        options.attr_key_mod    = rb_hash_lookup2(h, attr_key_mod_sym, options.attr_key_mod);
-
-        if (Qnil != (v = rb_hash_lookup(h, convert_special_sym))) {
-            options.convert_special = (Qfalse != v);
+    } else if (effort_sym == k) {
+        if (auto_define_sym == v) {
+            copts->effort = AutoEffort;
+        } else if (tolerant_sym == v) {
+            copts->effort = TolerantEffort;
+        } else if (strict_sym == v) {
+            copts->effort = StrictEffort;
+        } else {
+            rb_raise(ox_parse_error_class, ":effort must be :strict, :tolerant, or :auto_define.\n");
         }
-        if (Qnil != (v = rb_hash_lookup(h, no_empty_sym))) {
-            options.no_empty = (Qfalse != v);
+    } else if (skip_sym == k) {
+        if (skip_none_sym == v) {
+            copts->skip = NoSkip;
+        } else if (skip_off_sym == v) {
+            copts->skip = OffSkip;
+        } else if (skip_return_sym == v) {
+            copts->skip = CrSkip;
+        } else if (skip_white_sym == v) {
+            copts->skip = SpcSkip;
+        } else {
+            rb_raise(ox_parse_error_class, ":skip must be :skip_none, :skip_return, :skip_white, or :skip_off.\n");
         }
-
-        v = rb_hash_lookup(h, invalid_replace_sym);
+    } else if (trace_sym == k) {
+        Check_Type(v, T_FIXNUM);
+        copts->trace = FIX2INT(v);
+    } else if (symbolize_keys_sym == k) {
+        copts->sym_keys = (Qfalse == v) ? No : Yes;
+    } else if (element_key_mod_sym == k) {
+        copts->element_key_mod = v;
+    } else if (attr_key_mod_sym == k) {
+        copts->attr_key_mod = v;
+    } else if (convert_special_sym == k) {
+        copts->convert_special = (Qfalse != v);
+    } else if (no_empty_sym == k) {
+        copts->no_empty = (Qfalse != v);
+    } else if (invalid_replace_sym == k) {
         if (Qnil == v) {
-            if (Qtrue == rb_funcall(h, has_key_id, 1, invalid_replace_sym)) {
-                options.allow_invalid = Yes;
-            }
+            copts->allow_invalid = Yes;
         } else {
             long slen;
 
             Check_Type(v, T_STRING);
             slen = RSTRING_LEN(v);
-            if (sizeof(options.inv_repl) - 2 < (size_t)slen) {
+            if (sizeof(copts->inv_repl) - 2 < (size_t)slen) {
                 rb_raise(ox_parse_error_class,
                          ":invalid_replace can be no longer than %d characters.",
-                         (int)sizeof(options.inv_repl) - 2);
+                         (int)sizeof(copts->inv_repl) - 2);
             }
-            strncpy(options.inv_repl + 1, StringValuePtr(v), sizeof(options.inv_repl) - 1);
-            options.inv_repl[sizeof(options.inv_repl) - 1] = '\0';
-            *options.inv_repl                              = (char)slen;
-            options.allow_invalid                          = No;
+            strncpy(copts->inv_repl + 1, StringValuePtr(v), sizeof(copts->inv_repl) - 1);
+            copts->inv_repl[sizeof(copts->inv_repl) - 1] = '\0';
+            *copts->inv_repl                             = (char)slen;
+            copts->allow_invalid                         = No;
         }
-        v = rb_hash_lookup(h, strip_namespace_sym);
+    } else if (strip_namespace_sym == k) {
         if (Qfalse == v) {
-            *options.strip_ns = '\0';
+            *copts->strip_ns = '\0';
         } else if (Qtrue == v) {
-            *options.strip_ns   = '*';
-            options.strip_ns[1] = '\0';
+            *copts->strip_ns   = '*';
+            copts->strip_ns[1] = '\0';
         } else if (Qnil != v) {
             long slen;
 
             Check_Type(v, T_STRING);
             slen = RSTRING_LEN(v);
-            if (sizeof(options.strip_ns) - 1 < (size_t)slen) {
+            if (sizeof(copts->strip_ns) - 1 < (size_t)slen) {
                 rb_raise(ox_parse_error_class,
                          ":strip_namespace can be no longer than %d characters.",
-                         (int)sizeof(options.strip_ns) - 1);
+                         (int)sizeof(copts->strip_ns) - 1);
             }
-            strncpy(options.strip_ns, StringValuePtr(v), sizeof(options.strip_ns) - 1);
-            options.strip_ns[sizeof(options.strip_ns) - 1] = '\0';
+            strncpy(copts->strip_ns, StringValuePtr(v), sizeof(copts->strip_ns) - 1);
+            copts->strip_ns[sizeof(copts->strip_ns) - 1] = '\0';
         }
-        v = rb_hash_lookup(h, margin_sym);
-        if (Qnil != v) {
-            long slen;
+    } else if (margin_sym == k) {
+        long slen;
 
-            Check_Type(v, T_STRING);
-            slen = RSTRING_LEN(v);
-            if (sizeof(options.margin) - 1 < (size_t)slen) {
-                rb_raise(ox_parse_error_class,
-                         ":margin can be no longer than %d characters.",
-                         (int)sizeof(options.margin) - 1);
-            }
-            strncpy(options.margin, StringValuePtr(v), sizeof(options.margin) - 1);
-            options.margin[sizeof(options.margin) - 1] = '\0';
-            options.margin_len                         = strlen(options.margin);
-        }
-        if (Qnil != (v = rb_hash_lookup(h, with_cdata_sym))) {
-            options.with_cdata = (Qtrue == v);
+        Check_Type(v, T_STRING);
+        slen = RSTRING_LEN(v);
+        if (sizeof(copts->margin) - 1 < (size_t)slen) {
+            rb_raise(ox_parse_error_class,
+                     ":margin can be no longer than %d characters.",
+                     (int)sizeof(copts->margin) - 1);
         }
+        strncpy(copts->margin, StringValuePtr(v), sizeof(copts->margin) - 1);
+        copts->margin[sizeof(copts->margin) - 1] = '\0';
+        copts->margin_len                        = strlen(copts->margin);
+    } else if (with_cdata_sym == k) {
+        copts->with_cdata = (Qtrue == v);
+    }
+
+    return ST_CONTINUE;
+}
+
+static VALUE load(char *xml, size_t len, int argc, VALUE *argv, VALUE self, VALUE encoding, Err err) {
+    VALUE           obj;
+    struct _options options = ox_default_options;
+
+    if (1 == argc && rb_cHash == rb_obj_class(*argv)) {
+        rb_hash_foreach(*argv, load_options_cb, (VALUE)&options);
     }
     if ('\0' == *options.encoding) {
         if (Qnil != encoding) {
@@ -877,14 +861,10 @@ static VALUE load(char *xml, size_t len, int argc, VALUE *argv, VALUE self, VALU
     xml = defuse_bom(xml, &options);
     switch (options.mode) {
     case ObjMode:
-#ifdef RB_GC_GUARD
         rb_gc_disable();
-#endif
         obj = ox_parse(xml, len, ox_obj_callbacks, 0, &options, err);
-#ifdef RB_GC_GUARD
         RB_GC_GUARD(obj);
         rb_gc_enable();
-#endif
         break;
     case GenMode: obj = ox_parse(xml, len, ox_gen_callbacks, 0, &options, err); break;
     case LimMode: obj = ox_parse(xml, len, ox_limited_callbacks, 0, &options, err); break;
@@ -952,11 +932,7 @@ static VALUE load_str(int argc, VALUE *argv, VALUE self) {
     } else {
         xml = ALLOCA_N(char, len);
     }
-#if HAVE_RB_OBJ_ENCODING
     encoding = rb_obj_encoding(*argv);
-#else
-    encoding = Qnil;
-#endif
     memcpy(xml, StringValuePtr(*argv), len);
     xml[len - 1] = '\0';
     obj          = load(xml, len - 1, argc - 1, argv + 1, self, encoding, &err);
@@ -1201,21 +1177,13 @@ static void parse_dump_options(VALUE ropts, Options copts) {
         VALUE v;
 
         if (Qnil != (v = rb_hash_lookup(ropts, ox_indent_sym))) {
-#ifdef RUBY_INTEGER_UNIFICATION
             if (rb_cInteger != rb_obj_class(v) && T_FIXNUM != rb_type(v)) {
-#else
-            if (rb_cFixnum != rb_obj_class(v)) {
-#endif
                 rb_raise(ox_parse_error_class, ":indent must be a Fixnum.\n");
             }
             copts->indent = NUM2INT(v);
         }
         if (Qnil != (v = rb_hash_lookup(ropts, trace_sym))) {
-#ifdef RUBY_INTEGER_UNIFICATION
             if (rb_cInteger != rb_obj_class(v) && T_FIXNUM != rb_type(v)) {
-#else
-            if (rb_cFixnum != rb_obj_class(v)) {
-#endif
                 rb_raise(ox_parse_error_class, ":trace must be a Fixnum.\n");
             }
             copts->trace = NUM2INT(v);
@@ -1469,11 +1437,6 @@ void Init_ox() {
     ox_string_id            = rb_intern("string");
     ox_text_id              = rb_intern("text");
     ox_to_c_id              = rb_intern("to_c");
-    ox_to_s_id              = rb_intern("to_s");
-    ox_to_sym_id            = rb_intern("to_sym");
-    ox_tv_nsec_id           = rb_intern("tv_nsec");
-    ox_tv_sec_id            = rb_intern("tv_sec");
-    ox_tv_usec_id           = rb_intern("tv_usec");
     ox_value_id             = rb_intern("value");
 
     encoding_id = rb_intern("encoding");
@@ -1675,8 +1638,6 @@ _ox_raise_error(const char *msg, const char *xml, const char *current, const cha
             xline++;
         }
     }
-#ifdef RB_GC_GUARD
     rb_gc_enable();
-#endif
     rb_raise(ox_parse_error_class, "%s at line %d, column %d [%s:%d]\n", msg, xline, col, file, line);
 }

data/ext/ox/ox.h

@@ -146,7 +146,7 @@ struct _pInfo {
     VALUE              *marked;
     int                 mark_size;  // allocated size
     int                 mark_cnt;
-    char                last;       // last character read, rarely set
+    char                last;  // last character read, rarely set
 };
 
 extern VALUE ox_parse(char *xml, size_t len, ParseCallbacks pcb, char **endp, Options options, Err err);
@@ -205,11 +205,6 @@ extern ID ox_start_element_id;
 extern ID ox_string_id;
 extern ID ox_text_id;
 extern ID ox_to_c_id;
-extern ID ox_to_s_id;
-extern ID ox_to_sym_id;
-extern ID ox_tv_sec_id;
-extern ID ox_tv_nsec_id;
-extern ID ox_tv_usec_id;
 extern ID ox_value_id;
 
 extern rb_encoding *ox_utf8_encoding;

data/ext/ox/parse.c

@@ -167,7 +167,7 @@ ox_parse(char *xml, size_t len, ParseCallbacks pcb, char **endp, Options options
             helper_stack_cleanup(&pi.helpers);
             return Qnil;
         }
-        pi.s++;    // past <
+        pi.s++;  // past <
         switch (*pi.s) {
         case '?':  // processing instruction
             pi.s++;

data/ext/ox/sax.c

@@ -55,7 +55,7 @@ static char read_text(SaxDrive dr);
 static char read_jump(SaxDrive dr, const char *pat);
 static char read_attrs(SaxDrive dr, char c, char termc, char term2, int is_xml, int eq_req, Hint h);
 static char read_name_token(SaxDrive dr);
-static char read_quoted_value(SaxDrive dr);
+static char read_quoted_value(SaxDrive dr, bool inst);
 
 static void hint_clear_empty(SaxDrive dr);
 static Nv   hint_try_close(SaxDrive dr, const char *name);
@@ -469,7 +469,7 @@ DONE:
         Nv   sp;
 
         for (sp = dr->stack.tail - 1; dr->stack.head <= sp; sp--) {
-            snprintf(msg, sizeof(msg) - 1, "%selement '%s' not closed", EL_MISMATCH, sp->name);
+            snprintf(msg, sizeof(msg) - 1, "%selement '%s' not closed", EL_MISMATCH, nv_name(sp));
             ox_sax_drive_error_at(dr, msg, dr->buf.pos, dr->buf.line, dr->buf.col);
             end_element_cb(dr, sp->val, dr->buf.pos, dr->buf.line, dr->buf.col, sp->hint);
         }
@@ -1219,6 +1219,7 @@ static char read_attrs(SaxDrive dr, char c, char termc, char term2, int is_xml,
             c = buf_next_non_white(&dr->buf);
         }
         if ('=' != c) {
+            // TBD allow in smart mode
             if (eq_req) {
                 dr->err = 1;
                 return c;
@@ -1230,7 +1231,7 @@ static char read_attrs(SaxDrive dr, char c, char termc, char term2, int is_xml,
             pos        = dr->buf.pos + 1;
             line       = dr->buf.line;
             col        = dr->buf.col + 1;
-            c          = read_quoted_value(dr);
+            c          = read_quoted_value(dr, '?' == termc);
             attr_value = dr->buf.str;
 
             if (is_encoding) {
@@ -1297,10 +1298,11 @@ static char read_name_token(SaxDrive dr) {
     return '\0';
 }
 
-/* The character after the quote or if there is no quote, the character after the word is returned. dr->buf.tail is one
- * past that. dr->buf.str will point to the token which will be '\0' terminated.
+/* The character after the quote or if there is no quote, the character after
+ * the word is returned. dr->buf.tail is one past that. dr->buf.str will point
+ * to the token which will be '\0' terminated.
  */
-static char read_quoted_value(SaxDrive dr) {
+static char read_quoted_value(SaxDrive dr, bool inst) {
     char c;
 
     c = buf_get(&dr->buf);
@@ -1324,19 +1326,27 @@ static char read_quoted_value(SaxDrive dr) {
     }
     // not quoted, look for something that terminates the string
     dr->buf.str = dr->buf.tail - 1;
-    ox_sax_drive_error(dr, WRONG_CHAR "attribute value not in quotes");
+    // TBD if smart or html then no error
+    if (!(dr->options.smart && ox_hints_html() != dr->options.hints)) {
+        ox_sax_drive_error(dr, WRONG_CHAR "attribute value not in quotes");
+    }
     while ('\0' != (c = buf_get(&dr->buf))) {
         switch (c) {
         case ' ':
             // case '/':
         case '>':
-        case '?':  // for instructions
         case '\t':
         case '\n':
         case '\r':
             *(dr->buf.tail - 1) = '\0'; /* terminate value */
             // dr->buf.tail is in the correct position, one after the word terminator
             return c;
+        case '?':
+            if (inst) {
+                *(dr->buf.tail - 1) = '\0'; /* terminate value */
+                return c;
+            }
+            break;
         default: break;
         }
     }

data/ext/ox/sax_as.c

@@ -43,11 +43,7 @@ static VALUE parse_double_time(const char *text) {
     for (; text - dot <= 9; text++) {
         v2 *= 10;
     }
-#if HAVE_RB_TIME_NANO_NEW
     return rb_time_nano_new(v, v2);
-#else
-    return rb_time_new(v, v2 / 1000);
-#endif
 }
 
 typedef struct _tp {
@@ -102,11 +98,7 @@ static VALUE parse_xsd_time(const char *text) {
     tm.tm_hour = (int)cargs[3];
     tm.tm_min  = (int)cargs[4];
     tm.tm_sec  = (int)cargs[5];
-#if HAVE_RB_TIME_NANO_NEW
     return rb_time_nano_new(mktime(&tm), cargs[6]);
-#else
-    return rb_time_new(mktime(&tm), cargs[6] / 1000);
-#endif
 }
 
 /* call-seq: as_s()

data/ext/ox/sax_buf.c

@@ -79,7 +79,7 @@ int ox_sax_buf_read(Buf buf) {
         } else {
             shift = buf->pro - buf->head - 1;  // leave one character so we cab backup one
         }
-        if (0 >= shift) {                      /* no space left so allocate more */
+        if (0 >= shift) { /* no space left so allocate more */
             char  *old  = buf->head;
             size_t size = buf->end - buf->head + BUF_PAD;
 

data/ext/ox/slotcache.c

@@ -82,8 +82,8 @@ slot_cache_get(SlotCache cache, const char *key, VALUE **slot, const char **keyp
                     orig->key    = form_key(key);
                     orig->value  = Qundef;
                 }
-            } else {                                                           /* not exact match but on the path */
-                if (0 != cache->key) {                                         /* there is a key/value here already */
+            } else {                   /* not exact match but on the path */
+                if (0 != cache->key) { /* there is a key/value here already */
                     if (depth == *cache->key || (255 <= depth && 0 == strncmp(cache->key, key, depth) &&
                                                  '\0' == cache->key[depth])) { /* key belongs here */
                         continue;
@@ -145,7 +145,7 @@ static void slot_print(SlotCache c, unsigned int depth) {
                     vs   = "undefined";
                     clas = "";
                 } else {
-                    VALUE rs = rb_funcall2((*cp)->value, rb_intern("to_s"), 0, 0);
+                    VALUE rs = rb_String((*cp)->value);
 
                     vs   = StringValuePtr(rs);
                     clas = rb_class2name(rb_obj_class((*cp)->value));

data/lib/ox/version.rb

@@ -1,4 +1,4 @@
 module Ox
   # Current version of the module.
-  VERSION = '2.14.16'
+  VERSION = '2.14.18'
 end

data/lib/ox/xmlrpc_adapter.rb

@@ -12,6 +12,7 @@ module Ox
   class StreamParser < XMLRPC::XMLParser::AbstractStreamParser
     # Create a new instance.
     def initialize
+      super
       @parser_class = OxParser
     end
 

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: ox
 version: !ruby/object:Gem::Version
-  version: 2.14.16
+  version: 2.14.18
 platform: ruby
 authors:
 - Peter Ohler
 autorequire:
 bindir: bin
 cert_chain: []
-date: 2023-04-12 00:00:00.000000000 Z
+date: 2024-03-21 00:00:00.000000000 Z
 dependencies: []
 description: "A fast XML parser and object serializer that uses only standard C lib.\n\nOptimized
   XML (Ox), as the name implies was written to provide speed optimized\nXML handling.
@@ -105,7 +105,7 @@ required_rubygems_version: !ruby/object:Gem::Requirement
     - !ruby/object:Gem::Version
       version: '0'
 requirements: []
-rubygems_version: 3.4.1
+rubygems_version: 3.4.10
 signing_key:
 specification_version: 4
 summary: A fast XML parser and object serializer.