leptris 1.9.397.0 → 1.9.399.0

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 6e79b74d747baf367807485fc6bbba8847884f96f9c047aefbe4a8eb3f2b7a5c
-  data.tar.gz: 367a1a42e960bcc98446621a1a6c599ef2b71c7999af7dbe5592bcfa332cab8f
+  metadata.gz: 1e31ed00747161d7d79be2e6f15aabd83a46b50eb8159d976250107e40d9a1e6
+  data.tar.gz: e2586c567b0fa8b5bc5959dc45842ccbea67062f9749dff591964a8ce2483280
 SHA512:
-  metadata.gz: 1e1970af4e1d10542f749cb7d90032d1c4bc0a6cff4e9229ce250431c963829727724279e872c8a5ca153c9ccec098b19dfaf8c5fee57cf9216aa69ef4c3e5f2
-  data.tar.gz: 5d982afa969fa2a3a1a75b2dc70924d94f912cccac18193c5fe1a67631bb008cec07576a02f8e35146579461a10956f754d796f9b7659dcae7ff1ba1b5b10a75
+  metadata.gz: f86ad40fa4c704561b8f3325a6feaa830961a7f013409af5f97dd9c41d3adaf5ee43dd394ad84d45f6088d3105378a8deee34a065b83288d6873b6363a4db691
+  data.tar.gz: 15b3a2b9a5196bdb9e3317a4ea07c7f8fef601f9c0158dea318d02bd3dff413ad26138c4cd8e64243d07add8697431c4a2e3903d04fcbeecd30c7ea16e55c419

data/Rakefile

@@ -8,7 +8,7 @@ RSpec::Core::RakeTask.new(:spec)
 # Pin for `rake compile` and the platform-gem builds. Keep in lockstep
 # with .github/workflows/build.yml (which calls `rake compile`) and the
 # CHANGELOG when libleptris releases.
-LIBLEPTRIS_VERSION = "1.9.397"
+LIBLEPTRIS_VERSION = "1.9.399"
 # Vendored alongside libleptris for fn:normalize-unicode (TODO
 # .restructure/20): built per platform with a RELOCATABLE @rpath
 # install name, loaded by ffi.rb before libleptris so the

data/lib/leptris/version.rb

@@ -1,5 +1,5 @@
 # frozen_string_literal: true
 
 module Leptris
-  VERSION = "1.9.397.0"
+  VERSION = "1.9.399.0"
 end

data/vendor-src/README.md

@@ -1,7 +1,7 @@
 Sources vendored in this gem (recompile rights; the ruby
 variant builds them at install via extconf.rb):
 
-- libleptris v1.9.397
+- libleptris v1.9.399
 - utf8proc v2.11.0
 
 Rebuild by hand:

data/vendor-src/libleptris/CHANGELOG.md

@@ -1,5 +1,43 @@
 ## [Unreleased]
 
+## [1.9.399] - 2026-10-31
+
+### Fixed
+
+- **`\\d` matches the Unicode-3.1 legacy digits (leptris#1936,
+  PR #1940).** XSD 1.0 Part 2 cites Unicode 3.1, where
+  U+1369–U+1371 (Ethiopic digits one-nine) were Nd — Unicode 4.0.1
+  demoted them to No, so the utf8proc-backed Nd mask missed them
+  and the suite's reS38 (`\\d` vs `፱`) / reT51 (`\\D`) rows
+  failed. The legacy-Nd set now joins the class as a pure-codepoint
+  range, outside the utf8proc guard so every build (windows
+  included) answers it; `\\D` inherits the exclusion through the
+  complement's negation. RED-first spec
+  (`XsdDatatypes.DigitClassPinsLegacyNd`), and the shared-fixture
+  collision behind an arm-lane flake got unique names.
+
+
+
+## [1.9.398] - 2026-10-31
+
+### Fixed
+
+- **gMonth accepts both XSD 1.0 spellings (leptris#1936, PR
+  #1937).** The first edition defined the lexical as `--MM--` (the
+  W3C suite's gMonth002/004 carry `--03--`); the errata added
+  `--MM`. Both validate now, each with its optional timezone.
+- **Absent timezone must end the lexical.** `tz_optional`'s absent
+  path returned success without reaching end of string, so the
+  g-regional/date family accepted trailing garbage (`--03garbage`,
+  `2026-10-08x`). Pinned by a 15-row red-first spec; the XSD
+  compile suite stays 138/138.
+- **The automaton size cap is inert at instance validation
+  (moxml#514, PR #1938).** The #1894 cap is a resource guard, not a
+  validity signal: a document whose content model hits the cap
+  validates on its merits instead of failing.
+
+
+
 ## [1.9.397] - 2026-10-31
 
 ### Performance

data/vendor-src/libleptris/CMakeLists.txt

@@ -3,7 +3,7 @@
 cmake_minimum_required(VERSION 3.20)
 
 project(leptris
-    VERSION 1.9.397
+    VERSION 1.9.399
     DESCRIPTION "Fast XML parser and XPath evaluator in pure C"
     LANGUAGES C CXX
 )

data/vendor-src/libleptris/src/leptris/rng/rng_regex.c

@@ -112,6 +112,17 @@ static void cls_add_builtin(RreNode* n, char c) {
             n->cat_set = 1;
             n->cat_mask |= 1u << UTF8PROC_CATEGORY_ND;
 #endif
+            /* XSD 1.0 cites Unicode 3.1, where U+1369-1371
+             * (Ethiopic digits one-nine) were Nd — Unicode 4.0.1
+             * demoted them to No, so modern tables miss them
+             * (#1936 reS38/reT51). Pin the legacy-Nd set outside
+             * the utf8proc guard: the ranges are pure codepoints
+             * and \d must answer them on every build. */
+            if (n->uni_n < 4) {
+                n->uni_lo[n->uni_n] = 0x1369;
+                n->uni_hi[n->uni_n] = 0x1371;
+                n->uni_n++;
+            }
             break;
         case 'w':
             for (int i = '0'; i <= '9'; i++) cls_add(n, (unsigned char)i);

data/vendor-src/libleptris/src/leptris/xsd/datatypes.c

@@ -146,7 +146,9 @@ static int two_digits(const char** p) {
     return 1;
 }
 static int tz_optional(const char** p) {
-    if (**p != 'Z' && **p != '+' && **p != '-') return 1; /* absent */
+    /* #1936: the absent timezone is only absent at END of string —
+     * "2000-01-01garbage" used to pass the g-regional family. */
+    if (**p != 'Z' && **p != '+' && **p != '-') return **p == 0;
     if (**p == 'Z') { (*p)++; return **p == 0; }
     (*p)++;
     return two_digits(p) && *(*p) == ':' && (*p)++ && two_digits(p) &&
@@ -393,7 +395,15 @@ int xsd_builtin_valid(const char* type, const char* v) {
         const char* p = v;
         if (*p++ != '-') return 0;
         if (*p++ != '-') return 0;
-        return two_digits_bounded(&p, 1, 12) && tz_optional(&p);
+        if (!two_digits_bounded(&p, 1, 12)) return 0;
+        /* #1936: XSD 1.0 first edition spelled the lexical --MM--
+         * (the errata added --MM) — both are valid 1.0 forms, each
+         * carrying the optional timezone. */
+        if (p[0] == '-' && p[1] == '-') {
+            const char* q = p + 2;
+            return tz_optional(&q);
+        }
+        return tz_optional(&p);
     }
     if (strcmp(type, "xs:dateTime") == 0) {
         const char* p = v;

data/vendor-src/libleptris/src/leptris/xsd/validate.c

@@ -392,14 +392,13 @@ static void validate_children(struct xsd_validator* v,
     int ok = xsd_content_valid(s, model, xsd_target_ns(s), names, uris,
                                n);
     if (ok == -2) {
-        /* #1894: pathological occurrence product — the automaton
-         * exceeds the compile-time size cap. Loud, not a hang. */
-        const char* l2 = NULL;
-        leptris_element_expanded_name(elem, &l2, NULL, NULL);
-        verrf_atk(v, elem, "content-model", l2 ? l2 : "", "element",
-                  l2 ? l2 : "",
-                  "content model exceeds the automaton size cap "
-                  "(pathological occurrence product)");
+        /* #1894/#514: pathological occurrence product — the automaton
+         * exceeds the compile-time size cap. The cap is a resource
+         * guard, not a validity verdict: the content model is
+         * engine-INERT for this element (no instance finding — the
+         * consuming validator decides), the same posture as the
+         * compile-time guard. The element's content check is skipped
+         * and the walk continues; under-report, never false-report. */
         free(names);
         free(uris);
         return;

data/vendor-src/libleptris/test/xsd/test_xsd_compile.cpp

@@ -390,6 +390,32 @@ TEST(XsdDatatypes, LengthFacetCountsOctetsForHexBinary) {
 // byte world and character-class SUBTRACTION — the OPC
 // ST_ContentType shape. A pattern that fails to parse rejects
 // everything, so media types failed wholesale.
+// leptris#1936 (reS38/reT51): XSD 1.0 cites Unicode 3.1, where
+// U+1369-1371 (Ethiopic digits one-nine) were Nd — \\d matches
+// ፱ and \\D rejects it even though the host tables (4.0.1+)
+// carry them as No.
+TEST(XsdDatatypes, DigitClassPinsLegacyNd) {
+    const char xsd[] =
+        "<xs:schema xmlns:xs=\"http://www.w3.org/2001/XMLSchema\">"
+        "<xs:simpleType name=\"Digs\"><xs:restriction base=\"xs:string\">"
+        "<xs:pattern value=\"\\d\"/></xs:restriction></xs:simpleType>"
+        "<xs:simpleType name=\"NonDigs\"><xs:restriction base=\"xs:string\">"
+        "<xs:pattern value=\"\\D\"/></xs:restriction></xs:simpleType>"
+        "</xs:schema>";
+    LeptrisStatus st = LEPTRIS_OK;
+    LeptrisXsdSchema s = leptris_xsd_compile(xsd, strlen(xsd), &st);
+    ASSERT_EQ(st, LEPTRIS_OK);
+    ASSERT_NE(s, (LeptrisXsdSchema)0);
+    EXPECT_EQ(leptris_xsd_simple_valid(s, "Digs", "\xE1\x8D\xB1"), 1);
+    EXPECT_EQ(leptris_xsd_simple_valid(s, "Digs", "\xE1\x8D\xA9"), 1);
+    EXPECT_EQ(leptris_xsd_simple_valid(s, "Digs", "5"), 1);
+    EXPECT_EQ(leptris_xsd_simple_valid(s, "Digs", "x"), 0);
+    EXPECT_EQ(leptris_xsd_simple_valid(s, "NonDigs", "\xE1\x8D\xB1"), 0);
+    EXPECT_EQ(leptris_xsd_simple_valid(s, "NonDigs", "5"), 0);
+    EXPECT_EQ(leptris_xsd_simple_valid(s, "NonDigs", "x"), 1);
+    leptris_xsd_free(s);
+}
+
 TEST(XsdDatatypes, PatternPropertyClassesAndSubtraction) {
     // The OPC ST_ContentType shape, minus the quoted-string member
     // (same property-class + subtraction machinery): type/subtype
@@ -460,6 +486,35 @@ TEST(XsdDatatypes, BuiltinsLexicalMatrix) {
     }
 }
 
+/* leptris#1936: gMonth accepts BOTH XSD 1.0 spellings — the
+ * errata's --MM and the first edition's --MM-- — each with the
+ * optional timezone; and the g-regional/date family rejects
+ * trailing garbage after an absent timezone. */
+TEST(XsdDatatypes, GMonthSpellingsAndStrictTail) {
+    const BuiltinCase rows[] = {
+        {"xs:gMonth", "--03", 1},
+        {"xs:gMonth", "--03--", 1},
+        {"xs:gMonth", "--03Z", 1},
+        {"xs:gMonth", "--03--Z", 1},
+        {"xs:gMonth", "--03--+05:00", 1},
+        {"xs:gMonth", "--03-05:00", 1},
+        {"xs:gMonth", "--12", 1},
+        {"xs:gMonth", "--13", 0},
+        {"xs:gMonth", "--00", 0},
+        {"xs:gMonth", "--03garbage", 0},
+        {"xs:gMonth", "--03-x", 0},
+        {"xs:gMonth", "--03--05:00", 0},
+        {"xs:date", "2026-10-08garbage", 0},
+        {"xs:gDay", "---05garbage", 0},
+        {"xs:gYear", "2026x", 0},
+    };
+    for (const BuiltinCase& c : rows) {
+        SCOPED_TRACE(std::string(c.type) + " : " + c.value);
+        EXPECT_EQ(leptris_xsd_builtin_valid(c.type, c.value), c.expect)
+            << c.type << " : " + std::string(c.value);
+    }
+}
+
 /* leptris#1883: the builtin lexical table converges on Datatypes
  * Part 2 — collapse semantics, unbounded integers, +INF spelling,
  * 64-bit bounds, calendar/duration/binary grammars, xmlns prefix. */
@@ -3657,6 +3712,33 @@ TEST(XsdContent, PathologicalProductFailsLoudlyNotOom) {
     leptris_xsd_free(s);
 }
 
+/* lutaml/moxml#514: the cap is a resource guard, not a validity
+ * verdict. A content model whose NFA crosses the cap is engine-INERT
+ * at INSTANCE validation — no structured row flips a suite-valid
+ * document to invalid (the compile-time loud face above is the one
+ * place the guard reports). */
+TEST(XsdContent, Issue514CapIsInertAtInstanceValidation) {
+    std::string xsd =
+        "<xs:schema xmlns:xs=\"http://www.w3.org/2001/XMLSchema\">"
+        "<xs:element name=\"r\"><xs:complexType><xs:sequence"
+        " minOccurs=\"0\" maxOccurs=\"2000\"><xs:element name=\"e\""
+        " minOccurs=\"0\" maxOccurs=\"2000\"/></xs:sequence>"
+        "</xs:complexType></xs:element></xs:schema>";
+    LeptrisStatus st = LEPTRIS_OK;
+    LeptrisXsdSchema s =
+        leptris_xsd_compile(xsd.data(), xsd.size(), &st);
+    ASSERT_EQ(st, LEPTRIS_OK);
+    ASSERT_NE(s, (LeptrisXsdSchema)0);
+    LeptrisDocument d = leptris_parse_string("<r><e/><e/></r>", 15, &st);
+    ASSERT_NE(d, nullptr);
+    /* The instance is suite-valid and stays valid: the capped model
+     * is skipped, not failed. */
+    EXPECT_EQ(leptris_xsd_validate(s, d), 1);
+    EXPECT_EQ(leptris_xsd_error_count(s), 0u);
+    leptris_document_free(d);
+    leptris_xsd_free(s);
+}
+
 TEST(XsdContent, Issue1894ReproStaysUnderCapAndCorrect) {
     /* The issue's exact shape — seq maxOccurs=100 x element
      * unbounded — lowers to a loop region (not copies) and must stay
@@ -3760,7 +3842,7 @@ TEST(XsdCrossNs, SingleDocImportChaseResolves) {
         " xmlns:r=\"urn:r\" xmlns=\"urn:a\" targetNamespace=\"urn:a\""
         " elementFormDefault=\"qualified\">"
         "<xs:import namespace=\"urn:r\""
-        " schemaLocation=\"blip_repro_rel.xsd\"/>"
+        " schemaLocation=\"blip_chase_rel.xsd\"/>"
         "<xs:attributeGroup name=\"AG_Blob\">"
         "<xs:attribute ref=\"r:embed\" use=\"optional\"/>"
         "</xs:attributeGroup>"
@@ -3776,8 +3858,8 @@ TEST(XsdCrossNs, SingleDocImportChaseResolves) {
     if (!dir) dir = std::getenv("TEMP"); /* windows runners define TEMP/TMP */
     if (!dir) dir = std::getenv("TMP");
     std::string base = dir ? dir : ".";
-    std::string rel_path = base + "/blip_repro_rel.xsd";
-    std::string main_path = base + "/blip_repro_main.xsd";
+    std::string rel_path = base + "/blip_chase_rel.xsd";
+    std::string main_path = base + "/blip_chase_main.xsd";
     {
         std::ofstream f(rel_path);
         f << rel;

data/vendor-src/libleptris/vcpkg.json

@@ -1,7 +1,7 @@
 {
   "$schema": "https://raw.githubusercontent.com/microsoft/vcpkg/master/scripts/vcpkg.schema.json",
   "name": "leptris",
-  "version": "1.9.397",
+  "version": "1.9.399",
   "description": "High-performance XML parser and XPath 1.0 engine in C (libleptris)",
   "homepage": "https://github.com/leptris/leptris",
   "license": "MIT",

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: leptris
 version: !ruby/object:Gem::Version
-  version: 1.9.397.0
+  version: 1.9.399.0
 platform: ruby
 authors:
 - Ribose Inc.