leptris 1.9.242.0 → 1.9.248.0

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 4de40885119a37c57132fe2a16bb048198afa93b8e4077f8c3b4d428e176c753
-  data.tar.gz: 55353281eb7cd09b749905b35172a40d1be65821758ddffcac687fe77df44e52
+  metadata.gz: 1445126f3a7bced66c8ccb08c13e6d61f0a8ffc95f154bb091aedd07e54ef706
+  data.tar.gz: c0c39cf6600337d568dfae5237f650a5b3aa1ebab16b42a648fff456c9d8f050
 SHA512:
-  metadata.gz: 62f07379ee5b97c1695845e7afd61eb0b1135171eb558feb7f8def8789d86488348c9f14616a67124cfbed8f4608c7f1d0af90793e37d54da755669687777f2e
-  data.tar.gz: e086c44b5d76e6e557c19c37cde928c7337e633f20dde340bf46e9ee3c95dd5c2285ea8c51ee5ca9a04a81b04e2f7b2cc605779268bb1c586240c4e58d6524cb
+  metadata.gz: 35a1e9c1637cdfb8d2556d6ce9e10b5fce62b369f5824ae4b9a941026dee156bc54aa5d8dc9d30ba2fe25e8403d096c45449234ed29b422ce5d12febeca9d7ec
+  data.tar.gz: 36a555006ae5e9bb900eaa2f4955c337e4251dbfdd58be20ba91b4825b79de86a6fd8a8b161608c4f71c3d49422d68f3c97e6720e0baceac47e6de7cd0a8e26d

data/CHANGELOG.md

@@ -5,6 +5,22 @@ All notable changes to Leptris will be documented in this file.
 The format is based on [Keep a Changelog](https://keepachangelog.com/1.0.0/),
 and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0.html).
 
+## [1.9.248.0] - 2026-09-26
+
+### Fixed
+
+- Vendored libleptris 1.9.248 (riding 1.9.247 + 1.9.248):
+  - xpath: filtering a non-nodeset base treats the atom as a
+    one-item sequence — `(1)[1]` keeps the atom and `(1)[1] + 1`
+    is 2 (was an unconditional empty nodeset).
+  - html: the document arena span is sized from the input once
+    (#1218) — borrowed pointers stay in the int32 compact range,
+    removing ~776k overflow-table inserts on a 2.6 MB parse;
+    measured 31.4 -> 50.9 MB/s (+62%) on the table-rows corpus.
+- Restores the naming lockstep: the previous release vendored
+  engine 1.9.246 under the version 1.9.242.0; version.rb now
+  matches the vendored engine semver again.
+
 ## [1.9.241.0] - 2026-09-25
 
 ### Fixed

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 = "1.9.246"
+LIBLEPTRIS_VERSION = "1.9.248"
 # 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 = "1.9.242.0"
+  VERSION = "1.9.248.0"
 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 v1.9.246
+- libleptris v1.9.248
 - utf8proc v2.11.0
 
 Rebuild by hand:

data/vendor-src/libleptris/CHANGELOG.md

@@ -1,5 +1,21 @@
 ## [Unreleased]
 
+## [1.9.248] - 2026-09-26
+
+<!-- Edit this section with the actual release notes. -->
+<!-- See https://keepachangelog.com for format guidance. -->
+
+### Changed
+
+- (describe changes here)
+
+
+## [1.9.247] - 2026-09-26
+
+### Fixed
+
+- filter on a non-nodeset base treats the atom as a one-item sequence (xpath)
+
 ## [1.9.246] - 2026-09-25
 
 ### Added

data/vendor-src/libleptris/CMakeLists.txt

@@ -3,7 +3,7 @@
 cmake_minimum_required(VERSION 3.20)
 
 project(leptris
-    VERSION 1.9.246
+    VERSION 1.9.248
     DESCRIPTION "Fast XML parser and XPath evaluator in pure C"
     LANGUAGES C CXX
 )

data/vendor-src/libleptris/src/leptris/html/html_parse.c

@@ -5453,8 +5453,17 @@ static LeptrisDocument html_parse_shared(
         if (status) *status = LEPTRIS_ERROR_NULL_ARG;
         return NULL;
     }
+    /* #1218: size the arena span from the input up front — the
+     * HTML tree expands ~8x (nodes + attributes + the owned input
+     * copy), and an undersized span spills later allocations into
+     * extension blocks in a different malloc region, pushing every
+     * borrowed pointer out of the int32 compact range (measured
+     * 776k overflow-table inserts on the 2.6 MB table shape). */
+    extern struct leptris_document* leptris_document_create_with_arena_size(
+        size_t arena_size);
     struct leptris_document* doc =
-        (struct leptris_document*)leptris_document_create();
+        (struct leptris_document*)leptris_document_create_with_arena_size(
+            len * 10 + (1u << 18));
     if (!doc) {
         if (status) *status = LEPTRIS_ERROR_MEMORY;
         return NULL;

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

@@ -173,7 +173,17 @@ void leptris_parse_options_init(leptris_parse_options* opts) {
  * pool-allocated (TODO 154) zeroed struct, strict mode inherited
  * from the thread-local setting, standalone unset (-1).
  */
-LEPTRIS_API LeptrisDocument leptris_document_create(void) {
+/* #1218: parse entry points that know their input size reserve the
+ * arena span ONCE at content-derived scale — an undersized span
+ * overflows into tracked extension blocks, which malloc places in a
+ * different region (measured 105 TB away on macOS, ~5.5 GB from the
+ * span). Every borrowed pointer (text runs, element names, attribute
+ * strings) then exceeds the int32 compact range and pays an
+ * overflow-table insert per field plus a hash lookup per decode for
+ * the REST of the parse. Sized spans keep the whole parse in one
+ * contiguous region. */
+struct leptris_document* leptris_document_create_with_arena_size(
+    size_t arena_size) {
     extern LeptrisMemoryPool* leptris_pool_create_arena_backed(
         LeptrisArena*, int);
     extern void* leptris_pool_alloc(LeptrisMemoryPool* pool, size_t size);
@@ -185,7 +195,7 @@ LEPTRIS_API LeptrisDocument leptris_document_create(void) {
      * documents take the same arena-backed pool the parser uses
      * (overflow extends via tracked blocks; the pool's page
      * machinery no longer runs for documents). */
-    LeptrisArena* arena = leptris_arena_create(64 * 1024);
+    LeptrisArena* arena = leptris_arena_create(arena_size);
     if (!arena) return NULL;
     LeptrisMemoryPool* pool = leptris_pool_create_arena_backed(arena, 1);
     if (!pool) {
@@ -210,6 +220,10 @@ LEPTRIS_API LeptrisDocument leptris_document_create(void) {
     return doc;
 }
 
+LEPTRIS_API LeptrisDocument leptris_document_create(void) {
+    return leptris_document_create_with_arena_size(64 * 1024);
+}
+
 LEPTRIS_API LeptrisDocument leptris_document_create_html(void) {
     struct leptris_document* doc = leptris_document_create();
     if (doc) doc->html_mode = 1;

data/vendor-src/libleptris/src/leptris/xpath/evaluator.c

@@ -985,8 +985,48 @@ struct leptris_xpath_result* evaluate_expr(XPathContext* ctx, XPathASTNode* ast)
             if (!base) return NULL;
             if (base->type != XPATH_RESULT_NODESET ||
                 !base->value.nodeset_value) {
+                /* Non-nodeset base: the atom is a one-item sequence
+                 * — (1)[1] keeps it, (1)[true()] keeps it, (1)[2]
+                 * empties (the old code returned empty
+                 * unconditionally). The predicate sees position 1
+                 * of 1; the context node stays the outer one. */
+                int keep = 1;
+                for (size_t p = 1; p < ast->child_count && keep; p++) {
+                    XPathASTNode* pred = ast->children[p];
+                    if (pred->type == XPATH_AST_NUMBER) {
+                        keep = pred->number_value == 1.0;
+                        continue;
+                    }
+                    LeptrisElement old_node = ctx->context_node;
+                    size_t old_pos = ctx->context_position;
+                    size_t old_size = ctx->context_size;
+                    void* old_pred = ctx->current_predicate_node;
+                    ctx->context_position = 1;
+                    ctx->context_size = 1;
+                    struct leptris_xpath_result* pr =
+                        evaluate_expr(ctx, pred);
+                    ctx->context_node = old_node;
+                    ctx->context_position = old_pos;
+                    ctx->context_size = old_size;
+                    ctx->current_predicate_node = old_pred;
+                    if (!pr) {
+                        keep = 0;
+                        continue;
+                    }
+                    if (pr->type == XPATH_RESULT_NUMBER)
+                        keep =
+                            (size_t)pr->value.number_value == 1;
+                    else
+                        keep = xpath_to_boolean(pr) ? 1 : 0;
+                    xpath_result_free(pr);
+                }
+                if (keep) return base;
                 xpath_result_free(base);
-                return xpath_result_new(XPATH_RESULT_NODESET);
+                struct leptris_xpath_result* empty =
+                    xpath_result_new(XPATH_RESULT_NODESET);
+                if (empty && empty->value.nodeset_value)
+                    empty->value.nodeset_value->is_sequence = 1;
+                return empty;
             }
             if (ast->child_count > 1) {
                 apply_predicates(ctx, base->value.nodeset_value,

data/vendor-src/libleptris/test/xquery/test_qt3.cpp

@@ -941,7 +941,7 @@ TEST(Qt3Subset, FnReverse) {
 }
 
 TEST(Qt3Subset, FnSubsequence) {
-    run_test_set("fn/subsequence.xml", {}, 67, {},
+    run_test_set("fn/subsequence.xml", {}, 68, {},
                  {
                    "cbcl-subsequence-011",
                    "cbcl-subsequence-012",
@@ -949,7 +949,6 @@ TEST(Qt3Subset, FnSubsequence) {
                    "cbcl-subsequence-014",
                    "cbcl-subsequence-025",
                    "cbcl-subsequence-019", /* document{} + `!` map chain */
-                   "K-SeqSubsequenceFunc-34", /* mixed-type seq item eq */
                    "fn-subsequence-mix-args-025", /* `!` + @id pair */
 
                  });

data/vendor-src/libleptris/test/xquery/test_xquery.cpp

@@ -893,6 +893,51 @@ TEST(XQueryCtors, SequenceItemEbv) {
 /* unordered{E} / ordered{E} order-mode constructors: order is
  * irrelevant to this engine's sequence model, so the splice strips
  * them to (E) (QT3 K2-SeqReverseFunc-2 nests three deep). */
+/* Filter on a non-nodeset base: the atom is a one-item sequence —
+ * (1)[1] keeps it, (1)[2] empties, boolean predicates see position
+ * 1 of 1 (QT3 K-SeqSubsequenceFunc-34 chains (1)[deep-equal(...)]).
+ * The old code returned an empty nodeset unconditionally. */
+TEST(XPathFilter, ScalarBase) {
+    LeptrisStatus st = LEPTRIS_OK;
+    LeptrisDocument doc = leptris_parse_string("<e/>", 4, &st);
+    ASSERT_NE(doc, nullptr);
+    struct {
+        const char* q;
+        const char* want;
+    } cases[] = {
+        {"(1)[1]", "1"},
+        {"(1)[true()]", "1"},
+        {"(2)[1]", "2"},
+        {"(1)[2]", ""},
+        {"(1)[false()]", ""},
+        {"(1)[deep-equal(1, 1)] eq 1", "true"},
+        {"(1)[1] + 1", "2"},
+    };
+    for (auto& c : cases) {
+        LeptrisXQuery xq = leptris_xquery_parse(c.q, strlen(c.q));
+        ASSERT_NE(xq, nullptr) << c.q;
+        LeptrisXPathResult r = leptris_xquery_eval(xq, doc, NULL);
+        ASSERT_NE(r, nullptr) << c.q;
+        std::string got;
+        if (leptris_xpath_result_type(r) == LEPTRIS_XPATH_NODESET) {
+            size_t n = leptris_xpath_result_count(r);
+            for (size_t i = 0; i < n; i++) {
+                const char* v = leptris_xpath_result_node_value(r, i);
+                if (i) got += ' ';
+                got += v ? v : "";
+            }
+        } else {
+            char* s2 = leptris_xpath_result_string(r);
+            got = s2 ? s2 : "";
+            leptris_free_string(s2);
+        }
+        EXPECT_EQ(got, c.want) << c.q;
+        leptris_xpath_result_free(r);
+        leptris_xquery_free(xq);
+    }
+    leptris_document_free(doc);
+}
+
 TEST(XQueryCtors, OrderModeStrip) {
     LeptrisStatus st = LEPTRIS_OK;
     LeptrisDocument doc = leptris_parse_string("<e/>", 4, &st);

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": "1.9.246",
+  "version": "1.9.248",
   "description": "High-performance XML parser and XPath 1.0 engine in C (libleptris)",
   "homepage": "https://github.com/leptris/leptris",
   "license": "MIT",

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: leptris
 version: !ruby/object:Gem::Version
-  version: 1.9.242.0
+  version: 1.9.248.0
 platform: ruby
 authors:
 - Ribose Inc.
 autorequire:
 bindir: bin
 cert_chain: []
-date: 2026-09-25 00:00:00.000000000 Z
+date: 2026-09-26 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: ffi