justhtml 3.6.1__tar.gz → 3.7.0__tar.gz

{justhtml-3.6.1 → justhtml-3.7.0}/CHANGELOG.md

@@ -5,7 +5,16 @@ All notable changes to this project will be documented in this file.
 The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.0.0/),
 and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0.html).
 
-## [Unreleased]
+## [3.7.0] - 2026-07-15
+
+### Fixed
+
+- Keep templates open in the head when they contain a processing instruction, matching the latest web-platform-tests tree-construction fixture.
+
+### Changed
+
+- Refresh the web-platform-tests tree-construction fixtures and comparison scores.
+- Make the correctness benchmark skip an optional `html5-parser` installation when its libxml2 ABI conflicts with `lxml`.
 
 ## [3.6.1] - 2026-07-13
 

{justhtml-3.6.1 → justhtml-3.7.0}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: justhtml
-Version: 3.6.1
+Version: 3.7.0
 Summary: A pure Python HTML5 parser that just works.
 Project-URL: Homepage, https://github.com/emilstenstrom/justhtml
 Project-URL: Issues, https://github.com/emilstenstrom/justhtml/issues
@@ -40,7 +40,7 @@ Requires-Dist: html5-parser; extra == 'benchmark'
 Requires-Dist: html5lib; extra == 'benchmark'
 Requires-Dist: lxml; extra == 'benchmark'
 Requires-Dist: markupever; extra == 'benchmark'
-Requires-Dist: selectolax>=0.4.8; extra == 'benchmark'
+Requires-Dist: selectolax>=0.4.11; extra == 'benchmark'
 Requires-Dist: turbohtml>=1.3.1; extra == 'benchmark'
 Requires-Dist: zstandard; extra == 'benchmark'
 Provides-Extra: dev

{justhtml-3.6.1 → justhtml-3.7.0}/benchmarks/correctness.py

@@ -148,7 +148,7 @@ def check_parser_available(parser_name):
             import html5_parser  # noqa: F401
 
             return True
-        except ImportError:
+        except (ImportError, RuntimeError):
             return False
     if parser_name == "markupever":
         try:

{justhtml-3.6.1 → justhtml-3.7.0}/docs/comparison.md

@@ -11,22 +11,22 @@ Use a different tool when one narrow requirement matters more than the whole pip
 | Tool | HTML5 parsing [1] | Speed | Query | Build | Sanitize | Notes |
 |------|------------------------------------------|-------|----------|-------|------------------|-------|
 | **JustHTML**<br>Pure Python | ✅ 100% | ⚡ Fast | ✅ CSS selectors | ✅ `element()` | ✅ Built-in | Correct, secure, easy to install, and fast enough. |
-| **`selectolax`**<br>Python wrapper of C-based Lexbor | 🟡 95.4% | 🚀 Very Fast | ✅ CSS selectors | ✅ `create_node()` | ❌ Needs sanitization | Very fast. |
-| **Chromium**<br>browser engine | 🟡 95.1% [2] | 🚀 Very Fast | — | — | — | Current browser-harness result. |
-| **`turbohtml`**<br>Python wrapper of a C core | 🟡 94.3% | 🚀 Very Fast | ✅ CSS selectors, XPath | ✅ `E.*` builder | ✅ Built-in | Broad, compiled alternative with parsing, querying, and sanitization. |
-| **WebKit**<br>browser engine | 🟡 94.0% [2] | 🚀 Very Fast | — | — | — | Current browser-harness result. |
-| **Firefox**<br>browser engine | 🟡 93.2% [2] | 🚀 Very Fast | — | — | — | Current browser-harness result. |
+| **`selectolax`**<br>Python wrapper of C-based Lexbor | ✅ 100% | 🚀 Very Fast | ✅ CSS selectors | ✅ `create_node()` | ❌ Needs sanitization | Very fast. |
+| **Chromium**<br>browser engine | 🟡 95.0% [2] | 🚀 Very Fast | — | — | — | Current browser-harness result. |
+| **`turbohtml`**<br>Python wrapper of a C core | 🟡 94.2% | 🚀 Very Fast | ✅ CSS selectors, XPath | ✅ `E.*` builder | ✅ Built-in | Broad, compiled alternative with parsing, querying, and sanitization. |
+| **WebKit**<br>browser engine | 🟡 93.9% [2] | 🚀 Very Fast | — | — | — | Current browser-harness result. |
+| **Firefox**<br>browser engine | 🟡 93.1% [2] | 🚀 Very Fast | — | — | — | Current browser-harness result. |
 | **`html5lib`**<br>Pure Python | 🟡 82.3% | 🐢 Slow | 🟡 XPath (lxml) | 🟡 Tree API | 🔴 [Deprecated](https://github.com/html5lib/html5lib-python/issues/443) | Unmaintained reference implementation; incomplete coverage of the tree-construction fixtures. |
-| **`markupever`**<br>Python wrapper of Rust-based html5ever | 🟡 79.4% | 🚀 Very Fast | ✅ CSS selectors | ✅ `TreeDom .create_*()` | ❌ Needs sanitization | Fast, but 105 fixture cases abort its current parser process. |
+| **`markupever`**<br>Python wrapper of Rust-based html5ever | 🟡 79.4% | 🚀 Very Fast | ✅ CSS selectors | ✅ `TreeDom .create_*()` | ❌ Needs sanitization | Fast, but many fixtures cases abort its current parser process. |
 | **`html5_parser`**<br>Python wrapper of C-based Gumbo | 🔴 47.8% | 🚀 Very Fast | 🟡 XPath (lxml) | 🟡 `etree` (lxml) | ❌ Needs sanitization | Fast, but its public tree API loses information needed by many fixtures. |
 | **`BeautifulSoup`**<br>Pure Python | 🔴 0.3% (default) | 🐢 Slow | 🟡 Custom API | ✅ `new_tag()` API | ❌ Needs sanitization | Wraps `html.parser` (default). Can use lxml or html5lib. |
 | **`html.parser`**<br>Python stdlib | 🔴 0.3% | ⚡ Fast | ❌ None | ❌ None | ❌ Needs sanitization | Standard library. Chokes on malformed HTML. |
-| **`lxml`**<br>Python wrapper of C-based libxml2 | 🔴 0.3% | 🚀 Very Fast | 🟡 XPath | ✅ `etree` / E-factory | ❌ Needs sanitization | Fast but not HTML5 compliant. Context-fragment cases are skipped; supported cases still perform poorly. Don't use the old lxml.html.clean module! |
+| **`lxml`**<br>Python wrapper of C-based libxml2 | 🔴 0.9% | 🚀 Very Fast | 🟡 XPath | ✅ `etree` / E-factory | ❌ Needs sanitization | Fast but not HTML5 compliant. Context-fragment cases are skipped; supported cases still perform poorly. Don't use the old lxml.html.clean module! |
 
-[1]: Correctness scores are counted against the [web-platform-tests tree-construction tests](https://github.com/web-platform-tests/wpt/tree/master/html/syntax/parsing/resources). They were run locally at [2026-07-11](https://github.com/web-platform-tests/wpt/commit/ba458f52c8a848bbd9c525ac98fb1f7b30e41dc2).
+[1]: Library correctness scores are counted against the [web-platform-tests tree-construction tests](https://github.com/web-platform-tests/wpt/tree/master/html/syntax/parsing/resources). They were run locally at [2026-07-15](https://github.com/web-platform-tests/wpt/commit/4830edb033cb486fd0cd6f85b5e937cfc718704d) and cover 1,880 cases after excluding scripting-directed fixtures. JustHTML and Selectolax 0.4.11 passed all 1,880 cases; MarkupEver's score counts its 103 process-aborting cases as errors.
 
 
-[2]: Current local rerun with [`justhtml-html5lib-tests-bench`](https://github.com/EmilStenstrom/justhtml-html5lib-tests-bench) against WPT commit [`ba458f52`](https://github.com/web-platform-tests/wpt/commit/ba458f52c8a848bbd9c525ac98fb1f7b30e41dc2). The browser harness skips 12 `#script-on` cases but includes `#script-off` cases, so these scores are not directly comparable to the 1,879-case Python-parser scores above.
+[2]: Current local rerun with [`justhtml-html5lib-tests-bench`](https://github.com/EmilStenstrom/justhtml-html5lib-tests-bench) against WPT commit [`4830edb`](https://github.com/web-platform-tests/wpt/commit/4830edb033cb486fd0cd6f85b5e937cfc718704d). Chromium 143.0.7499.4, WebKit 26.0, and Firefox 144.0.2 were compared against 1,908 cases; the harness skips 12 `#script-on` cases but includes `#script-off` cases, so these scores are not directly comparable to the 1,880-case Python-parser scores above.
 
 ## Why JustHTML
 

{justhtml-3.6.1 → justhtml-3.7.0}/pyproject.toml

@@ -1,7 +1,7 @@
 [project]
 name = "justhtml"
 authors = [{ name = "Emil Stenström", email = "emil@emilstenstrom.se" }]
-version = "3.6.1"
+version = "3.7.0"
 description = "A pure Python HTML5 parser that just works."
 readme = "README.md"
 license = { file = "LICENSE" }
@@ -18,7 +18,7 @@ benchmark = [
     "html5lib",
     "lxml",
     "beautifulsoup4",
-    "selectolax>=0.4.8",
+    "selectolax>=0.4.11",
     "html5-parser",
     "markupever",
     "turbohtml>=1.3.1",

{justhtml-3.6.1 → justhtml-3.7.0}/src/justhtml/parser/engine.py

@@ -1253,32 +1253,8 @@ class ParseEngine:
         if self._has_null and "\0" in data:
             data = data.replace("\0", "\ufffd")
         node = ProcessingInstruction(data=data)
-        self._move_head_template_to_body_for_processing_instruction()
         self._append_misc_node(node, source_pos)
 
-    def _move_head_template_to_body_for_processing_instruction(self) -> None:
-        if self._fragment or self._head is None or self._html is None:
-            return
-        stack = self._stack
-        template_idx = self._open_template_index()
-        if template_idx is None:
-            return
-        outer_template_idx = template_idx
-        for idx in range(template_idx, 0, -1):
-            candidate = stack[idx]
-            if type(candidate) is Template and candidate.parent is self._head:
-                outer_template_idx = idx
-        template = stack[outer_template_idx]
-        if type(template) is not Template or template.parent is not self._head:
-            return
-        head_children = self._head.children
-        if head_children is not None:  # pragma: no branch - Document head always owns a children list
-            head_children.remove(template)
-        self._append(self._body, template)
-        self._stack = _CountingStack([self._doc, self._html, self._body, *stack[outer_template_idx:]])  # type: ignore[list-item]
-        self._body_mode_seen = True
-        self._after_head = False
-
     def _processing_instruction_data(self, content_start: int, content_end: int) -> str | None:
         html = self._html_input
         if content_start >= content_end or not _is_processing_instruction_name_start(html[content_start]):

{justhtml-3.6.1 → justhtml-3.7.0}/tests/test_parser_engine.py

@@ -14,6 +14,7 @@ from justhtml.parser.engine import (
 )
 from justhtml.parser.options import ParserOptions
 from justhtml.sanitizer import DEFAULT_DOCUMENT_POLICY, DEFAULT_POLICY, SanitizationPolicy, UrlPolicy, UrlRule
+from justhtml.serializer import to_test_format
 from tests.harness.tree import TestRunner
 
 
@@ -48,6 +49,12 @@ class TestParserSyntaxAndRecovery(_ParserEngineTestCase):
         document = JustHTML("<?\ud83d\ude80\ud83dX\ud83d\ude80>", sanitize=False)
         assert document.root.children[0].data == "?🚀\ud83dX🚀"
 
+    def test_processing_instruction_keeps_open_head_template_in_head(self) -> None:
+        document = JustHTML("<template><?pi>", sanitize=False)
+        assert to_test_format(document.root) == (
+            "| <html>\n|   <head>\n|     <template>\n|       content\n|         <?pi ?>\n|   <body>"
+        )
+
     def test_rawtext_processing_instruction_branch(self) -> None:
         script = Element("script", {}, "html")
         engine = ParseEngine("<?x>", fragment=True, plan=compile_raw_engine_plan(fragment=True))