GitHub Issue: html5lib/html5lib-python

Title

<template> element placed in <body> instead of <head> (tree builder bug)

Body

Summary

html5lib-python's tree builder places <template> elements in <body> instead of <head> when encountered before the <body> start tag. The WHATWG spec (section 13.2.6.4.7 "The 'in body' insertion mode") requires <template> start tags to be processed using the "in head" rules, which places them in <head>.

Reproduction

import html5lib

doc = html5lib.parse("<template></template>", treebuilder="dom", namespaceHTMLElements=False)

# Walk the tree to find where <template> ended up
for child in doc.documentElement.childNodes:
    print(f"<{child.localName}>")
    for grandchild in child.childNodes:
        print(f"  <{grandchild.localName}>")

Actual output:

<head>
<body>
  <template>

Expected output (per WHATWG spec):

<head>
  <template>
<body>

Spec reference

WHATWG HTML spec, section 13.2.6.4.7:

A start tag whose tag name is one of: "base", "basefont", "bgsound", "link", "meta", "noframes", "script", "style", "template", "title"

Process the token using the rules for the "in head" insertion mode.

This means <template> in "in body" mode should be handled identically to <script>, <style>, <title>, etc. -- all of which html5lib-python correctly places in <head>.

Additional context

There is also a secondary issue: html5lib-python's treebuilders (minidom, etree) do not implement the template content document fragment. The spec says <template> elements store their children in a separate "content" document fragment, not as regular child nodes. html5lib-python stores template children as regular childNodes instead. This is noted in existing code comments but compounds the placement bug above.

Environment

• html5lib version: 1.1 (also reproduced with 1.2-dev from git)
• Python: 3.13
• Treebuilder: dom (minidom) -- also affects etree
• Compared against: html5ever 0.38 (Servo) -- which places <template> in <head>