[FIX] util/snippets: fix conversion of full html docs

upg-1158494
opw-3577439

```
  File "/tmp/tmpjzebte5h/migrations/account_online_synchronization/saas~15.1.1.0/pre-migrate.py", line 6, in migrate
    util.remove_field(cr, "account.link.journal.line", "action")
  File "/tmp/tmpjzebte5h/migrations/util/fields.py", line 163, in remove_field
    adapt_domains(cr, model, fieldname, "ignored", adapter=adapter, skip_inherit=skip_inherit, force_adapt=True)
  File "/tmp/tmpjzebte5h/migrations/util/domains.py", line 316, in adapt_domains
    with suppress(_Skip), edit_view(cr, view_id=view_id, active=None) as view:
  File "/usr/lib/python3.8/contextlib.py", line 113, in __enter__
    return next(self.gen)
  File "/tmp/tmpjzebte5h/migrations/util/records.py", line 210, in edit_view
    arch_etree = lxml.etree.fromstring(arch["en_US"])
  File "src/lxml/etree.pyx", line 3257, in lxml.etree.fromstring
  File "src/lxml/parser.pxi", line 1916, in lxml.etree._parseMemoryDocument
  File "src/lxml/parser.pxi", line 1796, in lxml.etree._parseDoc
  File "src/lxml/parser.pxi", line 1085, in lxml.etree._BaseParser._parseUnicodeDoc
  File "src/lxml/parser.pxi", line 618, in lxml.etree._ParserContext._handleParseResultDoc
  File "src/lxml/parser.pxi", line 728, in lxml.etree._handleParseResult
  File "src/lxml/parser.pxi", line 657, in lxml.etree._raiseParseError
  File "<string>", line 14
lxml.etree.XMLSyntaxError: Extra content at the end of the document, line 14, column 17
```
The Traceback from the lxml etree parser is caused by a previous corruption of
the document in question by a previous upgrade script. The document as is in
the origin DB is a full HTML doc with an `html` root tag and can be parsed by
lxml.etree just fine. The culprit script is
`website/16.0.1.0/pre-convert_html.py`, which ultimately calls the
`HTMLConverter` in `snippets.py`. The HTMLConverter always encloses the
document with tags `<wrap>` and `</wrap>`. This breaks a certain logic in the
`lxml.html.fromstring()` (see
https://github.com/lxml/lxml/blob/2ac88908ffd6df380615c0af35f2134325e4bf30/src/lxml/html/html5parser.py#L184)
and leads to a corrupted result when the converted document is recreated via
`etree.tostring()`: The tags `html`, `head` and `body` are lost.

To fix this, do not add the `wrap` tags if the document looks like a full HTML
doc according to the test and logic of `lxml.html.fromstring()`.

Part of odoo/upgrade#5352

Signed-off-by: Christophe Simonis (chs) <[email protected]>

Author: cawo-odoo

src/util/snippets.py

@@ -199,9 +199,13 @@ def __call__(self, content):
         if not content:
             return (False, content)
         content = re.sub(r"^<\?xml .+\?>\s*", "", content.strip())
-        # Wrap in <wrap> node before parsing to preserve external comments and
-        # multi-root nodes
-        els = html.fromstring(f"<wrap>{content}</wrap>", parser=utf8_parser)
+        # Wrap in <wrap> node before parsing to preserve external comments and multi-root nodes,
+        # except for when this looks like a full html doc, because in this case the wrap tag breaks the logic in
+        # https://github.com/lxml/lxml/blob/2ac88908ffd6df380615c0af35f2134325e4bf30/src/lxml/html/html5parser.py#L184
+        els = html.fromstring(
+            content if content.strip()[:5].lower() == "<html" else f"<wrap>{content}</wrap>",
+            parser=utf8_parser,
+        )
         has_changed = self.has_changed(els)
         new_content = (
             re.sub(r"(^<wrap>|</wrap>$)", "", etree.tostring(els, encoding="unicode").strip())