# ixml parser: skip a comment

[oblomov-dev]

What is wrong

A document that carries a comment does not parse. <!-- falls into the branch
for a start or close tag, where the tag regex does not match it:

ELSEIF lv_xml CP '<*'.
* start or close tag
  FIND FIRST OCCURRENCE OF REGEX lc_regex_tag IN lv_xml RESULTS ls_match.
  ASSERT ls_match-offset = 0.

so the parse ends in ASSERTION_FAILED — an assertion, not an exception, so
no CATCH cx_root around the parse absorbs it:

parse( |<root><!-- c --><item>A</item></root>| ).
" ASSERTION_FAILED

A comment sits in a hand-written or tool-generated document as a matter of
course — a header above the root element, a note next to a setting, an element
commented out — and none of those documents can be read at all today.

The change

The loop takes a comment ahead of the tag branch and skips it as a whole, up
to -->:

ELSEIF lv_xml CP '<!--*'.
  FIND FIRST OCCURRENCE OF '-->' IN lv_xml MATCH OFFSET lv_offset.
  ASSERT sy-subrc = 0.
  lv_offset = lv_offset + 3.
  lv_in_element = abap_false.

A comment carries no content, so it becomes no node — and what stands inside
it is text: the tags of a commented-out element are never read as markup. A
document without a comment is not touched, the new branch is only reached by
the literal <!--.

Tests

• parse_comment — a comment between two elements, the first child of the
  root is still the element
• parse_comment_before_root — a comment between the prolog and the root
• parse_comment_with_markup — <!-- <item>A</item> --> is text, the
  element that follows is the one that is read
• parse_comment_only_child — an element whose whole content is a comment has
  no children

npm test is green: abaplint: 0 issue(s) found, 735 file(s) analyzed, and
the full unit run passes.