diff --git a/CHANGELOG.md b/CHANGELOG.md
index 374133d6..4dbe3803 100644
--- a/CHANGELOG.md
+++ b/CHANGELOG.md
@@ -127,6 +127,10 @@
   빈칸이 여러 개 이어진 자리에서 줄이 한 글자 일찍 나뉘었다.
 - FormFit과 쪽 수 추정이 탭 바로 뒤의 닫는 문장 부호(`)`·`.`·`!`·`」` 등)로 줄을 시작할 수 있게 한다. 빈칸 뒤처럼 탭은 앞 줄에
   남고 부호가 다음 줄을 시작한다. 한/글이 그렇게 놓는다. 전에는 탭까지 다음 줄로 보냈다.
+- 쪽 수 추정(실험, `estimate_pages`)이 머리 정보 파트(`Contents/header.xml`)가 없는 문서에서 `IndexError`로 멈추던 것을
+  지원 밖(`a document without a header part (Contents/header.xml)`)으로 알린다. 줄 배치 정보 쓰기도 그런 문서에는 쓰지 않는다.
+- (실험) 줄 배치 정보 쓰기가 문단 모양이 한글·영문/숫자 자동 간격(`hh:autoSpacing`)을 켜고 한글 음절 바로 옆에 영문 글자나
+  숫자가 있는 문단을 건너뛴다. 한/글은 그 사이를 넓히는데 그 폭을 아직 재지 않아, 쓴 줄 나눔이 한/글과 다를 수 있었다.
 - FormFit과 쪽 수 추정이 글꼴 폭 표에 없는 가나(히라가나·가타카나)를 그 글꼴의 한글 음절 폭으로 잡는다(함초롬바탕
   972, 한컴 고딕 932). 한/글이 그렇게 놓는다. 전에는 전각(1 em)으로 셌다.
 - 쪽 수 추정(실험, `estimate_pages`)이 탭 정의를 한 번만 읽는다. 전에는 글자 모양을 볼 때마다 머리 정보의 탭 정의 전체를
diff --git a/src/hwpx/layout/line_cache.py b/src/hwpx/layout/line_cache.py
index 44d0dd71..6eb9dd34 100644
--- a/src/hwpx/layout/line_cache.py
+++ b/src/hwpx/layout/line_cache.py
@@ -295,6 +295,8 @@ def _lines(measure: Any, paragraph: Any, width: int, *, first: bool,
     if headed and (_pages._head_size(measure, paragraph, sizes) or shape.indent):  # a label of its own size
         return None
     text = _pages._run_text(runs)
+    if _auto_spaced(measure, paragraph.get("paraPrIDRef"), text):
+        return None  # the gap Hancom adds between Hangul and Latin or digits was not measured
     style = measure.style(paragraph.get("paraPrIDRef"), refs, paragraph)
     if style is not None and style.head and not (headed and style.head_hangs):
         return None
@@ -383,6 +385,28 @@ def _sure_of(text: str, starts: list[int], style: Any, looks: list[Any] | None =
     return all(start in breaks or text[start - 1] == "\n" for start in starts[1:])
 
 
+def _auto_spaced(measure: Any, para_pr_id: str | None, text: str) -> bool:
+    """Whether the paragraph shape turns on Hancom's automatic spacing between Hangul and Latin letters or
+    digits (``hh:autoSpacing``) and the text puts a Hangul syllable right next to one of those."""
+
+    spacing = getattr(measure, "_line_cache_auto_spacing", None)
+    if spacing is None:
+        spacing = {}
+        for shape in measure._header.iter(f"{HH}paraPr"):
+            auto = next(shape.iter(f"{HH}autoSpacing"), None)
+            if auto is not None:
+                spacing[shape.get("id")] = (auto.get("eAsianEng") in ("1", "true"), auto.get("eAsianNum") in ("1", "true"))
+        measure._line_cache_auto_spacing = spacing
+    latin, digits = spacing.get(para_pr_id, (False, False))
+    if not latin and not digits:
+        return False
+    for before, after in zip(text, text[1:]):
+        for hangul, other in ((before, after), (after, before)):
+            if "가" <= hangul <= "힣" and other.isascii() and (latin and other.isalpha() or digits and other.isdigit()):
+                return True
+    return False
+
+
 def _bold_sure(measure: Any, paragraph: Any, runs: list[Any]) -> bool:
     """Whether every bold run of *paragraph* holding text is set in faces of :data:`_BOLD_SURE_FACES`."""
 
diff --git a/src/hwpx/layout/pages.py b/src/hwpx/layout/pages.py
index 9299b69b..30654bf6 100644
--- a/src/hwpx/layout/pages.py
+++ b/src/hwpx/layout/pages.py
@@ -282,6 +282,8 @@ def _estimate(document: Any, *, first_reason: bool = False) -> PageEstimate:
     one given) instead of laying every section out to name them all."""
 
     root = getattr(_open(document), "_root")
+    if not root.headers:  # the shapes every paragraph names live there
+        return PageEstimate(None, (), ("a document without a header part (Contents/header.xml)",))
     measure = _Measure(root)
     layouts: list[_SectionLayout] = []
     reasons: list[str] = []
diff --git a/tests/test_layout_line_cache.py b/tests/test_layout_line_cache.py
index f2060c37..1b99fe75 100644
--- a/tests/test_layout_line_cache.py
+++ b/tests/test_layout_line_cache.py
@@ -403,3 +403,26 @@ def test_a_tab_among_runs_of_several_faces_gets_its_width() -> None:
     assert len(tabbed) >= 2
     for mine, theirs in tabbed:
         assert [dict(tab.attrib) for tab in mine.iter(f"{HP}tab")] == [dict(tab.attrib) for tab in theirs.iter(f"{HP}tab")]
+
+
+@pytest.mark.parametrize(("eng", "num", "text", "written"), [
+    ("1", "0", "가나 ABC 다라마바사 " * 3, True),     # Latin only beside spaces
+    ("1", "0", "가나ABC다라마바사 " * 3, False),      # a Hangul syllable right next to a Latin letter
+    ("0", "1", "제15호에 따른 정의 " * 3, False),     # ... and next to a digit
+    ("0", "0", "제15호에 따른 정의 " * 3, True),      # automatic spacing off
+])
+def test_a_paragraph_with_automatic_spacing_beside_hangul_is_left_for_hancom(
+        eng: str, num: str, text: str, written: bool) -> None:
+    # Hancom widens the gap between Hangul and Latin letters or digits when the paragraph shape turns that on;
+    # how much was not measured, so such a paragraph keeps no cache.
+    document = HwpxDocument.new()
+    paragraph = document.add_paragraph(text)
+    for shape in document.oxml.headers[0].element.iter(f"{HH}paraPr"):
+        if shape.get("id") == paragraph.element.get("paraPrIDRef"):
+            auto = next(shape.iter(f"{HH}autoSpacing"))
+            auto.set("eAsianEng", eng)
+            auto.set("eAsianNum", num)
+
+    write_line_caches(document)
+
+    assert bool(_segments(paragraph.element)) is written
diff --git a/tests/test_layout_page_estimate.py b/tests/test_layout_page_estimate.py
index 56b5b928..0a3b30d9 100644
--- a/tests/test_layout_page_estimate.py
+++ b/tests/test_layout_page_estimate.py
@@ -2277,3 +2277,32 @@ def test_footnotes_beneath_text_remain_unsupported_with_endnote_support(written:
 
     assert estimate.pages is None
     assert estimate.unsupported == ("section 0: footnotes beneath the text",)
+
+
+def _without_header(data: bytes) -> bytes:
+    out = io.BytesIO()
+    with zipfile.ZipFile(io.BytesIO(data)) as src, zipfile.ZipFile(out, "w", zipfile.ZIP_DEFLATED) as dst:
+        for item in src.infolist():
+            if item.filename == "Contents/header.xml":
+                continue
+            payload = src.read(item.filename)
+            if item.filename == "Contents/content.hpf":
+                payload = re.sub(rb"<opf:item [^>]*href=\"Contents/header.xml\"[^>]*/>", b"", payload)
+            dst.writestr(item, payload)
+    return out.getvalue()
+
+
+def test_a_document_without_a_header_part_is_left_unsupported() -> None:
+    # The paragraph and character shapes every paragraph names live in the header part; without it there is
+    # nothing to lay the text out by, so the estimate says so instead of failing.
+    document = HwpxDocument.new()
+    document.add_paragraph("머리 정보가 없는 문서")
+    data = _without_header(document.to_bytes())
+
+    estimate = estimate_pages(data)
+
+    assert estimate.pages is None and estimate.lines == ()
+    assert estimate.unsupported == ("a document without a header part (Contents/header.xml)",)
+    from hwpx.layout.line_cache import write_line_caches
+    report = write_line_caches(HwpxDocument.open(io.BytesIO(data)))
+    assert report.written == 0 and report.unsupported == estimate.unsupported