round-5 bisection: updateFields-path / mini-28-distinct / full-1field probes

tests/make_bisect_files.py

@@ -19,6 +19,9 @@ feature family:
   19 no_validate    - +w:doNotValidateAgainstSchema in settings.xml
   20 shared_hf_parts - every section points at section 1's 6 header/footer parts
   21 mini_28fields  - the MINI book with the TOC entry replicated to 28 fields
+  22 updatefields_nodirty - w:dirty off + <w:updateFields w:val="true"/> in settings
+  23 mini_28distinct - mini with 28 fields at 28 DISTINCT bookmarks
+  24 full_1field    - full book with exactly 1 dirty TOC field
 
 Open each in Word; whichever show the error dialog pinpoint the feature
 family at fault.
@@ -61,6 +64,9 @@ NAMES = [
     'bisect_19_no_validate.docx',
     'bisect_20_shared_hf_parts.docx',
     'bisect_21_mini_28fields.docx',
+    'bisect_22_updatefields_nodirty.docx',
+    'bisect_23_mini_28distinct.docx',
+    'bisect_24_full_1field.docx',
 ]
 
 
@@ -358,6 +364,90 @@ def make_mini_28fields(doc):
     return mini
 
 
+def make_updatefields_nodirty(doc):
+    """Full book: remove w:dirty from every field run AND set
+    <w:updateFields w:val="true"/> in settings (Word updates all fields on
+    open through the settings path instead of per-field dirty flags)."""
+    for fc in doc.element.body.iter(qn('w:fldChar')):
+        r = fc.getparent()
+        rpr = r.find(qn('w:rPr'))
+        if rpr is not None:
+            d = rpr.find(qn('w:dirty'))
+            if d is not None:
+                rpr.remove(d)
+    settings = doc.settings.element
+    el = etree.Element(qn('w:updateFields'))
+    el.set(qn('w:val'), 'true')
+    # canonical CT_Settings order: ... alwaysMergeEmptyNamespace, updateFields,
+    # hdrShapeDefaults, footnotePr, endnotePr, compat, rsids, mathPr
+    anchor = None
+    for tag in ('w:hdrShapeDefaults', 'w:footnotePr', 'w:endnotePr',
+                'w:compat', 'w:rsids', 'w:mathPr'):
+        anchor = settings.find(qn(tag))
+        if anchor is not None:
+            break
+    if anchor is not None:
+        anchor.addprevious(el)
+    else:
+        settings.append(el)
+
+
+def make_mini_28_distinct(doc):
+    """Return a copy of the MINI book with 28 dirty PAGEREF fields pointing at
+    28 DISTINCT bookmarks: replicate the first TOC entry 25 more times with
+    rewritten targets, plus 25 copies of the chapter-1 title paragraph each
+    carrying a renumbered, renamed bookmark (_UF_ch_3..27). Un-confounds the
+    same-target multiplicity in bisect_21."""
+    mini = Document(os.path.join(HERE, 'word_test_output.docx'))
+    body = mini.element.body
+    paras = [c for c in body if c.tag == qn('w:p')]
+    first_entry = None
+    for p in paras:
+        if p.find(".//" + qn('w:instrText')) is not None:
+            first_entry = p
+            break
+    title_p = None
+    for p in paras:
+        ppr = p.find(qn('w:pPr'))
+        st = ppr.find(qn('w:pStyle')) if ppr is not None else None
+        if st is not None and st.get(qn('w:val')) == 'WAChapter':
+            title_p = p
+            break
+    if first_entry is None or title_p is None:
+        return mini
+    used = {int(b.get(qn('w:id'))) for b in body.iter(qn('w:bookmarkStart'))}
+    nid = (max(used) + 1) if used else 0
+    for k in range(3, 28):
+        ec = deepcopy(first_entry)
+        ec.find(".//" + qn('w:instrText')).text = f" PAGEREF _UF_ch_{k} "
+        body.insert(list(body).index(first_entry) + 1, ec)
+        tc = deepcopy(title_p)
+        for bs in tc.iter(qn('w:bookmarkStart')):
+            bs.set(qn('w:name'), f"_UF_ch_{k}")
+            bs.set(qn('w:id'), str(nid))
+            nid += 1
+        for be in tc.iter(qn('w:bookmarkEnd')):
+            be.set(qn('w:id'), str(nid))
+            nid += 1
+        body.insert(list(body).index(title_p) + 1, tc)
+    return mini
+
+
+def make_first1_toc(doc):
+    """Full book: keep only the FIRST TOC entry's field; plain-text the rest
+    (count threshold probe: does full + exactly 1 dirty field open?)."""
+    seen = 0
+    for p in doc.element.body.iter(qn('w:p')):
+        if p.find(".//" + qn('w:instrText')) is None:
+            continue
+        if seen < 1:
+            seen += 1
+            continue
+        for r in list(p.findall(qn('w:r'))):
+            if r.find(qn('w:fldChar')) is not None or r.find(qn('w:instrText')) is not None:
+                p.remove(r)
+
+
 MAKERS = [
     make_no_running,
     make_no_oddpage,
@@ -380,6 +470,9 @@ MAKERS = [
     make_no_validate,
     make_shared_hf_parts,
     make_mini_28fields,
+    make_updatefields_nodirty,
+    make_mini_28_distinct,
+    make_first1_toc,
 ]