Jules - various

changes.txt

@@ -2,6 +2,26 @@ Change Log
 ==========
 
 
+Fixed issues:
+
+* **Fixed** `4670 <https://github.com/pymupdf/PyMuPDF/issues/4670>`_: scrub fails to remove hidden text after clean_contents stopped including line breaks (\u2265 1.24.0)
+* **Fixed** `4943 <https://github.com/pymupdf/PyMuPDF/issues/4943>`_: enh: applying redactions with image cropping for currently unsupported colorspaces
+* **Fixed** `5030 <https://github.com/pymupdf/PyMuPDF/issues/5030>`_: find_tables() with layout enabled can return a zero-cell Table, and Table.bbox then raises "ValueError: min() iterable argument is empty"
+* **Fixed** `5044 <https://github.com/pymupdf/PyMuPDF/issues/5044>`_: Outline (TOC) parsing bug
+* **Fixed** `5049 <https://github.com/pymupdf/PyMuPDF/issues/5049>`_: font subsetting segfaults in 1.28.0, regression from 1.27.2.2 #5049
+* **Fixed** `5042 <https://github.com/pymupdf/PyMuPDF/issues/5042>`_: Page.get_texttrace() leaks None references \u2014 Fatal Python error (none_dealloc) in long-running processes (1.27.2.3 & 1.28.0)
+
+
+Other:
+
+* Output warning when legacy `fitz` module is imported.
+* Cope better with markdown containing illegal utf8 sequences.
+* Fixed building with PYMUPDF_SETUP_MUPDF_VS_UPGRADE.
+* pymupdf.Page.find_tables():
+  * new args `use_layout: bool = True` `union: bool = False` `refine: bool = False`.
+  * Improved speed.
+
+
 **Changes in version 1.28.0** (2026-06-29)
 
 Fixed issues:
@@ -21,6 +41,7 @@ Other:
 * Support Windows builds with free thread python.
 * pymupdf.Document.save() now saves non-PDF documents in PDF format.
 * New method pymupdf.Document.apply_css().
+* Pyodide wheel is available on pypi.org.
 
 
 **Changes in version 1.27.2.3** (2026-04-24)

src/table.py

@@ -2983,7 +2983,8 @@ def find_tables(
                 tp2 = page.get_textpage(flags=TABLE_DETECTOR_FLAGS)
             for rect in my_boxes:
                 cells = make_table_from_bbox(tp2, word_rects, rect)  # pylint: disable=E0606
-                tbf.tables.append(Table(page, cells))
+                if cells:
+                    tbf.tables.append(Table(page, cells))
         if refine:
             # Grid refinement + reconstruction. Runs while the page is still
             # derotated (before the finally block resets rotation) so word and

tests/test_annots.py

@@ -753,3 +753,40 @@ def test_5033():
         annot.update()
         annot.set_rotation(0)
         annot.update()
+
+
+def test_4943():
+    if pymupdf.pymupdf_version_tuple == (1, 28, 0) and pymupdf.mupdf_version_tuple == (1, 29, 0):
+        print(f'test_4943():not running because pymupdf-1.28.0 has ambiguous mupdf version.')
+        return
+    path = os.path.normpath(f'{__file__}/../../tests/resources/test_4943.pdf')
+    with pymupdf.open(path) as document:
+        page = document[0]
+        page.add_redact_annot(page.rect)
+        expect_fail = pymupdf.mupdf_version_tuple <= (1, 28, 0)
+        print(f'{pymupdf.pymupdf_version_tuple=} {pymupdf.mupdf_version_tuple=} {expect_fail=}.')
+        try:
+            page.apply_redactions(images=pymupdf.PDF_REDACT_IMAGE_PIXELS)
+        except pymupdf.mupdf.FzErrorArgument as e:
+            print(f'Exception: {e}')
+            assert expect_fail, f'{pymupdf.mupdf_version_tuple=} {pymupdf.pymupdf_version_tuple=}'
+        else:
+            print(f'No exception.')
+            assert not expect_fail, f'{pymupdf.mupdf_version_tuple=} {pymupdf.pymupdf_version_tuple=}'
+
+
+def test_4936():
+    path = os.path.normpath(f'{__file__}/../../tests/resources/test_4936.pdf')
+    with pymupdf.open(path) as document:
+        page = document[0]
+        rect = pymupdf.Rect(
+                pymupdf.FZ_MIN_INF_RECT,
+                pymupdf.FZ_MIN_INF_RECT,
+                pymupdf.FZ_MAX_INF_RECT,
+                pymupdf.FZ_MAX_INF_RECT,
+                )
+        page.add_redact_annot(rect, fill=False)
+        page.apply_redactions(graphics=2)
+        drawings = page.get_drawings()
+    print(f'{len(drawings)=}')
+    assert len(drawings) == 0

tests/test_tables.py

@@ -302,6 +302,9 @@ def test_find_tables_state_is_call_local_for_threads():
     if platform.python_implementation() == "GraalVM":
         print("test_find_tables_state_is_call_local_for_threads(): not running because slow on GraalVM.")
         return
+    if os.environ.get('PYODIDE_ROOT'):
+        print('test_find_tables_state_is_call_local_for_threads(): not running on Pyodide - threads unsupported.')
+        return
 
     pdf_bytes = _make_find_tables_state_doc()
     expected = _find_tables_use_layout_false_signature(pdf_bytes)