test 707560

MuPDF redaction used to leave behind glyph particles with an empty boundary box. This frequently happens in scripting systems like Devanagari.
This script performs a MuPDF regression test in this respect.

tests/test_general.py

@@ -797,6 +797,49 @@ def test_2957_2():
         bbox0 = fitz.Rect(w0[:4]).irect  # its IRect coordinates
         assert bbox0 == bbox1  # must be same coordinates
 
+
+def test_707560():
+    """https://bugs.ghostscript.com/show_bug.cgi?id=707560
+    Ensure that redactions also remove characters with an empty width bbox.
+    """
+    # Make text that will contain characters with an empty bbox.
+
+    greetings = (
+        "Hello, World!",  # english
+        "Hallo, Welt!",  # german
+        "سلام دنیا!",  # persian
+        "வணக்கம், உலகம்!",  # tamil
+        "สวัสดีชาวโลก!",  # thai
+        "Привіт Світ!",  # ucranian
+        "שלום עולם!",  # hebrew
+        "ওহে বিশ্ব!",  # bengali
+        "你好世界！",  # chinese
+        "こんにちは世界！",  # japanese
+        "안녕하세요, 월드!",  # korean
+        "नमस्कार, विश्व !",  # sanskrit
+        "हैलो वर्ल्ड!",  # hindi
+    )
+    text = " ... ".join([g for g in greetings])
+    where = (50, 50, 400, 500)
+    story = fitz.Story(text)
+    bio = io.BytesIO()
+    writer = fitz.DocumentWriter(bio)
+    more = True
+    while more:
+        dev = writer.begin_page(fitz.paper_rect("a4"))
+        more, _ = story.place(where)
+        story.draw(dev)
+        writer.end_page()
+    writer.close()
+    doc = fitz.open("pdf", bio)
+    page = doc[0]
+    text = page.get_text()
+    assert text, "Unexpected: test page has no text."
+    page.add_redact_annot(page.rect)
+    page.apply_redactions()
+    assert not page.get_text(), "Unexpected: text not fully redacted."
+
+
 def test_3070():
     with fitz.open(os.path.abspath(f'{__file__}/../../tests/resources/test_3070.pdf')) as pdf:
       links = pdf[0].get_links()