From 94219f9068e23c7a454de5e4cd1f3c84ece1059b Mon Sep 17 00:00:00 2001
From: eeshsaxena <chinmaymd72@gmail.com>
Date: Sun, 16 Aug 2026 10:31:25 +0530
Subject: [PATCH] Do not crash on empty WebVTT input or empty SRT cue text

Two subtitle readers raised on trivial input the parsers otherwise
tolerate:

- vtt.to_model on an empty document iterated straight to the None
  terminator in the START state and called None.startswith, raising
  AttributeError.
- srt.to_model hit UnboundLocalError on a cue whose text body is empty
  (a time line immediately followed by a blank line), because
  subtitle_text was only initialized when a non-empty text line was
  seen.

Guard the None line in the WebVTT START state, and initialize
subtitle_text when the paragraph is created.
---
 src/main/python/ttconv/srt/reader.py |  1 +
 src/main/python/ttconv/vtt/reader.py |  2 ++
 src/test/python/test_srt_reader.py   | 13 +++++++++++++
 src/test/python/test_vtt_reader.py   |  5 +++++
 4 files changed, 21 insertions(+)

diff --git a/src/main/python/ttconv/srt/reader.py b/src/main/python/ttconv/srt/reader.py
index 5b1b89eb..ed2e4309 100644
--- a/src/main/python/ttconv/srt/reader.py
+++ b/src/main/python/ttconv/srt/reader.py
@@ -270,6 +270,7 @@ def to_model(data_file: typing.IO, _config: SRTReaderConfiguration = None, progr
         return None
 
       current_p = model.P(doc)
+      subtitle_text = ""
 
       current_p.set_begin(
         int(m.group('begin_h')) * 3600 + 
diff --git a/src/main/python/ttconv/vtt/reader.py b/src/main/python/ttconv/vtt/reader.py
index 7682632e..4ff7985d 100644
--- a/src/main/python/ttconv/vtt/reader.py
+++ b/src/main/python/ttconv/vtt/reader.py
@@ -547,6 +547,8 @@ class _State(Enum):
   for line_index, line in enumerate(_none_terminated(lines)):
 
     if state is _State.START:
+      if line is None:
+        break
       if not line.startswith("WEBVTT"):
         LOGGER.warning("The first line of the file does not start with WEBVTT")
       state = _State.LOOKING
diff --git a/src/test/python/test_srt_reader.py b/src/test/python/test_srt_reader.py
index e1c3f9df..97fb5b9f 100644
--- a/src/test/python/test_srt_reader.py
+++ b/src/test/python/test_srt_reader.py
@@ -546,6 +546,19 @@ def test_alignment_multiple_tags_uses_first(self):
     self.assertNotIn("{\\an", text_content)
     self.assertIn("Multiple tags here", text_content)
 
+  def test_empty_cue_text(self):
+    # A cue whose text body is empty (a time line immediately followed by a
+    # blank line) used to raise UnboundLocalError.
+    f = io.StringIO(
+      "1\n"
+      "00:00:01,000 --> 00:00:04,000\n"
+      "\n"
+      "2\n"
+      "00:00:05,000 --> 00:00:08,000\n"
+      "Second\n"
+    )
+    self.assertIsNotNone(to_model(f))
+
 
 if __name__ == '__main__':
   unittest.main()
diff --git a/src/test/python/test_vtt_reader.py b/src/test/python/test_vtt_reader.py
index 76e57805..769a3900 100644
--- a/src/test/python/test_vtt_reader.py
+++ b/src/test/python/test_vtt_reader.py
@@ -438,5 +438,10 @@ def test_default_positioning(self):
     self.assertEqual(o.x.value, 100*1/40)
     self.assertEqual(e.width.value, 100 - 2*100*1/40)
 
+  def test_empty_input(self):
+    # An empty document used to raise AttributeError on the None sentinel.
+    self.assertIsNotNone(to_model(io.StringIO("")))
+    self.assertIsNotNone(to_model(io.StringIO("   \n")))
+
 if __name__ == '__main__':
   unittest.main()