From 1daf57eae54695b0ace2cde4a90a986a660a4095 Mon Sep 17 00:00:00 2001
From: "copilot-swe-agent[bot]" <198982749+Copilot@users.noreply.github.com>
Date: Wed, 4 Mar 2026 09:32:03 +0000
Subject: [PATCH 1/3] Initial plan


From 5efce6518d5a182bfc4dc19ed52d509a5a085600 Mon Sep 17 00:00:00 2001
From: "copilot-swe-agent[bot]" <198982749+Copilot@users.noreply.github.com>
Date: Wed, 4 Mar 2026 09:38:32 +0000
Subject: [PATCH 2/3] Fix SCC captions out of order when short text followed by
 longer text

In PASS 2 of SCCWriter.write(), buffer time calculations could push
a longer caption's adjusted start time before the previous shorter
caption's start time. Two fixes applied:

1. Also adjust the first caption's start time for buffering (was
   skipped due to early `continue`)
2. Clamp each caption's adjusted start time to be at least as late
   as the previous caption's adjusted start time

Co-authored-by: lorandvarga <7048551+lorandvarga@users.noreply.github.com>
---
 pycaption/scc/__init__.py    | 11 ++++++-----
 tests/test_scc_conversion.py | 33 +++++++++++++++++++++++++++++++++
 2 files changed, 39 insertions(+), 5 deletions(-)

diff --git a/pycaption/scc/__init__.py b/pycaption/scc/__init__.py
index 5846f551..ed7424b4 100644
--- a/pycaption/scc/__init__.py
+++ b/pycaption/scc/__init__.py
@@ -563,11 +563,12 @@ def write(self, caption_set):
             code_words = len(code) / 5 + 8
             code_time_microseconds = code_words * MICROSECONDS_PER_CODEWORD
             code_start = start - code_time_microseconds
-            if index == 0:
-                continue
-            previous_code, previous_start, previous_end = codes[index - 1]
-            if previous_end + 3 * MICROSECONDS_PER_CODEWORD >= code_start:
-                codes[index - 1] = (previous_code, previous_start, None)
+            if index > 0:
+                previous_code, previous_start, previous_end = codes[index - 1]
+                if code_start < previous_start:
+                    code_start = previous_start
+                if previous_end + 3 * MICROSECONDS_PER_CODEWORD >= code_start:
+                    codes[index - 1] = (previous_code, previous_start, None)
             codes[index] = (code, code_start, end)
 
         # PASS 3:
diff --git a/tests/test_scc_conversion.py b/tests/test_scc_conversion.py
index ed7a81c9..9ee70517 100644
--- a/tests/test_scc_conversion.py
+++ b/tests/test_scc_conversion.py
@@ -1,3 +1,5 @@
+import re
+
 import pytest
 
 from pycaption import (
@@ -6,6 +8,7 @@
     SCCWriter,
     SRTReader,
     SRTWriter,
+    WebVTTReader,
     WebVTTWriter,
 )
 from tests.mixins import CaptionSetTestingMixIn
@@ -65,6 +68,36 @@ def test_dfxp_is_valid_xml_when_scc_source_has_ampersand_character(
         assert dfxp == sample_dfxp_with_ampersand_character
 
 
+class TestSCCTimestampOrdering:
+    def test_scc_captions_are_in_order_when_short_text_followed_by_long(self):
+        """When short caption text is followed by longer caption text,
+        the SCC output timestamps should remain in chronological order.
+        See: https://github.com/pbs/pycaption/issues/XXX
+        """
+        vtt_input = (
+            "WEBVTT\n\n"
+            "0\n"
+            "00:00:02.200 --> 00:00:02.359\n"
+            "you know,\n\n"
+            "1\n"
+            "00:00:02.400 --> 00:00:03.760\n"
+            "the way he kind of looked at me.\n\n"
+            "2\n"
+            "00:00:04.700 --> 00:00:05.169\n"
+            "And I said,\n\n"
+            "3\n"
+            "00:00:05.210 --> 00:00:05.520\n"
+            "oh\n"
+        )
+        captions = WebVTTReader().read(vtt_input)
+        scc_output = SCCWriter().write(captions)
+        timestamps = re.findall(r"(\d+:\d+:\d+:\d+)", scc_output)
+        for i in range(1, len(timestamps)):
+            assert timestamps[i] >= timestamps[i - 1], (
+                f"Timestamps out of order: {timestamps[i - 1]} > {timestamps[i]}"
+            )
+
+
 class TestSCCToWebVTT:
     def test_webvtt_newlines_are_properly_rendered(
         self,

From 27795144382aff212ed66b05503ce83b3a13f1ed Mon Sep 17 00:00:00 2001
From: "copilot-swe-agent[bot]" <198982749+Copilot@users.noreply.github.com>
Date: Wed, 4 Mar 2026 09:39:38 +0000
Subject: [PATCH 3/3] Address code review: remove placeholder issue reference,
 add format comment

Co-authored-by: lorandvarga <7048551+lorandvarga@users.noreply.github.com>
---
 tests/test_scc_conversion.py | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/tests/test_scc_conversion.py b/tests/test_scc_conversion.py
index 9ee70517..99c3714c 100644
--- a/tests/test_scc_conversion.py
+++ b/tests/test_scc_conversion.py
@@ -72,7 +72,6 @@ class TestSCCTimestampOrdering:
     def test_scc_captions_are_in_order_when_short_text_followed_by_long(self):
         """When short caption text is followed by longer caption text,
         the SCC output timestamps should remain in chronological order.
-        See: https://github.com/pbs/pycaption/issues/XXX
         """
         vtt_input = (
             "WEBVTT\n\n"
@@ -91,6 +90,7 @@ def test_scc_captions_are_in_order_when_short_text_followed_by_long(self):
         )
         captions = WebVTTReader().read(vtt_input)
         scc_output = SCCWriter().write(captions)
+        # SCC timestamps use HH:MM:SS:FF format (FF = frames)
         timestamps = re.findall(r"(\d+:\d+:\d+:\d+)", scc_output)
         for i in range(1, len(timestamps)):
             assert timestamps[i] >= timestamps[i - 1], (