From cbb5155060f597fd1e2fb54defd76bc97e94267b Mon Sep 17 00:00:00 2001
From: Niklas <dev@n1klas.net>
Date: Tue, 2 Oct 2018 18:21:48 +0200
Subject: [PATCH 1/4] Merge TTML subtitle cues with same timecodes while
 converting to SRT

---
 youtube_dl/utils.py | 21 ++++++++++++++++-----
 1 file changed, 16 insertions(+), 5 deletions(-)

diff --git a/youtube_dl/utils.py b/youtube_dl/utils.py
index e84d35d4d..17ea7bc09 100644
--- a/youtube_dl/utils.py
+++ b/youtube_dl/utils.py
@@ -2831,6 +2831,9 @@ def dfxp2srt(dfxp_data):
             continue
         default_style.update(style)
 
+    last_begin_time = None
+    last_end_time = None
+
     for para, index in zip(paras, itertools.count(1)):
         begin_time = parse_dfxp_time_expr(para.attrib.get('begin'))
         end_time = parse_dfxp_time_expr(para.attrib.get('end'))
@@ -2841,12 +2844,20 @@ def dfxp2srt(dfxp_data):
             if not dur:
                 continue
             end_time = begin_time + dur
-        out.append('%d\n%s --> %s\n%s\n\n' % (
-            index,
-            srt_subtitles_timecode(begin_time),
-            srt_subtitles_timecode(end_time),
-            parse_node(para)))
 
+        if begin_time == last_begin_time and end_time == last_end_time:
+            out.append('%s\n' % (parse_node(para)))
+        else:
+            out.append('\n%d\n%s --> %s\n%s\n' % (
+                index,
+                srt_subtitles_timecode(begin_time),
+                srt_subtitles_timecode(end_time),
+                parse_node(para)))
+
+        last_begin_time = begin_time
+        last_end_time = end_time
+
+    out.append('\n')
     return ''.join(out)
 
 

From 405947c32e6975c854e28e69836393b275e88539 Mon Sep 17 00:00:00 2001
From: Niklas <dev@n1klas.net>
Date: Tue, 2 Oct 2018 19:10:33 +0200
Subject: [PATCH 2/4] Reflect merged cues in indexes

---
 youtube_dl/utils.py | 5 +++--
 1 file changed, 3 insertions(+), 2 deletions(-)

diff --git a/youtube_dl/utils.py b/youtube_dl/utils.py
index 17ea7bc09..c9586e9b1 100644
--- a/youtube_dl/utils.py
+++ b/youtube_dl/utils.py
@@ -2833,6 +2833,7 @@ def dfxp2srt(dfxp_data):
 
     last_begin_time = None
     last_end_time = None
+    index_offset = 0
 
     for para, index in zip(paras, itertools.count(1)):
         begin_time = parse_dfxp_time_expr(para.attrib.get('begin'))
@@ -2846,10 +2847,10 @@ def dfxp2srt(dfxp_data):
             end_time = begin_time + dur
 
         if begin_time == last_begin_time and end_time == last_end_time:
+            index_offset += 1
             out.append('%s\n' % (parse_node(para)))
         else:
-            out.append('\n%d\n%s --> %s\n%s\n' % (
-                index,
+                index - index_offset,
                 srt_subtitles_timecode(begin_time),
                 srt_subtitles_timecode(end_time),
                 parse_node(para)))

From 7ecd95aa9164ad47252c6fa134692f3d9e3e8d49 Mon Sep 17 00:00:00 2001
From: Niklas <dev@n1klas.net>
Date: Tue, 2 Oct 2018 19:11:12 +0200
Subject: [PATCH 3/4] Don't add preceding newline when converting TTML to SRT

---
 youtube_dl/utils.py | 3 +++
 1 file changed, 3 insertions(+)

diff --git a/youtube_dl/utils.py b/youtube_dl/utils.py
index c9586e9b1..cec206742 100644
--- a/youtube_dl/utils.py
+++ b/youtube_dl/utils.py
@@ -2850,6 +2850,9 @@ def dfxp2srt(dfxp_data):
             index_offset += 1
             out.append('%s\n' % (parse_node(para)))
         else:
+            if out:
+                out.append('\n')
+            out.append('%d\n%s --> %s\n%s\n' % (
                 index - index_offset,
                 srt_subtitles_timecode(begin_time),
                 srt_subtitles_timecode(end_time),

From 582e1d10e7ee52bd177fca3298792c329e22340f Mon Sep 17 00:00:00 2001
From: Niklas <dev@n1klas.net>
Date: Tue, 2 Oct 2018 20:33:40 +0200
Subject: [PATCH 4/4] Update TTML->SRT conversion test

---
 test/test_utils.py | 7 ++-----
 1 file changed, 2 insertions(+), 5 deletions(-)

diff --git a/test/test_utils.py b/test/test_utils.py
index 9e28e008f..589a1c2bb 100644
--- a/test/test_utils.py
+++ b/test/test_utils.py
@@ -1196,18 +1196,15 @@ The first line
         srt_data = '''1
 00:00:02,080 --> 00:00:05,839
 <font color="white" face="sansSerif" size="16">default style<font color="red">custom style</font></font>
-
-2
-00:00:02,080 --> 00:00:05,839
 <b><font color="cyan" face="sansSerif" size="16"><font color="lime">part 1
 </font>part 2</font></b>
 
-3
+2
 00:00:05,839 --> 00:00:09,560
 <u><font color="lime">line 3
 part 3</font></u>
 
-4
+3
 00:00:09,560 --> 00:00:12,359
 <i><u><font color="yellow"><font color="lime">inner
  </font>style</font></u></i>