[pitivi] pipeline: Seek where the user intends after recovering

[Alexandru Băluț <alexbalut src gnome org>]

commit a3243a0ecf34a4a5cb5e888832122aa0b7e6833f
Author: Alexandru Băluț <alexandru balut gmail com>
Date:   Wed Mar 15 23:29:07 2017 +0100

    pipeline: Seek where the user intends after recovering
    
    After recovery, the pipeline used to seek at the last successful seek
    position. The user's intention to seek somewhere else was ignored.
    
    Now, the first attempt is to seek where the user intends to, then if
    that fails at the last successful seek position.
    
    Reviewed-by: Thibault Saunier <tsaunier gnome org>
    Differential Revision: https://phabricator.freedesktop.org/D1692

 pitivi/utils/pipeline.py |   16 +++++++++-------
 1 files changed, 9 insertions(+), 7 deletions(-)
---
diff --git a/pitivi/utils/pipeline.py b/pitivi/utils/pipeline.py
index 1530c6d..ec5163d 100644
--- a/pitivi/utils/pipeline.py
+++ b/pitivi/utils/pipeline.py
@@ -95,6 +95,7 @@ class SimplePipeline(GObject.Object, Loggable):
         self._last_position = 0 * Gst.SECOND
         self._recovery_state = self.RecoveryState.NOT_RECOVERING
         self._attempted_recoveries = 0
+        # The position where the user intends to seek.
         self._next_seek = None
         self._timeout_async_id = 0
         self._force_position_listener = False
@@ -344,11 +345,11 @@ class SimplePipeline(GObject.Object, Loggable):
             self.info("Setting next seek to %s", self._next_seek)
             return
 
-        self.debug("position: %s", format_ns(position))
+        self._next_seek = None
 
         # clamp between [0, duration]
         position = max(0, min(position, self.getDuration()))
-
+        self.debug("Seeking to position: %s", format_ns(position))
         res = self._pipeline.seek(1.0,
                                   Gst.Format.TIME,
                                   Gst.SeekFlags.FLUSH | Gst.SeekFlags.ACCURATE,
@@ -361,7 +362,6 @@ class SimplePipeline(GObject.Object, Loggable):
 
         self._addWaitingForAsyncDoneTimeout()
 
-        self.debug("seeking successful")
         self.emit('position', position)
 
     def seekRelative(self, time_delta):
@@ -400,10 +400,12 @@ class SimplePipeline(GObject.Object, Loggable):
                             self.error("Too many tries to seek back to right position, "
                                        "not trying again, and going back to 0 instead")
                         else:
+                            self.info("Performing seek after pipeline recovery")
                             self._recovery_state = self.RecoveryState.SEEKED_AFTER_RECOVERING
-                            self.simple_seek(self._last_position)
-                            self.info(
-                                "Seeked back to the last position after pipeline recovery")
+                            position = self._last_position
+                            if self._next_seek is not None:
+                                position = self._next_seek
+                            self.simple_seek(position)
                     self._listenToPosition(self._force_position_listener)
                 elif prev == Gst.State.PAUSED and new == Gst.State.PLAYING:
                     self._listenToPosition(True)
@@ -433,8 +435,8 @@ class SimplePipeline(GObject.Object, Loggable):
                 self._attempted_recoveries = 0
             self.__emitPosition()
             if self._next_seek is not None:
+                self.info("Performing seek after ASYNC_DONE")
                 self.simple_seek(self._next_seek)
-                self._next_seek = None
         else:
             self.log("%s [%r]", message.type, message.src)