Seeing the following two stacktraces while testing #733:
2014-11-19 12:59:45,328 setup_pipeline failed for (66, (1, 1), None, 0, 0, None, 'BGRX', (1, 1), 600, 400, codec_spec(nvenc:nvenc4)) Traceback (most recent call last): File "/usr/lib64/python2.7/site-packages/xpra/server/window_video_source.py", line 1095, in setup_pipeline if self.setup_pipeline_option(width, height, src_format, *option): File "/usr/lib64/python2.7/site-packages/xpra/server/window_video_source.py", line 1156, in setup_pipeline_option ve.init_context(enc_width, enc_height, enc_in_format, dst_formats, encoder_spec.encoding, quality, speed, encoder_scaling, self.encoding_options) File "encoder.pyx", line 1254, in xpra.codecs.nvenc4.encoder.Encoder.init_context (xpra/codecs/nvenc4/encoder.c:8739) NVENCException: opening session - returned 10: This indicates that the API call failed because it was unable to allocate enough memory to perform the requested operation. 2014-11-19 12:59:45,853 error on window 88 Traceback (most recent call last): File "/usr/lib64/python2.7/site-packages/xpra/server/source.py", line 362, in recalculate_delays ws.reconfigure() File "/usr/lib64/python2.7/site-packages/xpra/server/window_source.py", line 746, in reconfigure self.update_encoding_options(force_reload) File "/usr/lib64/python2.7/site-packages/xpra/server/window_video_source.py", line 630, in update_encoding_options self.check_pipeline_score(force_reload) File "/usr/lib64/python2.7/site-packages/xpra/server/window_video_source.py", line 670, in check_pipeline_score if ve is None or csce is None or ve.is_closed() or csce.is_closed(): AttributeError: 'xpra.codecs.csc_swscale.colorspace_converter.Color' object has no attribute 'is_closed'
I believe this was caused by r8108: although the calls to
is_closed were there before, we never actually hit that codepath because we never kept the same csc module whilst changing the encoder. That's why this only occurs now with nvenc enabled as per #733.
Backports for v0.14.x were in r8128, I assume this is fixed so closing. Feel free to re-open if I've missed anything.
this ticket has been moved to: https://github.com/Xpra-org/xpra/issues/736