Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[bob]

The latest build is up at:

http://files.jriver.com/mediacenter/channels/v21/latest/MediaCenter-21.0.54.deb (also in the latest apt repository)

21.0.54 (3/4/2016)

Changes from the main branch. Should be considered for stable repo but is in latest for now.

21.0.51 (2/26/2016)

1) More work on eliminating hard killed threads. Needs testing for stability.
2) More complete deletion of dynamic servers and zones when those functions are executed.

21.0.48 (2/15/2016)

1. NEW: Software Deinterlacing using the YADIF algorithm during video playback and transcoding.
2. Changed: Many internal changes to remove thread killing on stuck threads. Should eliminate segfaults, might cause hangs. Need feedback.
3. NEW: Added switches under Media Network to disable disk buffering of audio and video files when the content is being received for rendering. This is intended for low power/sdcard based machines.

21.0.39 (1/25/2016)

1. Fixed: Bug in calculating the offset from UTC.
2. Changed: MC linux will download the mp3 encoder instead of using the system mp3 encoder.
3. Fixed: Allocation of some files failed on linux when the filesystem didn't support fallocate(). Affected Thumbnails at the very least.
4. Fixed: ARM System Identification.
5. Changed: deb packaging and post-install to integrate MC more into Unity (and probably other desktops).

See the windows 21 log for general non-linux related changes.
21.0.37 (1/15/2016)

Changes from the main branch

21.0.28 (12/17/2015)

1. Fixed: Potential worker thread hang in CListener class, also a potential memory leak.
2. Fixed: Fixed potential hangs/crashes and memory leaks in networking code.

21.0.23 (11/23/2015)

1. Fixed: Memory leak bug introduced in 20.0.123. More apparent if there are lots of windows (from any application) open at once.
2. New: The internal function that provides the OS information to MC is now functional. This fixed a bug in Noire where the window action buttons could have been windows 10 lookalikes.
3. Changed: Internal event handling.

21.0.16 (10/16/2015)

1. Fixed: Thumbnail creation for Videos would often crash JRWorker.
2. Fixed: The TVInfo Expression did not work.
3. Fixed: Change in eventing fixes maximize button on ARM, possibly other subtle changes.

21.0.7 (9/15/2015)

1. NEW: Support for external sidecar subtitles.
2. NEW: Automatic stream selection based on the preferred languages configured in Video -> Subtitles & Languages.
3. Changed: Improved font metrics handling of descents on letters like "g" or "j".
4. Fixed: Leaving fullscreen mode (ie when playing video fullscreen) returns MC to the state is was in before entering fullscreen mode.
5. Fixed: MC behavior with multiple monitors (xrandr) and fullscreen mode. Still needs a bit more work to fully support multiple monitors.
6: Changed: MC will print an error message (to stderr) when unable to open the X display instead of segfaulting.
7. Fixed: The cursor will be hidden in fullscreen mode.
8. Changed: The custom Data is used from the users MC resource area instead of from MC's app library area.

21.0.6 (9/4/2015)

1. Fixed: Issue with crashes when closing the tree during playback when the cover art is displaying in the action window. This could possibly help with other sizing issues.
2. Changed: Skins are copied into the users MC resource area from MC's app library area upon updating to this build and are used from there going forward so they can be modified without requiring root permissions.

21.0.5 (9/1/2015)

1. Changed: Added graphical confirmation of the license installation through the file association.

21.0.4 (8/14/2015)

1. Fixed: The "View Current Log" button on the logging dialog wasn't working.
2. Fixed: Exporting a playlist to HTML didn't open a browser to view the playlist properly.
3. Fixed: Viewing the html of a gallery in preview mode didn't open a browser properly to view the preview.
4. Changed: Internal, method of running an external command.
5. Fixed: Can purchase an upgrade via the usual method (install MC 20 license on MC 21 and follow the instructions).

21.0.3 (8/10/2015)

1. NEW: Improved font support.
2. Fixed: Trial period was broken.

21.0.2 (8/7/2015)

1. NEW: Subtitle Support during Video Playback.
2. Changed: Windows redraw better.
3. NEW: Image support. 2D viewing effects only at this time.

[bob]

This build could use some testing before putting into the repo. There are a lot of internal changes trying to eliminate crashes that could inadvertently cause hangs.

[Mike Noe]

Very stable here on TW.  Per that thread on segfaults, on .39 , I changed the audio device format from "auto" to a specific format for my S/PDIF device and I haven't had a single segfault since and so far, .48 has been solid, essentially in server mode being accessed by a few MCWS clients.

[Mike Noe]

Another 24 hrs with this build....this time banging on the GUI at bit and playing video, in addition to the MCWS clients (other desktop apps) that control the different zones.  I've also started and stopped it numerous times.  Not a single segfault, no real problems at all, so far.  And yes, I too notice that shutdown is much faster.  Whatever threading changes you made seems to have done wonders -- at least on openSUSE TW.

As a side note, I'd like to test x86_64, but now, everytime I uninstall i386 and install x86_64, my master license requires a restore.  I don't remember this happening before.  I've even tried saving the ~/.jriver/Media Center 21/Settings/ and putting it back, still no joy, it wants a restore.  Is this a path issue to ca_certs stuff?

[bob]

Very stable here on TW.  Per that thread on segfaults, on .39 , I changed the audio device format from "auto" to a specific format for my S/PDIF device and I haven't had a single segfault since and so far, .48 has been solid, essentially in server mode being accessed by a few MCWS clients.

Thanks for the report.
I'm wondering about that auto setting but the code looks clean. It just restarts the open with the next settings if the first one fails.

[bob]

Another 24 hrs with this build....this time banging on the GUI at bit and playing video, in addition to the MCWS clients (other desktop apps) that control the different zones.  I've also started and stopped it numerous times.  Not a single segfault, no real problems at all, so far.  And yes, I too notice that shutdown is much faster.  Whatever threading changes you made seems to have done wonders -- at least on openSUSE TW.

As a side note, I'd like to test x86_64, but now, everytime I uninstall i386 and install x86_64, my master license requires a restore.  I don't remember this happening before.  I've even tried saving the ~/.jriver/Media Center 21/Settings/ and putting it back, still no joy, it wants a restore.  Is this a path issue to ca_certs stuff?

Thanks for the report.

I don't think the license storage should be any different between the two architectures. Does the library go back and forth between the two?

[Mike Noe]

Yeah, library seems intact either way, no problems with the library.

[bob]

I'm hoping we are about done with the threading changes. Would appreciate some testing before it gets into the repository.

[Mike Noe]

Bob,
Couple of segfaults here over the weekend since going to .51.

I have a GUI response tester app that kicks off a worker thread that then hits MCWS with a request as such (essentially a complete database load):

Code: [Select]

Files/Search?query=-[media type]=[]

It gets the results, populates a model, waits 500ms, then does it again, forever.  In addition, on another thread, it queries each zone on the MCWS server(6 zones on this server) for Playing Now Info every 5 seconds (Playback/Info?Zone=xxx).

One segfault occurred about 30 secs after MediaCenter startup when my tester started polling the server for the zone info.  The second segfault occurred hours later, this time when kicking off the tester thread mentioned above, after the 3rd or 4th iteration.  Memory usage appeared normal and CPU usage for MC was less than ~12%.  I should mention that playback is occurring on one zone while running this tester.  Testing library size is ~25K tracks, combo of video and audio files.

No other issues since, fwiw.

[bob]

Bob,
Couple of segfaults here over the weekend since going to .51.

I have a GUI response tester app that kicks off a worker thread that then hits MCWS with a request as such (essentially a complete database load):

Code: [Select]

Files/Search?query=-[media type]=[]

It gets the results, populates a model, waits 500ms, then does it again, forever.  In addition, on another thread, it queries each zone on the MCWS server(6 zones on this server) for Playing Now Info every 5 seconds (Playback/Info?Zone=xxx).

One segfault occurred about 30 secs after MediaCenter startup when my tester started polling the server for the zone info.  The second segfault occurred hours later, this time when kicking off the tester thread mentioned above, after the 3rd or 4th iteration.  Memory usage appeared normal and CPU usage for MC was less than ~12%.  I should mention that playback is occurring on one zone while running this tester.  Testing library size is ~25K tracks, combo of video and audio files.

No other issues since, fwiw.

Could you pass along the script/program so I could try to duplicate that here?

[Mike Noe]

It's all QT-ified C++....QThreads/QNetworkAccessManager and all that.  WIth the binaries, you would need at least QT5.5.1 runtime....would that work?

[bob]

It's all QT-ified C++....QThreads/QNetworkAccessManager and all that.  WIth the binaries, you would need at least QT5.5.1 runtime....would that work?

I'll set it up on one of my test boxes...

[Mike Noe]

Okay, I sent you a more robust CLI version, still requires QT5.5.

Forgot to add, fwiw, only one segfault since my last report.  It's hard to reproduce, but seems related to mcws.

[bob]

Okay, I sent you a more robust CLI version, still requires QT5.5.

Forgot to add, fwiw, only one segfault since my last report.  It's hard to reproduce, but seems related to mcws.

Dealing with server stuff now but will get to it asap. Thanks.

[Mike Noe]

...had one more segfault...maybe this is helpful.

Basically, I rebooted.  On this box, MC is set to start up via the Plasma desktop, not thru the MC option, fwiw.

All is normal, everything starts up fine, my desktop session is restored and then I use my mcws app (similar to the tester I sent you) to connect and start audio playback on the S/PDIF zone (built-in).  As soon as I issue the Playback/Play, MC disappears (segfault).  Started MC again, connected with the app, issue Play again and all is well. 

Rebooted, tried the same again, no segfault, all puppy dogs and rainbows.  This is the same device that I set to S32_LE explicitly from "auto" from before, fwiw. 

As a side note, I've had MPV fail to open this device on first play a few times.  Not until I would play some audio with MC, then go back to MPV, would MPV then open the device successfully.  However, this is rare and I haven't seen it recently.

[bob]

...had one more segfault...maybe this is helpful.

Basically, I rebooted.  On this box, MC is set to start up via the Plasma desktop, not thru the MC option, fwiw.

All is normal, everything starts up fine, my desktop session is restored and then I use my mcws app (similar to the tester I sent you) to connect and start audio playback on the S/PDIF zone (built-in).  As soon as I issue the Playback/Play, MC disappears (segfault).  Started MC again, connected with the app, issue Play again and all is well. 

Rebooted, tried the same again, no segfault, all puppy dogs and rainbows.  This is the same device that I set to S32_LE explicitly from "auto" from before, fwiw. 

As a side note, I've had MPV fail to open this device on first play a few times.  Not until I would play some audio with MC, then go back to MPV, would MPV then open the device successfully.  However, this is rare and I haven't seen it recently.

It really sounds like a driver bug, like it's not getting initialized on properly on the first open.

[Mike Noe]

It really sounds like a driver bug, like it's not getting initialized on properly on the first open.

Yep.  So, this afternoon I just did a boatload of catch-up TumbleWeed updates (incl Kernel 4.4.3) on this box, so I'll keep MC playing for few days like this.

[bob]

To be considered for the stable repo. Currently in latest.