Scythe42's fixes for 1.4.0 (4 Viewers)

[Scythe42]

released soon ?

https://github.com/MediaPortal/MediaPortal-1/commit/8ecff23599bfa5e39d2c663096d5e345a474c3cd

Do you think this related only for AMD Gfx Card? Or is Nvidia and Intel also affected?

Should be a general Windows thing. Too lazy to check the DDK.

 

[azzuro]

YEAHHH Baby LOL

 

[azzuro]

- System Specs : HTPC
- How many displays : 1
- What resolution does each display use : 1024x768
- More than one GPU used : NO
- How are displays connected : Svideo
- Skin Used : Default

this test are without you last commit !
Tested With Patched Core for Atmolight
Minidisplay Set to Extensive logging

***** Core used : Patched for Atmolight *****
1* Log Files : All plugins which I use are enabled (Atmowin = OK / Atmolight = OK) with LED ON
Minidisplay turn OFF = OK
MediaPortalLogs_20_04_13__16_23.zip

***** Core used : Patched for Atmolight *****
2* Log Files : All plugins which I use are enabled (Atmowin = OK / Atmolight = OK) with LED OFF
Minidisplay turn OFF = OK
MediaPortalLogs_20_04_13__16_33.zip

***** Core used : Official ****
3* Log Files : All plugins which I use are enabled (Atmowin = OK / Atmolight = OK) with LED OFF

Minidisplay turn OFF = WRONG
MediaPortalLogs_20_04_13__16_37.zip

***** Core used : Patched for Atmolight **** (similar to 2* Log, only for confirmation)
4* Log Files : All plugins which I use are enabled (Atmowin = OK / Atmolight = OK) with LED OFF
Minidisplay turn OFF = OK
MediaPortalLogs_20_04_13__16_41.zip

Ps : i can't compil your last commit, can you send me Core file, Please?

 

[Scythe42]

That is an outstanding bug from stock MP 1.3.0 that even uses the wrong backbuffer that is masked by the "half-full-screen" issue where MP did not honor the size of the display is shown on. Think of a reverse half-full screen issue that is now visible.

Quoting myself..

Ouch, that goes even deeper. Basically the whole MP code assumes that only one display is connected per adapter and that adapter and screen number always match. Ouch. That can be right but does not have to. Depends on how a GPU exposes the ports to windows. The assumption: adapter = display is at least wrong.

"Reverse half-full screen" in effect. Stock MP displayed a screen in larger resolutions than what was available on the display, but Windows changed the size. Ah, not trying to explain. Just a screw up in the code in regards to multi-display handling. This is now just visible...

I think by accident I found the reason why the refresh rate changes always uses the refresh rate of the primary screen. Or at least something that is related here. Most broken on XP and multi-display setups.

Will take me a couple of hours to get this right, possible that I break something here on the first try.

 

[elliottmc]

Do we know if the people who had problems with Intel GPUs in the previous build have tested this one? 90 downloads so far and very few issues, so I guess we are close!

 

[azzuro]

I use it now on productive, except Plugin error, no real problem appears !!! but i can't test the original bug ! "Half-fulscreen"
Tv Plugin :
DVB-T : more fast compared to 1.3
DVB-S : more fast compared to 1.3
Myvidéo : No issue found
Music : Not tested
Pictures : Not tested

 

[Scythe42]

Do we know if the people who had problems with Intel GPUs in the previous build have tested this one? 90 downloads so far and very few issues, so I guess we are close!

No reply yet.

 

[megahorst]

I have found an issue regarding the windows screensaver which is available with this release:
MePo does not go to sleep on it's own under certain conditions while the screensaver is on. I'm not sure where this thing has to be handelt. MePo or PS++ @michael_t: can you please have a look at this too?

Steps to reproduce:

• screensaver idle time shorter than power scheduler ++ idle time
• play music
• screensaver is triggered
• music stops
• PS++ idle time is over
• MePo does not go to sleep

I have attached the log

 

[Lightning303]

Just encountered a problem i have never seen before, and i dont think anybody else has described it :/.
Im running a stardard system (see system specs), 1 tv, 1 avr, no extra monitors or something like that.
As is side note, currently i have a problem where my htpc wakes up all the time by "itself", i have an rf remote, and the receiver is picking up signals from my neighbour. So thats why my system woke up often during the day. You can see that in the logs. And honestly to test hibernate/resume behaviour i think its helping ;P, as PS++ send my htpc after 15 minutes of doing nothing back to hibernate.

So started MP yesterday at 2:17, watched a video and then turned of my avr and tv. A download was still running until 6:24, thats when PS++ send my htpc to hibernate. Every wakeup from then on until the one at 20:00 was a wakeup by my rf receiver, and my avr and tv were off at that time.
At 20:00 my htpc woke up the fetch epg data. At around 20:30 i wanted to start watch a movie (htpc still running, fetching epg), so i started my avr and tv. From the logs it seems that it was 20:36.
Now, finaly, the problem .
I couldnt see mp at my tv. I saw my desktop, as if mp crashed or wmere closed. However i also didnt see my taskbar. Just desktop without taskbar. Moving the mouse didnt show me a cursor or anything.
It seems that i can not reproduce this problem, and as i said, first time i see this behaviour. Maybe you can find something in the logs.
I will do more tests after the movie .

 

[Scythe42]

[2013-04-20 06:09:48,607] [Error  ] [MPMain  ] [ERROR] - Main: OnResume - OnResume called but MP is not in suspended state.
[2013-04-20 07:21:02,593] [Error  ] [MPMain  ] [ERROR] - Main: OnResume - OnResume called but MP is not in suspended state.
[2013-04-20 09:08:34,801] [Error  ] [MPMain  ] [ERROR] - Main: OnResume - OnResume called but MP is not in suspended state.
[2013-04-20 10:01:37,271] [Error  ] [MPMain  ] [ERROR] - Main: OnResume - OnResume called but MP is not in suspended state.
[2013-04-20 12:33:36,616] [Error  ] [MPMain  ] [ERROR] - Main: OnResume - OnResume called but MP is not in suspended state.
[2013-04-20 13:42:56,263] [Error  ] [MPMain  ] [ERROR] - Main: OnResume - OnResume called but MP is not in suspended state.
[2013-04-20 15:54:03,197] [Error  ] [MPMain  ] [ERROR] - Main: OnResume - OnResume called but MP is not in suspended state.
[2013-04-20 16:17:46,584] [Error  ] [MPMain  ] [ERROR] - Main: OnResume - OnResume called but MP is not in suspended state.
[2013-04-20 18:31:46,650] [Error  ] [MPMain  ] [ERROR] - Main: OnResume - OnResume called but MP is not in suspended state.
[2013-04-20 19:57:14,706] [Error  ] [MPMain  ] [ERROR] - Main: OnResume - OnResume called but MP is not in suspended state.

That's why you don't see anything. Will check the logs for more details what is missing in your setup.