68 replies to this topic

[WanWizard]

I don't see anything in that commit that would cause a change of the display string?

[blzr]

well, after reverting changes to VidroHardwre.py and HardwareInfo.py, voila:

Attached Files

•  hdmi.jpg   47.47KB   7 downloads

[athoik]

well, I'm pretty sure it's been displayed correctly - HDMI port for boxes that have it...
after taking a look in latest commits, this is my first suspect:
https://github.com/O...e2799a6760f9715

(changes in VideoHardware.py and HardwareInfo.py)

I don't see anything in that commit that would cause a change of the display string?

There is a logic error when checking model.
 

#echo "my super box" > /tmp/hwmodel
#python

>>> device_name = "???"
>>>
>>> try:
...     device_name = open("/tmp/hwmodel").read().strip()
...     print "hwmodel"
... except:
...     print "pass"
...     pass
... else:
...     device_name = open("/tmp/model").read().strip()
...     print "model"
...
hwmodel
Traceback (most recent call last):
  File "<stdin>", line 8, in <module>
IOError: [Errno 2] No such file or directory: '/tmp/model'
>>> print device_name
my super box

So the else part executes when there is no exception (http://stackoverflow.com/a/16138256)!

Anyway I think commit needs improvements to meet OpenPLi standards...

Edited by athoik, 25 September 2015 - 18:23.

[athoik]

I suggest to use something like the folowing:

SystemInfo["ScartYPbPr"] = not fileExists("/proc/stb/info/hwmodel") and not HardwareInfo().get_device_model() == "fusionhd"

By checking the get_device_model we don't need the following:

 		# Name ... bit odd, but history prevails
 		try:
-			self.device_name = open("/proc/stb/info/model").read().strip()
+			self.device_name = open("/proc/stb/info/hwmodel").read().strip()
 		except:
 			pass
+		else:
+			self.device_name = open("/proc/stb/info/model").read().strip()

[pop_eye]

 		# Name ... bit odd, but history prevails
 		try:
-			self.device_name = open("/proc/stb/info/model").read().strip()
+			self.device_name = open("/proc/stb/info/hwmodel").read().strip()
 		except:
 			pass
+		else:
+			self.device_name = open("/proc/stb/info/model").read().strip()

I think the original logic was that the following is needed for all boxes:

 

self.device_name = open("/proc/stb/info/model").read().strip()

 

Not sure why fusionhd needed to rename this.

Edited by pop_eye, 25 September 2015 - 19:12.

[athoik]

I think the original logic was that the following is needed for all boxes:
 
self.device_name = open("/proc/stb/info/model").read().strip()
 
Not sure why fusionhd needed to rename this.

Actually when model proc entry didn't exist there was no problem (pass in except), now it will fail!

Edited by athoik, 25 September 2015 - 19:13.

[pop_eye]

But we need the proc entry to identify the receiver model...

[pop_eye]

Actually is used for finding the device name which is passed then to find the model.

 

if self.device_model is None:
            self.device_model = self.device_name

        # HDMI capbility
        self.device_hdmi = (    self.device_name == 'dm7020hd' or
                    self.device_name == 'dm800se' or
                    self.device_name == 'dm500hd' or
                    (self.device_name == 'dm8000' and self.device_version != None))

        print "Detected: " + self.get_device_string()

[Rob van der Does]

well, I'm pretty sure it's been displayed correctly - HDMI port for boxes that have it...

I can confirm that for my Xtrends and VU's; it has always been 'HDMI', but recently changed into 'DVI'.

[littlesat]

I think in the specific commit there exists
changes we as openpli usually never should expected and fusion needs to adapt their drivers.....

Athoik at least has the good idea for the scart detection... And the model was always there to fake enigma 2 for detecting hdmi or dvi as it returns a dmm box....

Edited by littlesat, 26 September 2015 - 07:55.

[athoik]

The following seems to fixes the HDMI issue for me.

diff --git a/lib/python/Plugins/SystemPlugins/Videomode/VideoHardware.py b/lib/python/Plugins/SystemPlugins/Videomode/VideoHardware.py
index d76a820..127ef4f 100644
--- a/lib/python/Plugins/SystemPlugins/Videomode/VideoHardware.py
+++ b/lib/python/Plugins/SystemPlugins/Videomode/VideoHardware.py
@@ -59,7 +59,7 @@ class VideoHardware:
                "640x480" : { 60: "640x480" }
        }

-       if not fileExists("/proc/stb/info/hwmodel") and not HardwareInfo().get_device_name == "fusionhd":
+       if not fileExists("/proc/stb/info/hwmodel") and not HardwareInfo().get_device_model() == "fusionhd":
                modes["Scart"] = ["PAL", "NTSC", "Multi"]
                modes["YPbPr"] = ["720p", "1080i", "576p", "480p", "576i", "480i"]
        modes["DVI"] = ["720p", "1080p", "1080i", "576p", "480p", "576i", "480i"]
diff --git a/lib/python/Tools/HardwareInfo.py b/lib/python/Tools/HardwareInfo.py
index 2ec0f5b..bb58648 100644
--- a/lib/python/Tools/HardwareInfo.py
+++ b/lib/python/Tools/HardwareInfo.py
@@ -30,11 +30,9 @@ class HardwareInfo:

                # Name ... bit odd, but history prevails
                try:
-                       self.device_name = open("/proc/stb/info/hwmodel").read().strip()
+                       self.device_name = open("/proc/stb/info/model").read().strip()
                except:
                        pass
-               else:
-                       self.device_name = open("/proc/stb/info/model").read().strip()

                # Model
                for line in open((resolveFilename(SCOPE_SKIN, 'hw_info/hw_info.cfg')), 'r'):

[littlesat]

In fact we could get rid of that complete hardwareinfo.py.... Just keep a stripped version for backwards compatibility (plugins) and move all stuff to systeminfo... And add a scartavailable check...

We should check on features, not on boxes... At least not on multible places...

I'm willing to change this... But when doing so Imcannot guarantee it will work for all diferrent boxes the next day... So I really need help/testing here...

It seems devicd name is only used to detect hdmi... Probably we should change that completely... Only some dmm boxes the 8000 and the 800hd do nit have that...

I noted that he patch should be for the remote only.... But some additinal stuff went also through by accident...

Edited by littlesat, 26 September 2015 - 09:05.

[littlesat]

i just partly reverted the commit that causes this topic.... for all the other boxes it should be fine from now on...  https://github.com/O...1a1df6c249d0852

 

2Do ...

find a different way for the fusion hd boxes to detect it has HDMI... or even completely change this stuff......

Add a way to detect a box has Scart or not...

[Rob van der Does]

And the feature-detection is also about RCA & Component?

[littlesat]

Yep... And I think SystemInfo might be the best place....

[Rob van der Does]

And then the ET8&10k get WoL again.....?

[littlesat]

10K should get wol..... Dima blocked ET8K...

[Rob van der Does]

Yes, but I mean that this is an 'if 8k then no wol' and I thought PLi never wants to do that? And now you're going for the 'features' we get wol back on 8k because the proc-entry is there?

[littlesat]

In systeminfo.py this is somehow accepted als there features are determined for phython

[Erik Slagter]

I see two "proper" solutions.

 

1) just name the output "DIGITAL" (or something like that)

2) just name the output "HDMI" (because the only still supported model (how long...?) that has a DVI-connector is the DM8000)

 

Let's not add all sorts of hardware-dependend code just to make sure the name of the output is "correct".

 

For those who missed it, HDMI and DVI are completely identical for "home and garden" use, you can use a (passive!) adapter cable to connect one to another.