1748 replies to this topic

[Erik Slagter]

An alternative solution might be to stage the dvb headers from the kernel, instead of using the linux-libc-headers.

That sounds to me as an excellent idea. I never understood why one would want to duplicate kernel headers into libc so that they don't match the current kernel (unless maybe you want to compile for an older kernel, which in this case, imho, is irrelevant).

[athoik]

Is there a runtime check to get DVB_API_VERSION ?
 
Here they mention the following:

 

If you use newer kernel should work ok why is that a problem ? kernel
syscalls are backward compatible
and we use oldest kernel version of like 2.6.16 for eglibc to build
against so it should not be a problem
to use different versions for kernel and kernel-headers. I think your
problem lies elsewhere

[MiLo]

All boxes share the same libraries, so everything must be compiled against the same kernel headers, regardless of the actual kernel used.

You'll have to use runtime checks for everything that may be kernel dependent. Compile-time checks on kernel features will not work.

We've been doing this in the wrong way since the beginning, that's why we got away with it so long.

[mx3L]

Found issue in Input.py:

--- a/lib/python/Components/Input.py
+++ b/lib/python/Components/Input.py
@@ -62,7 +62,7 @@ class Input(VariableText, HTMLComponent, GUIComponent, NumericalTextInput):
                        self.currPos = 0
                        self.Text = u""
                else:
-                       self.Text = text.decode("utf-8", "ignore").decode("utf-8")
+                       self.Text = text.decode("utf-8", "ignore").encode("utf-8")
                self.update()
 
        def getText(self):

Attached Files

•  0001-Input-fixed-typo-don-t-decode-unicode-but-encode-it.patch   805bytes   4 downloads

Edited by mx3L, 19 February 2015 - 15:06.

[littlesat]

Pushed

 0001-Input-fixed-typo-don-t-decode-unicode-but-encode-it.patch   

[mx3L]

Thanks!

[peteru]

It looks like the same problem is also in insertChar. You may want to amend the patch.

Edited by peteru, 20 February 2015 - 05:13.

[Erik Slagter]

Commits cannot be ammend once they're pushed. But it's no problem to make a new commit.

[Taapat]

After the decode("utf-8") was changed to encode('utf-8') I have a problem with the bouquet name editing if the name is with cyrillic characters.
If I trying to do this, then I have gsod:

File "/usr/lib/enigma2/python/Screens/VirtualKeyBoard.py", line 323, in okClicked
    self["text"].deleteBackward()
  File "/usr/lib/enigma2/python/Components/Input.py", line 217, in deleteBackward
    self.update()
  File "/usr/lib/enigma2/python/Components/Input.py", line 58, in update
    self.text = self.Text.encode("utf-8") + " "
  File "/usr/lib/python2.7/encodings/utf_8.py", line 16, in decode
    return codecs.utf_8_decode(input, errors, True)
UnicodeDecodeError: 'utf8' codec can't decode byte 0xd0 in position 14: invalid continuation byte
(PyObject_CallObject(<bound method NumberActionMap.action of <Components.ActionMap.NumberActionMap instance at 0x2cd818cc>>,('OkCancelActions', 'ok')) failed)

I do not have original openpli image, but I think that on it is exactly the same problem.
Try create test bouquet, change the virtual keyboard layout to RU, and rename or delete any of the symbol.
If I simply remove encode('utf-8') then everything works without problems.

Edited by Taapat, 21 February 2015 - 18:12.

[ims]

I can confirm it. For lat2 is GSOD too.

 

There must be .decode("utf-8") for input nat. chars ! Why encode for input ? See http://sourceforge.n...dfeb9d45d3d4eb/

 

Please - revert it.

 

 

@mx3L - where you had problem, that you needed change it ?

Edited by ims, 21 February 2015 - 22:48.

[mx3L]

I'm passing utf-8 encoded string as text to VirtualKeyboard

Screen <class 'Screens.VirtualKeyBoard.VirtualKeyBoard'>(('Editova\xc5\xa5 n\xc3\xa1zov s\xc3\xbaboru',), {'text': 'HDTV by Lidu\xc5\xa1ka(Broadchurch.S02E04.HDTV'}): <type 'exceptions.UnicodeEncodeError'>
Traceback (most recent call last):
  File "/usr/local/e2/lib/enigma2/python/mytest.py", line 226, in create
    return screen(self, *arguments, **kwargs)
  File "/usr/local/e2/lib/enigma2/python/Screens/VirtualKeyBoard.py", line 74, in __init__
    self["text"] = Input(currPos=len(kwargs.get("text", "").decode("utf-8",'ignore')), allMarked=False, **kwargs)
  File "/usr/local/e2/lib/enigma2/python/Components/Input.py", line 25, in __init__
    self.setText(text)
  File "/usr/local/e2/lib/enigma2/python/Components/Input.py", line 65, in setText
    self.Text = text.decode("utf-8", "ignore").decode("utf-8")
  File "/usr/lib/python2.7/encodings/utf_8.py", line 16, in decode
    return codecs.utf_8_decode(input, errors, True)
UnicodeEncodeError: 'ascii' codec can't encode character u'\u0161' in position 12: ordinal not in range(128)
Traceback (most recent call last):
  File "/usr/local/e2/lib/enigma2/python/Components/ActionMap.py", line 46, in action
    res = self.actions[action]()
  File "/usr/local/e2/lib/enigma2/python/Plugins/Extensions/SubsSupport/subtitles.py", line 3760, in editFName
    self.session.openWithCallback(editFnameCB, VirtualKeyBoard, _("Edit Filename"), text= toString(self["fname"].text.strip()))
  File "/usr/local/e2/lib/enigma2/python/mytest.py", line 294, in openWithCallback
    dlg = self.open(screen, *arguments, **kwargs)
  File "/usr/local/e2/lib/enigma2/python/mytest.py", line 305, in open
    dlg.isTmp = True
AttributeError: 'NoneType' object has no attribute 'isTmp'

Sorry guys, you're right, self.Text has to be unicode, and self.text utf-8 encoded string.

This should be alright... please check

diff --git a/lib/python/Components/Input.py b/lib/python/Components/Input.py
index 680e95d..a1afe33 100644
--- a/lib/python/Components/Input.py
+++ b/lib/python/Components/Input.py
@@ -62,7 +62,7 @@ class Input(VariableText, HTMLComponent, GUIComponent, NumericalTextInput):
                        self.currPos = 0
                        self.Text = u""
                else:
-                       self.Text = text.decode("utf-8", "ignore").decode("utf-8")
+                       self.Text = text.decode("utf-8", "ignore")
                self.update()
 
        def getText(self):
@@ -152,7 +152,7 @@ class Input(VariableText, HTMLComponent, GUIComponent, NumericalTextInput):
                self.update()
 
        def insertChar(self, ch, pos=False, owr=False, ins=False):
-               self.Text = self.Text.decode("utf-8", "ignore").decode("utf-8")
+               self.Text = self.Text.decode("utf-8", "ignore")
                if not pos:
                        pos = self.currPos
                if ins and not self.maxSize:

Edited by mx3L, 21 February 2015 - 23:31.

[ims]

it is better

[mx3L]

Good ,

probably "self.Text = self.Text.decode("utf-8", "ignore")' in insertChar should'nt be there at all since self.Text should be still unicode, so no need to decode it, but it will not raise anything since errors are ignored.

[Taapat]

And if we have started talk about the input, here is a patch to add Latvian layout in VirtualKeyboard.

Attached Files

•  0001-Add-Latvian-layout-in-VirtualKeyBoard.patch   1.83KB   2 downloads

[littlesat]

Pushed

 

 

  0001-Add-Latvian-layout-in-VirtualKeyBoard.patch   1.83KB   1 downloads

[ims]

And if we have started talk about the input, here is a patch to add Latvian layout in VirtualKeyboard.

@tappat - you can add LAT into NumericalTextInput too. See into .../tools/NumericalTextInput.py

[Taapat]

Pushed

Thank you!
 

@tappat - you can add LAT into NumericalTextInput too.

I think about this, but in the end decided not to do so.
In Latvian is addition "soft" letters, and therefore increases the number of characters in a one digit. It seems to me therefore that not comfortable to use NumericalTextInput, and where used only it, could by used without national symbols, but maybe I'm wrong.
I know that other languages have a lot of character, and maybe you have any argument that I'm wrong and it does not interfere input.

[ims]

NumericalTextInput is same as SMS mod on GSM phone. May be, you could inspire there...

[Taapat]

Maybe you right. Looked to other languages, it seems that Latvian is not the worst case.

Attached Files

•  0001-Add-Latvian-layout-in-NumericalTextInput.patch   1.01KB   3 downloads

[Taapat]

I did not find an appropriate topic, therefore, decided to write here. I'm sorry if I had to write another place.
I see that in ServiceName been added a new option- TransponderInfo. Very well, now, to show this information will no longer be used non-standard components. 
I have three suggestions:
1. It seems that in the displayed of polarization have been forgotten about circular polarization which begins with the word "circular". The first patch fixed it.
2. I do not know what satellites.xml is in openpli feed, but I use it fromt the site http://satellites-xml.eu/ and there sat name is such as "Astra 4A / SES 5 (4.9E)", of course, that it is not properly displayed. The second patch fixed name, if in the sat name is "(".
3. I add name of the server if in the bouquet is IPTV service.

Attached Files

•  0001-ServiceName-fix-polarization-display-if-cirkular.patch   978bytes   3 downloads
•  0002-ServiceName-fix-orbital-position-on-non-standart-sat.patch   1001bytes   4 downloads
•  0003-ServiceName-show-server-info-in-stream-services.patch   947bytes   2 downloads

Edited by Taapat, 22 February 2015 - 18:14.