1748 replies to this topic

[Rob van der Does]

Of course I won't use it.
But as I said: it beats me that anybody would. For the reasons Erik mentioned.

[jpuigs]

Talking about channel lists there is a thing which I'd love to have it (and possibly some other users too) : to be able to "unsort" channel lists, especially in satellites (green key) section.

Then I'd sort them like I do with favorites.

If it's not possible, I'd love too to choose to get them sorted by frequency (from lower to higher) instead of names.

Edited by jpuigs, 8 September 2014 - 17:57.

[Erik Slagter]

But you can simply not use it.
A large part of what I suggest, I personally do not use, but it has asked for users, for which it is very important.

Maybe it's better to explain to the user why it works this way and also it may be better to find out what the user really wants (which often appears to be something not exactly the same) and solve that instead.

In this case it's often related to parental control. A better solution might be e.g. optionally locking the "all services" and "satellites" lists using a pin code, instead of damaging the database.

[Taapat]

Sorry, my English is not good, but I have already explained why it is necessary when propose the patch.
It seems to me, that you do not create your bouquet by yourself. Therefore, you can not imagine that there are users who are important to know, which of the channels are working with the same name, when creating a bouquet on computer.
If we accept all TV channels deletion, why not accept the deletion of channels one by one?

If this option is not acceptable, of course you can revert it. I only propose, and your choice is to accept it or not. In my receiver it works, and if anyone do not need it, for me is not very important if it is in openpli image or not.

Edited by Taapat, 9 September 2014 - 17:43.

[Erik Slagter]

As a matter of fact I do create all my bouquets myself.

 

I suggest you do this:

 

- select one by one all satellite positions

- press menu on it

- choose "create bouquet from satellite position"

 

Now you have "master bouquets" you can edit at will and derive other bouquets from.

 

We might want to implement something similar for the "all services", i.o.w. create a "master bouquet" as a copy of the "all services".

[jpuigs]

Thanks, I already knew this way.

By the way, the option I have on menu is called "copy to bouquet", not "create bouquet from satellite position".

This option lets you edit and order all satelliteas you want, but there is asmall problem.

When background scanning is working, new channels are added to satellite lists, not to favourites. So new channels won't appear on these new master bouquets.

That's why I asked to be able to choose to order channels into satellite lists either by name (actual behaviour) or by frequency.

Then If you choose "by frequency", channel lists could have 2 colour bars to see differences between transponders.

Look at the attachment I've made.

 

 

Attached Files

•  frequencysortchannelmenu.jpg   152.28KB   42 downloads

[Erik Slagter]

Yes exactly, the background scanning, which is going to overwrite any changes you make to the "all services" at some time.

[zakalibit]

Hi,

 

again something for httpsream, I found that not all redirect codes are handled, just 302, I came across at least one service returning 301 which is a valid redirect code so could we change 

if (statuscode == 302 && strncasecmp(linebuf, "location: ", 10) == 0)

as per rfc to 

if ((statuscode==301 || statuscode==302 || statuscode==303 || statuscode==307 || statuscode==308) && strncasecmp(linebuf, "location: ", 10) == 0)

Edited by zakalibit, 15 September 2014 - 18:27.

[Erik Slagter]

Applied.

[lonas]

Here it is.

 

Please!! have available package opera-hbbtv_20140725_0.tar.gz ???

[Taapat]

I see that in AudioSelection have added translation for languages​​, but the same is needed for subtitles.

Attached Files

•  0001-AudioSelection-Add-gettext-for-subtitles.patch   796bytes   4 downloads

[Taapat]

If we try python speed test we can see that if we use 'is' instead of '==', then we can get 2 times higher speed:

python -mtimeit -s 'a="tututu"' 'if a == "tatata": b=a'
1000000 loops, best of 3: 1.88 usec per loop

python -mtimeit -s 'a="tututu"' 'if a is "tatata": b=a'
1000000 loops, best of 3: 0.952 usec per loop

Of course 'is' not the same as '==', 'is' we can use only if the parameters are identical, and not use if equal is their values.
I have not had time to check all enigma2 python source code, but here are 3 patch in which I've changed in some Components, Converters and Screens '==' to 'is' and '!=' to 'is not' for better performance.

Attached Files

•  0001-Change-in-Components-to-is-and-to-is-not-for-better-.patch   26.66KB   11 downloads
•  0002-Change-in-Converters-to-is-and-to-is-not-for-better-.patch   20.26KB   6 downloads
•  0003-Change-in-Screens-to-is-and-to-is-not-for-better-per.patch   38.29KB   3 downloads

[littlesat]

Also when it is running from pyo files, like enigma 2 does....

Edited by littlesat, 14 October 2014 - 22:57.

[Taapat]

Of course I am no python expert, but do you want to say that running compiled files python using other methods, and performances will not be?

[Erik Slagter]

Time to try I guess...

[Taapat]

Time to try I guess...

 

Ok, create two files and compile them.

test1.py:

def Test():
  a = "tatata"
  for x in range(1000000):
    if a == "tututu":
      b = a
test = Test()

test2.py

def Test():
  a = "tatata"
  for x in range(1000000):
    if a is "tututu":
      b = a
test = Test()

Then run the test:

spark:~# time python /usr/test1.pyo
real    0m 3.44s
user    0m 2.84s
sys     0m 0.38s
spark:~# time python /usr/test2.pyo
real    0m 2.95s
user    0m 2.38s
sys     0m 0.37s

[addybo]

did you test the output?

 

because if i use "is" , it is not working

 

 

maybe this is reason, it works faster....

[littlesat]

Do you have some reference from the www why is is better performing than ==?....

[tension]

Of course 'is' not the same as '==', 'is' we can use only if the parameters are identical, and not use if equal is their values

 

the answer is in the question, so errors in the patches

>>> a="DVB-C"
>>> b="DVB-C"
>>> id(a)
140602432801120
>>> id(b)
140602432801216
>>> a is b
False
>>> a == b
True
>>>

[Taapat]

Do you have some reference from the www why is is better performing than ==?....

I am too lazy to search, but I think that the answer can be seen in the tension example.
Python do not need to get real value to compare, for example, he can use the id, and therefore work faster.
 

the answer is in the question, so errors in the patches

Therefore, I tried to change only for those variables, which are assigned a value, which compares them later. For example:

self.type = EPG_TYPE_MULTI
...
if self.type is EPG_TYPE_MULTI:

But perhaps I have something changed too much  .