• Hi,
    ich habe meine Feeds-Liste ein bißchen überarbeitet und da ist mir aufgefallen das die Umlaute mal stimmen mal nicht. Ich habe versucht dagegen was zu machen. Ich erzeuge den XMLParser jetzt auf Basis von encoding= aus dem XML-File. Es klappt bei mir mit heise (utf-8) und chip (ISO-8859-1). Ob auch andere Formate unterstützt werden kann ich nicht sagen.


    Anbei das Plugin zumindest für aktuellen CVS-Stand, die Quellen und meine Feeds. Ich habe Feeds von heise, chip, pc-welt, toms hardware, lycos, spiegel und tagesschau dazu gepackt. Falls ihr noch interessante Feeds kennt wäre ich dankbar.


    Wäre vielleicht nicht schlecht wenn jemand die Änderung einchecken könnte falls es zusagt.


    cu B.

  • Hi,
    gilt das auch für Dreambox?


    Irgendwie trau ich mich nicht so recht nachher verhunze ich mit einem Befehl das gesamte CVS:-)


    cu B.

    -- Die Lücke die ich hinterlasse ersetzt mich vollständig --

  • Hi,


    die Angst hatte wohl jeder am Anfang, aber bei solchen Sachen kann kaum 'was passieren. cvs ist ziemlich gut dokumentiert, google ist dein Freund. Und wenn du Fragen hast, kann ich dir - soweit es meine Möglichkeiten zulassen - gerne helfen.
    cu

  • Is it possible to have the encoding set by the feeds.xml, something like:


    Code
    <feed>
    	<name>Feed_name</name>
    	<url>http://feed_url/news.xml</url>
    	<encoding>ISO-8859-9</encoding>
    </feed>

    This way we could also use rss for feeds using other encoding than ISO-8859-1.

  • Hi OKA,
    this is more or less what I have changed with one small difference.
    Your feed like for instance


    http://feed_url/news.xml


    should have a line with encoding in it. I took a look and the feeds I saw had such a line. Do you have a feed without an encoding line? If you have can you give me an example with the correct encoding? Then I can try to take a look if it´s possible.


    There is also the question which encoding enigma supports.


    cu B.

    -- Die Lücke die ich hinterlasse ersetzt mich vollständig --

  • Hi Kat-CeDe,


    I have to find some more feeds, but this one I have now is working but not displaying the right encoding in the old RSS reader (X-newsreader).


    http://www.sabah.com.tr/rss/spor.xml


    But with your plugin choosing this feed produces an error: XML parse error (general xml).


    I looked into the source of my feed, there is an encoding line and a language line. Don't know if you meant that.


    Enigma supports at leaest the most common encodings. Have a look into /var/tuxbox/config/enigma/encoding.conf

  • Hi OKA,
    your link uses windows-1254 as encoding and I am not shure if enigma supports this. I tried to exchange windows-1254 with iso-8859-9 and I checked if new Parser worked but I always get the parser error.


    I have no solution at the moment.


    cu B.

    -- Die Lücke die ich hinterlasse ersetzt mich vollständig --

  • Hi OKA,
    bad news for the moment. It seems that the XMLTreeParser doesn´t like the encoding ISO-8859-9. I changed the program at the moment so that it only switches between ISO-8859-1 and UTF-8. Every other encoding is shown with -1. So at least you can see your feeds again. I tested it with setting the encoding fixed to -9 and got the parse error even without my changes.


    I found no hint which encodings are supportted fully.


    Has someone a list which encodings are supported or how I can enhance the encodings?


    Attached corrected version.


    Sorry for the moment Ralf

  • Hi Kat-CeDe,


    Too bad it didn't work out.
    I'm not a developer, but I do know that the language files (enigma-mo) are in UTF-8. Maybe you could make all ISO encoding convert to UTF-8. I found this:



    Also found something called "libiconv": http://www.gnu.org/software/libiconv/


    Thanks in advance.

    Einmal editiert, zuletzt von OKA ()

  • Hi OKA,
    I am still trying to get better language support.
    I already tried some internal functions of the DreamBox yesterday but it didn´t work as supposed. There is one function called convertUTF8DVB I used but up to now without much success.


    I found out that the Parser only seems to support 8859-1, UTF-8, UTF-16 and something like ACSII-US (??).


    cu B.

    -- Die Lücke die ich hinterlasse ersetzt mich vollständig --

  • Hi OKA,
    could you please test the attached version? I think I found a way and turkish looks OK for me but I hardly speak german and english so what can I know.


    Because I didn´t find a library function I copied the routine for converting ISO to UTF-8 from eStrings. If it works for you than RSS should support ISO-8859-1 up to 16 with the exception of 12 and UTF-8. For your first test link with windows-1254 I use 8859-9 as a work around.


    hope that works for you


    Ralf

  • Hi Kat-CeDe,


    I haven't visited this forum since months, so I just noticed you made a new version.


    Just tested the last version, and it works OK now with Turkish characters.
    Great work and thanks for your time.


    Is this version also being pushed into CVS?

  • Hi OKA,
    i cannot tell if it´s in the CVS and at the moment I have no dream:-(


    I hope the dream is not over for me:-)


    cu B.

    -- Die Lücke die ich hinterlasse ersetzt mich vollständig --