special characters tvguide.xml (1 Viewer)

[boeloeboeloe]

When I grab the program guide, I have always problems with special characters.
Something like this in the file tvguide.xml: "Clips van ��n artiest op een rij"

Is there a solutions for that?

 

[gibman]

first off, your xml file needs to be encoded in utf-8 format ?

How do u obtain ur xml file ?

webegp parsing a website ??

try and post a small section of your file.

/gibman

 

[boeloeboeloe]

The xml-file is encoded in utf-8 format.

I grab the file with the mediaportal webepg (from de site www.teveblad.be, but same problem with other sites).

When opening with notepad, you can see that there is a little space in the title-tag "America s" and in desc-tag "homevideo s". The required character is a single quote.

When opening with dreamweaver, the character is showed correctly.

When loading in Mediaportal is represented as a square.

Hope that you have enough information?

Thanks in advance.



"<programme start="20071207182500" stop="20071207192000" channel="Kanaaltwee-kanaaltwee.be">
<title>Buffy, the Vampire Slayer</title>
<sub-title>Surprise</sub-title>
<desc>Komische horrorserie.
Buffy heeft een nachtmerrie waarin ze ziet hoe Drusilla afrekent met Angel. Ze vraagt zich af of haar droom realiteit is en gaat naar Angels appartement. Hij blijkt ongedeerd en stelt haar gerust dat Drusilla dood is. Zij is er echter niet van overtuigd dat dat wel zo is. Over de Meester had ze net zo’n nachtmerrie en even later viel hij haar aan. Dat ze het bij het rechte eind heeft, blijkt wanneer blijkt dat Spike en Drusilla een demoon op haar afsturen die niet gedood kan worden...</desc>
<category>serie</category>
<episode-num>13</episode-num>
</programme>
<programme start="20071207192000" stop="20071207200500" channel="Kanaaltwee-kanaaltwee.be">
<title>Most Shocking</title>
<desc>Video-opnames van waargebeurde en schokkende feiten.</desc>
</programme>
<programme start="20071207200500" stop="20071207203500" channel="Kanaaltwee-kanaaltwee.be">
<title>America’ s Funniest Home Videos</title>
<desc>Show (VS) met grappige homevideo’ s.</desc>
</programme>
<programme start="20071207203500" stop="20071207232500" channel="Kanaaltwee-kanaaltwee.be">
<title>Bad Boys II</title>
<desc>Actiekomedie van Michael Bay.
Mike Lowrey en Marcus Burnett, twee detectives in Miami, maken jacht op de beruchte drugsbaron Tapia alias Johnny, die van plan is de stad te bevoorraden met extacy. Hun taak wordt er niet eenvoudiger op wanneer Sidney, de zus van Marcus en een agente van de drugsbrigade, probeert te infiltreren in het witwasnetwerk van Alexei, een Russische zware jongen...</desc>
<category>film</category>
</programme>"

 

[THDBASED]

Can confirm this. I am using the exact same guide and ' is showing as a "?"...

 

[Furetto]

In another thread I posted a fix for this.

The reason behind this is that TeveBlad is not using the normal apostrophe (ascii code 39) but the right curly quote mark (ascii code 146) and I guess the WebEPG grabber currently does not interprete this well.

The fix I used was to use the ACTION command to replace all & # 39; by & # 146; [spaces added to prevent replacing of the codes] in both TITLE and DESCRIPTION.

A more permanent fix would be to treat the & # 146; correctly inside webepg.

*grumble* It would seem that now the description is sometimes added twice... Need to investigate

 

[Furetto]

Gibman, maybe indeed you can help out here.

Excerpt from the website

<td nowrap class='detailtop' width='100%'>Debbie Travis& #146; Facelift
</td>

Comment: added a space between & and # to prevent the forum from replacing the code

In the TVGuide.xml this becomes

 <title>Debbie Travis# Facelift</title>

where the hash replaces a character that is not represented correctly. Depending on the font, codepage etc used, you see either nothing (default font) or a square replacing a non-representable character (in Times New Roman) or the & # 198 ; character Æ (when in Uestudio I click on OEM to ANSI) or finally the apostrophe & # 146; ’ (when in Uestudio I click ANSI to OEM, but this borks all other special characters).

As I mentioned in the other thread, I can workaround the issue for the TITLE by modifying the end of the www_teveblad_be grabber into

  </Listing>
  <Actions>
    <Modify channel="*" field="#TITLE" search="& #146;" action="Replace">& #39;</Modify>
    <Modify channel="*" field="#DESCRIPTION" search="& #146;" action="Replace">&# 39;</Modify>
  </Actions>
</Grabber>

Comment: again added a space between & and #

But with this code, the #DESCRIPTION tag gets doubled in many cases. Instead of simply saying

<desc>Kookprogramma.</desc>

I suddenly get

<desc>Kookprogramma. Kookprogamma.</desc>

(This last is from memory, I had it happen like this several times yesterday but removed the resulting files and don't feel like rerunning all just to prove I'm right )

 

[Furetto]

*bump*

 

[Furetto]

To start: sorry for hijacking this thread

For a long time, the Belgian webepg script is giving a small issue. The single quote character is rendered as a question mark. I don't know enough of programming to go dig into the code and fix it myself, but I think in the thread where this is discussed I provide information that could allow a developer to fix this rather easily.

To sum up in short, this website uses & #146; (space added on purpose) instead of & #39; as single quote or apostrophe. The normal single quote gets 'escaped' normally, but the one used on this website does not.

James, do you think this could be fixed easily ? My workaround I mentioned in the thread is not workin as good as I hoped...

 

[James]

Hi Furetto,

I have seen this thread and I will make a change to fix this problem for you

The problem is not really with WebEPG, but that MP doesn't support UniCode characters

/James

 

[Furetto]

Aah wonderful

Well, you deserve my Best Wishes for the New Year !!!