Here is my shell script:

I have this script in my /home/mythtv/bin folder. Use chmod +e on this file to flag it executable. Then make a log directory, in my case mkdir /home/mythtv/log. The script will create a log file in this folder.
Then set up you web directory, in my case mkdir /var/www/html/xmltv, do this as root and chown mythtv.mythtv /var/www/html/xmltv to assign ownership on this. Next you edit the existing (http://www.better-access.com/KM/tv_grab_au) grab script to point to your local web server instead of minnie.tuhs.org, in my case http://kingtide/xmltv/, So my two modified lines read:
my $default_root_url = 'http://kingtide/xmltv/channels.xml.gz' and
my $default_root_url = 'http://kingtide/xmltv/channels.xml.gz'
Almost there, your channel data needs to be configured, in my case I live in Melbourne. Modify the pre_tv_grab_au shell script with your username/password and your channels. Then you need to create a channels.xml.gz. so I created a file called /var/www/html/xmltv/channels.xml as shown here:

Then zip it up, cd /var/www/html/xmltv, then gzip channels.xml. To check it do zcat channels.xml.gz.
Now ready to test, so execute the script and check the log file for result. Once you have seeded your local xmltv folder you need to run your modified grabber, do /usr/local/bin/tv_grab_au --configure. This will download from you local server xmltv area and prompt you for username, password, base url and channel info. Here is how my config file looks.

Now run the grabber /usr/local/bin/tv_grab_au and check it now works.
Then run mythfilldatabase to load in the new guide data to the mythtv database.
Here is my cron entry:

and here is my $HOME/bin/runmythfilldatabase script:

My setup is running fedora core 6 and my hostname is kingtide.
That is about it, I have just got mine going so I am doing final testing on the cron job.
This is a workaround, so I am wondering if I should use the Shepherd method when I have time to changeover? comments welcome on this.
Anyhow, let me know if you need more help,

Cheers,
deMyth

Correction, changes to tv_grab_au should read:

my $default_root_url = 'http://kingtide/xmltv/channels.xml.gz'; and
my $url = 'http://kingtide/xmltv/';

kingtide is my servername, use yours in place of mine.

hope this helps.

Cheers,
deMyth
--------

I am getting this same problem, is there an easy fix other than that complex script work around?

No, I was unable to find a fix, so I wrote this workaround until there is a fix available. Maybe try the Shepherd installation. In my work around I have just another minor correction, should read chmod +x for executable. I have now tested my cron job and looks OK. Good to see I have 7 days of guide info, I was almost out.

deMyth

What is the story? Mine is doing this too... just after I thought it was working fine

2008-04-20 18:49:21.293 Connected to database 'mythconverg' at host: 192.168.0.3
2008-04-20 18:49:21.295 Grabber Command: nice tv_grab_au --config-file '/home/mythtv/.mythtv/FreeToAir.xmltv' --output /tmp/mythnotM2t --days 1 --offset 1
2008-04-20 18:49:21.295 ----------------- Start of XMLTV output -----------------
Failed to fetch http://minnie.tuhs.org/xmltv/channels.xml.gz at /usr/bin/tv_grab_au line 346.
2008-04-20 18:49:21.826 ------------------ End of XMLTV output ------------------
2008-04-20 18:49:21.827 FAILED: xmltv returned error code 2304.
Error in 1:1: unexpected end of file


Why is it doing this? Surely it only needs a tweak not a full on script hack??!

Cheers

Bj

_________________
Backend/Front : Core2 1.86Ghz 1GBRAM 500GB HDD 32" LCD
Frontend1:VIA M10K Diskless 17" LCD Kitchen
Frontend2:VIA M10K Diskless 22" LCD
Frontend3:Centrino in cubby

Would be great if someone can supply a quick hack, I think the perl code is failing to authenticate as it works against my local site with any credentials. I found wget would authenticate so I had to write a script using this until someone can help fix the tv_grab_au perl script or help with a proper alternative.

deMyth

It *was* working fine. I don't know what changed but I suspect the TUHS web server may be blocking it because of some header in the request. I asked Warren about this but he hasn't responded yet. Very annoying for all of us. If not cleared up tomorrow, I may have to recommend that we all go the Shepherd way.

Mike

_________________
*********************
LinHES 7.4
Australian Dragon
*********************

I just installed and config'ed Shepherd. What a ripper! The thing is awesome because it's way more accurate - Now this is a brilliant thing 'cause now the missus doesn't whinge about the shows not lining up with "the real things" on tv. 2nd it works.

Lets make this the default because if you can't get data from Oztivo at least something else will fill it in! I love a backup plan!

Here is what I did:

OzTiVo Shepherd Changeover


Cheers

Bj

_________________
Backend/Front : Core2 1.86Ghz 1GBRAM 500GB HDD 32" LCD
Frontend1:VIA M10K Diskless 17" LCD Kitchen
Frontend2:VIA M10K Diskless 22" LCD
Frontend3:Centrino in cubby

Hi BJ et al,

I agree with you. I've had two probs with oztivo and don't like the incomplete data thing. We should go to Shepherd.

Thanks for your howto bj, but you didn't need any more than three packages installed, and I think it can be scripted for upgrade.

I also think it should be the default installation for next release.

Will attempt an auto changeover script later today.

Mike

_________________
*********************
LinHES 7.4
Australian Dragon
*********************

In case anyone is still interested, the problem is indeed that the perl code is not authenticating. Replacing the three occurrences of http://minnie.tuhs.org/ in tv_grab_au with http://YOURUSERNAME:YOURPASSWORD\@minnie.tuhs.org/ and the root url in the config file in a similar way (but without the \ before the @) is a quick fix. Not sure exactly that all of those are needed, but it definitely works this way.

before we add shepherd to the next release, one issue needs addressing:

My status page now alwas reads as follows:



This may just be because i have 2 seperate sources adn the second one contains no new data. Is anyone with only one source seeing this?

Nope. I also have 2 sources and using Shepherd, but I'm not seeing that. I typically get this:

(The reason for the quick runtime (usually one or two minutes) is because it's the output following the second source run, which uses cached data.)

Thanks for your help with this simple hack, seems like the minnie.tuhs.org was not enforcing authentication until mid last week when I noticed the problem. I put my shell script workaround to rest and spent time last night setting up shepherd. It works well, although I couldn't get the cpan JavaScript to complete. I caught up with a mate yesterday who also created his own workaround to the problem, he used curl with auth params:
.

funny how a broken tv grabber script brings script hackers out of the woodwork.

Cheers,
deMyth

I have the same issue with Shepherd, tried all sorts of ways to get JavaScript.pm to complete but to no avail It only affects the running of the Rex grabber though. If you ever work out how to get that pm installed properly please post.

Yes, truly hilarious!

I fixed my script to use wget rather than curl. Also, the fix mentioned earlier only allows the retrieval of the channel listing. I applied the wget mod for the data also.

The new script is at http://www.better-access.com/KM/tv_grab_au
The diff is as above, but append .diff

I tested it, and it works quite well. The next version of KM won't have my Perl script, though (my setup scripts choose Shepherd as the oztivo default).
The reason is this: My script totally relies on oztivo's people telling me when they change something (which they don't) and there is no redundancy. You have the choice, but I'll be sticking with Shepherd for now.

Cheers

Mike

_________________
*********************
LinHES 7.4
Australian Dragon
*********************