DSM-G600, DNS-3xx and NSA-220 Hack Forum

Opperpanter wrote:

hmm ok, didn't know that, BUT a restart will sometimes indeed cause you to loose some good connections and stuff..

I just wrote a script to handle a watch folder myself.

If interest is there, I can post it here on the forum.

As requested, I made two simple scripts.

/ffp/scripts/tranmissionwatcher.sh

#!/ffp/bin/sh

WAIT_SECONDS=300
WATCH_DIR=/usb/transmission/torrent

while true 
do
    for file in $WATCH_DIR/*.torrent
    do
        echo $file
        if [ "$file" != "$WATCH_DIR/*.torrent" ]
        then
            echo adding "$file"
            echo [`date`] "$file" added to queue. >> /usb/transmission/log/transmissionwatcher.log
            transmission-remote -a "$file" >> /usb/transmission/log/transmissionwatcher.log
            mv "$file" "$file".queued
                fi
        sleep $WAIT_SECONDS
    done
done

I start this script at reboot by using

/ffp/start/starttorrentwatch.sh (make sure it is chmod a+x to survive reboot)

#!/ffp/bin/sh

cd /ffp/scripts
./transmissionwatcher.sh >> /usb/tmp/transmissionwatcher.log &

It checks every 5 minutes (300s) for new torrent files to add. If any files found, they're added to transmission without restart. The torrent files are renamed to .torrent.queued so they're skipped the next time.

I build this script based on somebody else's, I am not sure who/where, but credits go the original author.

Sorry guys for my abstinence, but I switched the ISP and that apparently takes 2 weeks. *grml*
I'm back online now, and will be updating Automatic shortly. I already found a rather serious bug some of you might already have run into.
When the download history gets full, Automatic crashes
It's already fixed and currently in testing, if it works well a release shall follow tomorrow.

I've also made first tests with a watch folder implementation which looks very good, it just needs some kinks here and there.

I'll keep you guys posted!


PS: For those who tried to download while the server was offline, you should be able to download the binaries again.

lekeur wrote:

I telnet : "/mnt/HD_a2/ffp/start/tranmissionwatcher.sh start" but it doesn't work. It returns :

Code:

/ffp/bin/sh: /mnt/HD_a2/ffp/start/transmissionwatcher.sh: not found

How can I resolve it. ?
Sorry but I'm totally noob with linux.


Thank you in advance.

If you want to run it by hand, just type

sh /mnt/HD_a2/ffp/start/transmissionwatcher.sh &

If you follow my exact instructions in previous post it works fine.

Last edited by Opperpanter (2008-09-24 16:50:34)

thanks for your work!

unfortunately the link to the source code doesn't work. could you fix it please.

I finally found the time to update the source package.
So, anyone feeling inclined to take a look at bicker/laugh/whatever, go ahead

I'm always open for suggestions, and would really appreciate it if you find a bug (or two).

Hm, good question.
I use libcurl to do the HTTP stuff, and curl does support HTTPS. But it requires libssl, which Automatic currently is not linked against, so the current version definitely can't do it.

Do you know of a feed I can play around with? (preferrably a public one)

some private feed doesn't have "enclosure" so
show "Is this really a torrent feed?" infinitly
how do i do?

Muppet wrote:

Unfortunately the one I have in mind is a private tracker that I cannot supply a link for.

Sorry for the late reply, I've been rather busy lately.

That said, I actually did get to test whether Automatic can deal with HTTPS feeds, and to my surprise, it can.
That's because libcurl is compiled with openssl linkage, so as long as that library is present on your system (it should, Transmission requires it), you can add HTTPS feeds to Automatic.

Please note that Automatic cannot handle authentication, however. So if your feed requires a password of some kind, this is not available (yet). I've been planning to add authentication/cookie handling, but couldn't think of a nice way to do it yet.

Regular expressions (at least the POSIX version) is great at matching strings, but not at all usable to NOT match something.

So all you can do is try to find some difference in the comparison string that is present for non-720p releases but not for 720p releases. This depends on the feed you're using. If you let me know what feed you're using, or what each feed entry looks like in XML, I can try to create a regular expression that does not include specific characteristics.

I'm afraid not. The eztv feed doesn't list the file name of each episode, but rather a "readable" string.

You can use this:

"Heroes.*\[HDTV"
"Prison Break.*\[HDTV"

But please note that this will most likely only work for this feed.

Thanks for the help. How do you keep the search based feed from downloading every episode?

I guess you would have to set a starting episode in your regular expression to filter out all the old episodes.

Say you wanted to start from 3x6 (and yes you're getting into complicated territory here):

Heroes.*(3x([6-9]|[1-9][0-9])|[4-9]x[0-9]?[0-9]).*

That should cover you from 3x6 to 9x99. I'll leave it with you to work it out if they run to a 10th season!

I highly recommend http://www.regextester.com/ to test this sort of thing out

As mentioned above, excluding certain kinds of items is currently very difficult and only possible with rather complex regular expressions. I'm currently testing a new regex engine though, which does support exclusion. With that it will be rather simple.

I'll keep you guys posted.