Open GoogleCodeExporter opened 8 years ago
Can you send me the offending .amz file privately so I can look and see if the
format has changed?
Original comment by SCColbert@gmail.com
on 11 Sep 2012 at 4:20
sent
Original comment by cuc...@gmail.com
on 11 Sep 2012 at 4:27
I'm having the same issue. Per your comment to cuc...@gmail, 2 files are
attached:
<> A .amz file that produced the same result.
<> A text file (pymazonOutput) showing the console dialog when running Pymazon
from the command line. Although the console output claims something was saved
to /home/Will/Downloads/AmazonMP3, that directory remains empty.
I don't know how to get the traceback output that cuc... showed.
This is Pymazon 0.9.1 on Debian Squeeze.
May I suggest: Include the version number in the opening comments of the
pymazon program. I'm not aware of any way to tell what version one has, other
than to install the latest version - replacing what could already be the latest
version.
Original comment by Will.Sti...@gmail.com
on 15 Sep 2012 at 7:14
Attachments:
From what I could tell last time, Amazon is now giving out empty AMZ files on
the download page. There's nothing wrong with Pymazon, the AMZ files just don't
have any track records. You have to get the real AMZ file from the cloud player
directly. Sadly, they block this feature for linux users. I tried a lot of
fancy spoofing tricks to make Amazon think I was a normal windows users, but
somehow none of them worked. The only workaround I've found so far is to
download each track individually. Good luck.
Original comment by cuc...@gmail.com
on 15 Sep 2012 at 12:12
The AMZ files aren't empty, they're all 283 bytes. I don't know how large they
were before, though. I still have three of them in my /tmp folder, all for the
same song, in case you need more.
Original comment by darth.br...@gmail.com
on 17 Sep 2012 at 5:48
The file is effectively empty; all it contains is
<playlist>
<title>Cloud Player Download</title>
<extension application="http://www.amazon.com">
<endpoint>https://www.amazon.com/cirrus/</endpoint>
</extension>
<trackList/>
</playlist>
(normally trackList would have links to the songs in it!)
It works normally for me if I use an MSIE User-Agent. (Same goes for the Cloud
Player, although you need to log out and back in. With an Iceweasel
User-Agent, the Cloud Player will refuse to let you download more than one file
at once.)
See also:
https://code.google.com/p/clamz/issues/detail?id=35
https://code.google.com/p/clamz/issues/detail?id=36
Original comment by Benjamin...@gmail.com
on 17 Sep 2012 at 7:32
http://chrispederick.com/work/user-agent-switcher/
Works fine with UAS set to MSIE.
Original comment by rickr...@gmail.com
on 18 Sep 2012 at 1:20
I'm not sure if this is related.
I tried pymazon on an .amz file from a single file purchase (!), which did not
work either.
It did report the title of the track, But I also got a Value Error from the
download thread.
File "/usr/local/lib/python2.7/dist-packages/pymazon/core/downloader.py", line 92, in _download
fs = int(obj.filesize)
ValueError: invalid literal for int() with base 10: ''
I decrpyted the .amz file using the decrypter in pymazon.
The resulting xml file contained a non-empty playlist and a link. I did use the
link to download the MP3 with wget.
I used pymazon 0.9.1 (I think, files are from Oct. 15, 2011).
Original comment by Michael.Strecke
on 30 Sep 2012 at 5:16
Ok - after some digging into the source code:
The error described above occurs when using the cli version of pymazon.
The problem seems to be, that the .amz file does not contain the track's file
length anymore, which in turn messes with way the program is calculating the
download progress.
Original comment by Michael.Strecke
on 30 Sep 2012 at 8:02
Michael, I think the issue you listed is covered in Issue 53. (It's slightly
different than the one listed here, I think.) I've created a working patch in
that thread, though it's not very elegant.
I'd welcome anyone who had a better grasp of the language to produce something
more pythonic.
Original comment by jsphs...@gmail.com
on 8 Oct 2012 at 10:40
Original issue reported on code.google.com by
cuc...@gmail.com
on 11 Sep 2012 at 4:17