[Top][All Lists]
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [Pan-users] Checking server for missing parts
From: |
Wayne E. Nail |
Subject: |
Re: [Pan-users] Checking server for missing parts |
Date: |
Tue, 05 Jan 2010 14:10:36 -0500 |
On Sun, 2010-01-03 at 20:23 -0800, john wendel wrote:
> Sorry I didn't make myself clear in earlier posts.
>
> I'm trying to understand the problem described in the "crazy christmas
> lights" post. I have an NZB file (from binsearch.info) that should fetch
> the data, it fails spectacularly with my news server. Pan reports many
> missing articles. I'm suspicious of Pan at this point because some of
> the error status messages appear to be asking for articles that are not
> in the NZB file.
>
> There seem to be 2 possible causes of the failure, [1] Pan is doing the
> right thing and the articles are missing on the newsserver or [2] Pan is
> asking for the wrong articles. I thought it would be useful if I had a
> tool that could query the newsserver for each article listed in the NZB
> and receive a yes/no status, without trying to download 1.6 GB of data.
>
> I don't think that psychic divination is required to validate an NZB
> file against a particular server. You just need to be able to parse each
> "section" of the NZB file, and do a "STAT <article-id>" command to the
> server for each section. You could even do it by hand with a simple
> telnet connection to the server, but it would get tedious for an NZB
> file with 6000 sections.
>
> Regards,
>
> John
Okay, so you got my curiousity jones all excited ;-)
I grabbed what was probably the same .nzb you did, from a binsearch.info
on the subject string in your original post. Said return was oddly
lacking any detail about the multiparts.
I fed the .nzb to Pan 1.33 (from 09SEP09 git checkout [checkout probably
the wrong term for git but I'm new to it]), and it downloaded 1.3 GB in
6,124 files, almost all of which looked like this:
248 Crazy Christmas Lights-2009-12-06-0_copy_1000.tp
4 Crazy Christmas Lights-2009-12-06-0_copy_1000.tp.ERRORS
248 Crazy Christmas Lights-2009-12-06-0_copy_1001.tp
4 Crazy Christmas Lights-2009-12-06-0_copy_1001.tp.ERRORS
248 Crazy Christmas Lights-2009-12-06-0_copy_1002.tp
4 Crazy Christmas Lights-2009-12-06-0_copy_1002.tp.ERRORS
...etc. ad nauseum
The .ERRORS files all look like this:
Warning: Missing everything before part #6745
But hark! Note the size of the single .tp file at the very end of this
mess. Why, a CD-sized file all decoded and everything. What be it?
664388 Crazy Christmas Lights-2009-12-06-0.tp
24 Crazy Christmas Lights-2009-12-06-0.tp.ERRORS
628 Crazy Christmas Lights-2009-12-06-0.tp.nzb
The .ERRORS file for this CD-sized .tp file is full of 379 entries like
this:
ERROR: %Part 2554 missing. Decoded file probably corrupt.
...but Pan decoded what it could find on Easynews.
Next, the google showed me that .tp is a file type for 'MPEG-2 TV
recorded file: File extension is used for MPEG-2 TV recorded file (Use
mpeg-2 compression).'
So I fed the CD-sized .tp to Kaffeine and it plays just fine. It's the
first 14 minutes of an hour-long show about over-the-top xmas
decorations aired on The Learning Channel, featuring setups you've
probably already seen on YouTube and such.
Is this a borked Usenet post? Boy howdy, is it ever! What kind of idjit
would post in that format in the first place? A clueless one, is what
kind ;-)
I've never seen this particular flavor of borkitude before, but it
appears to me that Pan is as uncontaminated by blame for this mess as
the poster is uncontaminated by clues.