NZB Completion Checker: New Version=New Features. Help needed testing =)

[Hypatia]

http://www.zoon.dk

Not only interface has changed.(you can check it out here: http://www.zoon.dk/2012/02/01/nzb-completion-checker-1-0-beta-released/ )


But some new features have been introduced as well:


1) SSL support

2) Multi-server setup.

One can set a primary server and the rest will be backup servers.


3) Password checker

Speaks for itself

4) Validation Methods ( HEAD & STAT)

SOme providers have problems with STAT method (like readnews resellers) so we have a choice here.

Personally i have some problerms with alpha version, it keeps crashing when showing results, so ive been using a beta recently

Also, what needs to be noted is percentage of completeness shown on 2 colored bars in the middle is the same for multiserver and single server bar.

Anyways, all this needs testing. Unlike nzb download checker NZBCC is in active development and we have a chance to help its author and improve it. AFter all its us who need this software.

Report any bugs through this form:
https://558733f6.servage-customer.net/index.html

[zot]

Thanks for the news.

I must admit that I've backed off using completion checkers lately, as I've found they're often only vaguely accurate (and in the case of Readnews, it can be faster to download the file than check its completion) but no doubt they're improving. I wonder if any use or plan an EAC-like approach, checking multiple times to ensure accuracy and repeatability?

I've noticed that actual completion (per downloading) generally tends to be slightly higher than a checker's listed completion - does anyone else find that also? Do they intentionally err on the side of conservation or what?

I've wondered if any existing newsreaders (like UE?) could be modified somehow to start an article download and then immediately cancel it -- as it seems that checking for an article (as these NZB completion checkers do) and actually fetching it (as a newsreader does) are different processes with sometimes different outcomes.

I keep promising myself that one of these days I'll finally get DotNet working again ...

[Hypatia]

I've noticed that actual completion (per downloading) generally tends to be slightly higher than a checker's listed completion

i might be wrong but it seems to happen when you do checking via STAT method which is supposed to be faster but it might be inaccurate in some cases.

[B18C5]

Newsbin's NZB checker uses "HEAD" because "STAT" is unreliable. For some servers, the "STAT" only comes from headers. If you have a server like Easynews or Newshosting with 1300 day file retention but, a cap of 120 million headers per group, you don't really want to use anything that depends on the headers. An NZB checker should be much faster than actual download, particularly if it uses multiple connections and only does HEAD. HEAD might be 1K of data compared to 400-600K for a regular post.

Newsbin obeys your backup server settings too.


I've wondered if any existing newsreaders (like UE?) could be modified somehow to start an article download and then immediately cancel it -- as it seems that checking for an article (as these NZB completion checkers do) and actually fetching it (as a newsreader does) are different processes with sometimes different outcomes.

You can't cancel an article download. The protocol doesn't permit it. Disconnecting is all you can do once you've asked for the article if you don't want it. "HEAD" should be the same as downloading an article because XOVER data (headers) doesn't contain all the header fields. "HEAD" means they have to go to the actual article and extract the data. Some servers are organized as server farms and not all the servers in the farm will have all the posts. I suspect that might be why you see different results on subsequent runs. I'd suggest setting your checker to always use HEAD if you want reliable results.

[Hypatia]

it has nothing to do with stat and head in this particular case. it ws a bug of nzbcc and author will fix it as he said

colored bars showed multi-server 100% completenes but numbers below were telling different story

Just a bug=)

[B18C5]

The point though is that "STAT" can be hit or miss though, it's very fast. "HEAD" is slower but, more reliable. "STAT" could be reliable, if every server used the actual records for verification but, I know for a fact that not all servers do this.

[Hypatia]

Well, its a good thing then that NZBcc has HEAD verification as well

[Hypatia]

The next alpha is on its way=)


The next release will feature:

A fixet installer – my bad – sorry :)
Fixed the results being shown on the results screen. Multiserver/singleserver text was the same, though the bars showed different results (the bars were right!)
On odd occations the settings was not saved


PPS Its here!

It should now be possible to connect to a server without a username/password
Online checker was extended to 15 days
Changed the .NET framework from Client to Full framework (sorry, had to!)
Fixed the installer to install the full version of the framework if missing
You can now enable/disable a server. When it’s disabled it will not be used during server traversal. However, you cannot disable the primary server
Fixed a severe bug using multiple servers, and the server traversal tool which could lead to an infinite checking loop
Fixed the breakdown form (again!). The bars should now show the correct result for each completion bar. Also, the correct number of bytes is also displayed when using multiple servers.
Fixed other minor bugs!

http://www.zoon.dk/2012/03/28/nzb-completion-checker-update-3/#more-382


PPPSS Fixed version is released
http://www.zoon.dk/2012/03/29/nzb-completion-checker-alpha-update-4/