"Check before download"- methods?

[semel]

Hi.

Thanx for a very handy and long awaited feature=)

However i have a few questions regarding it

1) What method does sabnzbd use? HEAD or STAT checks?(newsbin pro uses HEAD, nzb download checker uses STAT, nzb completion checker allows users to choose one of them)

STAT is much faster BUT on certain providers it doesnt work properly(readnews resellers-blocknews\usenetnnow etc

HEAD is slower but more accurate and all providers afaik support it


2) Does this check utilize backup servers?


3)Will we have a chance to set what method should be used manually? Blocknews is quite popular these days as a backup provider and if it is STAT method that is used..well..you get the picture.


Thanx

[shypike]

STAT is used instead of HEADER.
The idea is to waste as little bandwidth as possible.
The pre-check is meant for people with download quotas or low bandwidths.
Using HEADER would mean more bytes to download.
Personally, I find STAT already slow.

Pre-check is a dry-run, so all servers are used in the normal pattern.

As the usefulness of pre-check is not certain, we don't go overboard offering options for it.
Your suggestion is noted.
You might suggest to blocknews to get their act together and conform to Usenet standards

[sander]

3)Will we have a chance to set what method should be used manually? Blocknews is quite popular these days as a backup provider and if it is STAT method that is used..well..you get the picture.


Thanx

If you run SABnzbd from source, technically you could replace STAT by HEAD yourself: in sabnzbd/newswrapper.py do this:

Code: Select all

            # command = 'STAT <%s>\r\n' % (self.article.article)
            command = 'HEAD <%s>\r\n' % (self.article.article)

HTH

[semel]

The pre-check is meant for people with download quotas or low bandwidths.

to be honest i think it will help mostly users who deal with all those dmca takedowns (ggianews\supernews\newshosting etc) and AW's bad range(~200days affected) rather than those with quotas.(im talking here in numbers overall =))

Your suggestion is noted.

thanx

You might suggest to blocknews to get their act together and conform to Usenet standards

if it was that easy=)

If you run SABnzbd from source, technically you could replace STAT by HEAD yourself

i run it on win 7 64 and,alas, not familiar with compiling on it at all

[sander]

STAT is much faster BUT on certain providers it doesnt work properly(readnews resellers-blocknews\usenetnnow etc

HEAD is slower but more accurate and all providers afaik support it

BTW: What is the result of a STAT on those non-compliant newsservers? A "500 What?", or something else?

[shypike]

I'll see if something can be done for the next 0.7.0 Beta.

[semel]

BTW: What is the result of a STAT on those non-compliant newsservers? A "500 What?", or something else?

hmm it seems they've "fixed" this, sort of. i remember getting article not found via STAT check on blocknews when in reality they were available.

ive given it another try and here what i found

On new files STAT checks work fine on blocknews.

But on older files it is somewhat troubling

1) very slow(but blocknews is known for its slower speed on older stuff so we can do nothing about it)
2)some articles dont get processed with nzb download checker showing "501 newsgroup" message

i suspect it has something to do with time-outs during the checking process on blocknews 500 days file which is REALLY slow

head check is slow as well but it doesnt give this message and all articles are checked on the same nzb


In other words its not a good idea to check old files on any readnews reseller and its totally their fault

Anyways

I'll see if something can be done for the next 0.7.0 Beta.

i suppose it would still be nice to be able to change methods. Perhaps through INI if it is easier to do

Cheers

PS Btw STAT is not that slow as you say it is.

AW, 700 days old, 8.8Gb file - appr. 2 min
Though it depends on completion- if there are missing articles, a lot, it will be slow=)

[Bren0man]

Hey all,

I know I'm necroing a thread to the max but I just wanted to say a big thank you for implementing this feature and for creating this thread that I stumbled across via google. I've only recently resumed usenet activities after a hiatus of quite a few years. Since returning back I've had no end of incomplete downloads, frustration and wasted quota from troubleshooting.

After much googling and trial and error, I've come to realise that DMCAs are definitely a thing in 2014 (5 years ago I rarely encountered an incomplete file) and that the 'check before/abort incomplete download' options are my saviour (that and my complementary Astraweb service that my ISP provides is riddled with incomplete files which I assume are caused by DMCAs).

There's probably a good reason for this but would it not make sense for the pre-download check to be operational by default? It sure would have wonderful for this n00b lol.

Anywho, thanks again. If I had not come across this post, I'd still be searching for a solution, inching ever closer to snapping my keyboard over my own head and being forced to rely solely on torrents.

Cheers,
Bren0

[shypike]

There's probably a good reason for this but would it not make sense for the pre-download check to be operational by default? It sure would have wonderful for this n00b lol.

The check is slow with many Usenet providers. Also it isn't always reliable (also depending on provider).
We think the "Abort if" options are better.
The choice is yours, but we enable neither option by default.

[Bren0man]

Fair enough. Thanks again!