Resuming incomplete download doesnt work

[Cosmicbertie]

Hi,

I've been attempting to download an ISO of approx 61Gb (split into 58 1Gb RAR files)...due to my slow connection I have to frequently leave my PC on overnight to achieve this. However I've noticed that on the occasion that my PC decides to spontaneously reboot during the night (updates/crashes/etc) when I restart Newsbin it doesnt resume the download, it restarts again.

This appears to only affect incomplete files....so any files which have been downloaded and converted into a binary/RAR file show up.

For instance, I'd downloaded approx 40Gb of the overall files, and this was represented in the 'Current Progress' section as being over half complete. But when Newsbin restarts, it shows it at 3Gb downloaded (and unrar'd into files) and starts downloading all the chunks again from any incomplete file.

My 'chunks' directory is over 14Gb in size, but Newsbin wont read in any of the data which its already downloaded. Is there any way to force this?

The version im running is: 6.73: 00 D5 71 97 66 12

Thanks
CB

[Quade]

That's to depredador I was actually able to track this down. It happens to file sets that lie about their filename in the headers for the download. The subject might or might not reflect the correct filename but the filename in the posts themselves are just random numbers and letters. So what happens is on restart the same random filenames are seen by Newsbin but because Newsbin has already renamed the files using the subject into a human readable filename, it couldn't find them after restart.

I'll have a fix in the next 6.80 beta.

As for the chunks, that might be incomplete files awaiting retries to force an assembly or it could be assembly is getting behind. Assembly can get behind when you have a really fast connection speed ( >400 Mbps) and you allow Newsbin to download and unrar/repair at the same time. The unrar stalls the file assembly process so, the chunks stack up.

I'm probably going to change that so that file assembly can continue during the unrar.

[depredador]

Does it make a difference if you download headers? Like should that be turned on or off or it doesn't matter either way?

And should either box be ticked in the Filename Options/Duplicate Filename Settings?

[Quade]

Doesn't matter where the files come from. The issue is really the random extension-less filenames in the posts.

[Cosmicbertie]

Cool. I saw the other post by depradador, but didnt know there was an outcome.

When is the 6.8 beta coming out? I'll give it a try and report back.

[Quade]

Not sure. I was just working on long (> 260 character) filename support. I'm contemplating a major change in how download and repair work. I probably want to put up a beta before I experiment with that change.

[depredador]

I'll give it a try and report back.

Same. There are some things I want to download that are over 40gig in size but I can't do it because of slow internet connection and can't keep my notebook on 24/7 so need it to be able to resume. It used to do it but I'm not sure if uploaders have changed file types that makes something in Newsbin work differently .

[depredador]

Any idea when a Beta with hopefully a fix for this will be out?

[Quade]

It's on IRC channel right now being tested. Might be tomorrow or the day after.

[depredador]

Cool.