Version 6.73 Release Available...

[dexter]

The spam flood from January slowed down the release schedule and we are now moving forward with the 6.73 final release. Only minor changes since the last version was rolled out that made it easier to apply header download filters. Also opening the door to anyone who is available to update the Non-English translations through our Translations Wiki. Anyone providing translations support is eligible for some free access to our Usenet Search service.

Change Summary:

• Translations enabled in the installer
• Local Search now respects "Search in Groups"
• Fixed issue where ctrl-r wouldn't read certain posts with non-standard filenames (typically spam)

The download link is on the Newsbin Beta Page. If you find any issues, please reply to this thread or use our Technical Support Contact Form.

[kirm]

Program installed fine, just wanted to mention that the installer asked me twice for what language I preferred.

Just letting you know.

Kirm

[dexter]

Yeah, it does that. It's an issue with the NSIS installer. It asks again once it requests admin rights to perform the install. Annoying, I know.

Glad everything else is working.

[kirm]

There seems to be one issue I have come across, but do not think it is just this build.

If I dl headers from a particular group, and let's say half of them were incomplete while the others were complete due to just being posted like say one hour ago. I grab the completed ones with no issue.

Say I wait an hour or more, update all headers, then go back to that particular group, the ones that were incomplete previously remain incomplete and any newer postings are all complete. I have tried using special download for just this group, trying every option and still these files show incomplete. I then dl all headers for this group, and this lets me finally dl the incomplete files because they are now all showing as complete.

I have noticed this behavior on previous builds also. One would think dl 10 million older headers in this small group would complete the files.

[Quade]

I'm not seeing it. It suggests you might want to use more "overlap" in the server options.

I had some incompletes in the anime groups, normally missing 1-2 chunks per files (which is fixable with PARS). I looked at the raw GZ files and the missing part data was missing in the raw data from both Giga and UNS. You still might want to make sure that you security software both ignores the data folder and ignores Newsbin's connections to the news server.

Might be interesting to see if they're incomplete in our search engine. What server are you downloading headers from?

[kirm]

Might be interesting to see if they're incomplete in our search engine. What server are you downloading headers from?

Am using newshosting. Think you may be missing the point here. They are incomplete at the time I opened the group and of course time for the servers to get all the parts due to still being uploaded like within the past hour.

It's when I come back in say a couple of hours after another update of headers, I get the remainder of the files that were not showing earlier and they are all complete. It's the ones that showed incomplete 2 hours ago, do not change and keep showing as incomplete. I tried special dl and go for every option, and still no change. I have to dl all headers from this small group once more, and then all the files come in completed. It's kinda like I jumped in too early as this poster generally uploads on 2 or 3 consecutive days. They are mostly apk files, so usually not that large other than the obb folders.

Now if I haven't jumped into the group early, I always get all the files as completed. It's just jumping in while still being processed by servers, so I wait several hours before going back and the previous incompletes always remain incomplete till I dl all headers for that group over again. Sorry for repeating myself somewhat.

Hope I explained myself properly.

[Fisherking]

Dear Dex & Quade

The bug below still exists in this 673F1

viewtopic.php?f=44&t=41264

Thanks

[johnisme]

It's not a bug in newsbin pro,it's the way newshosting is updating the files,it shows a lot of files with incomplete headers but they get filled after awhile.

[kirm]

It's not a bug in newsbin pro,it's the way newshosting is updating the files,it shows a lot of files with incomplete headers but they get filled after awhile.

The problem is they are complete after a couple of hours, just newsbin will not show them as complete. Like I mentioned if I wait couple of hours and download all headers for group, then they show completed. It's like they will never show up as complete unless I do this. Seems like newsbin has marked them as incomplete and they remain that way. Downloading special is not correcting this.

[dexter]

Quade answered this already:

It suggests you might want to use more "overlap" in the server options.

It means they might be backfilling missing headers which would fall outside the range of headers Newsbin will be downloading the next time you do an update. The overlap tells Newsbin to go back "x" records from the last record it downloaded to pick up anything new that might have been inserted as an older record number. You usually don't need to do this, that is why it is an option.

[kirm]

The overlap tells Newsbin to go back "x" records

Any suggestions on what size to try the overlap

[Quade]

I'd do 1000 to start. Honestly if you server really is back filling that might not be enough.

Am using newshosting. Think you may be missing the point here. They are incomplete at the time I opened the group and of course time for the servers to get all the parts due to still being uploaded like within the past hour.

No I'm not. Sometimes files really are incomplete so, checking from another source will let you know if they're incomplete everywhere or not.

You can force Newsbin to re-download N days of headers by setting a "Download Age" to something. Maybe 2 days. Then right click the group "Post Storage/Use Download Age" which makes it re-download the last "Download Age" worth of header. That's like a super-overlap.

[kirm]

I'd do 1000 to start

OK, thanks Quade

[don_colchon]

Local search continues to search in all groups instead of a single group with 6.73 Build 4821...
Last working release was 6.72 if I remember correctly.

1. Right click on 1 group inside a group-of-groups and select "Search in Groups" => Window opens.
2. Enter some search string and hit Enter. => It starts searching in all groups - instead of the single group i want to search in.

I can restrict the search a little by selecting the group-of-groups in the drop downlist inside the search window. But this does not help much.
Another workaround is to create 1 gog for each group.

I just found another workaround: [b]manually typing in the full name of the group in the drop downlist[/b], e.g. alt.binaries.e-book.blabla => then it searches in exactly the one group. But the group name is not stored in this drop down list. It must be keyed in for each and every search.

(I have a trillion headers, storage age 9999 days)

[Quade]

Yeah that's an easy way to reproduce it. Dex and I were trying to make it fail from the main search tab, not that one.

Thanks.

[Fisherking]

viewtopic.php?f=44&t=41264

Any news about the bug above ?

Thanks

[dexter]

Fell through the cracks, it is not in 6.73.

There is a workaround, you can hit Pause, queue the files for download, then create the nzb from the download list.

[johnisme]

Just curious on the way back filling works,if i look for a certain file on a.b.teevee after 45 minutes it will have a few missing pieces,if i use search for the same file it will show as complete for the same group,i click back to teevee and it will show as downloaded with the file size with pieces missing.

[Quade]

Newsbin lets you set an "overlap" where you tell it to download headers it already downloaded. Normally header download works like:

First Time:
Group: 1-20000
DownloadHeaders: 1-2000
Next Time:
Group: 1 - 25000;
DownloadHeaders 20000-25000
Next Time:
Group: 1 - 30000;
DownloadHeaders 25000-30000

If you have an overlap of 1000 on the server options.

First Time:
Group: 1-20000
DownloadHeaders: 1-20000
Next Time:
Group: 1 - 25000;
DownloadHeaders 19000-25000
Next Time:
Group: 1 - 30000;
DownloadHeaders 24000-30000

It re-downloads 1000 headers is already downloaded.

As long as any "back filling" in the server lands inside that 1000 posts it should pick them up.

If the server though claims to deliver 1-25000 but in fact some of them are missing and it later fills them into the 1-2500 range, there's no way for Newsbin to know that the server has done this. Some servers have operated like this in the past. One way to test for this is to download headers normally, not that incomplete, then right click the group "Post Storage/Use Download Age" then set a shortish download age like 2 which makes Newsbin re-download the last 2 days worth of headers on this group. If the files fill in, then it looks like the server is back filling the wrong way.

[kirm]

As long as any "back filling" in the server lands inside that 1000 posts it should pick them up.

Thanks Quade for explaining how the overlap actually works. Never really understood it till this explanation. I recall a few years ago you suggested the same thing to me, but with new installs of OS, totally forgot about it. I am trying 1000 for overlap right now, but what would you suggest as a max limit. (Consequences?)

Kirm

[Quade]

There's no real downside of a large overlap. You just end up downloading more headers.

One way you could reduce the impact of back filling is by reducing how often you download headers. If you give them more time to accumulate there should be fewer incompletes between header download.

I download headers every 30 minutes. If I had a back filly server, I might make it an hour or more.