0.26.5 : Newsleecher syndrome ?

[Deckard]

With last releases, something weird occurs, making me think about NL behaviour.
Alt.Binz stop downloading some archives when reaching 9X%, with X between 4 and 8.
I have to disconnect then reconnect to end downloads.
Don't know if it comes from newsgroups provider (thread stop) or new priority mechanism introduced.

I have a relative fast connection, with a Core2Duo. What is then the best setting for priority and article caching ?
This is new, previous versions were ok with that. Also, CPU usage increase a lot with these versions.

Thank you.

[Rdl]

I would have to take a look at the log, but I guess it has nothing to do with the program, but with the usenet provider.

Best settings for priority and article caching? It all depends on what are you looking for. If you do other stuff while downloading, set decoding thread priority to idle. Caching depends on few parameters. Size of the file you ussualy download, your download speed, your decoding speed, available memory. Whole purpose is to prevent any article part beeing written to HD. Ussualy good start is 2 x your ussual file size you download + 10%. Note: If you set your cache size to 220MB it DOESN't mean that alt.binz will use 220MB. If needed it will use UP TO 220MB of memory for article part caching

[Deckard]

I would have to take a look at the log, but I guess it has nothing to do with the program, but with the usenet provider.

May be you're right. My usenet provider is Usenetserver which experiencing a lot of problems at the moment.
I use many threads and some of them hang just before finishing download.
I've never seen that before USN problems, whatever alt.binz version I ran.
I can easily recognize this as I was used to it with NL.

Disconnecting then reconnect resolves the problem. But could you handle this directly ?
If an archive stay stuck at 95-99% of download for 10/20/30/... seconds, disconnect the involved thread then reconnect it ?

[Rdl]

Disconnecting then reconnect resolves the problem. But could you handle this directly ?
If an archive stay stuck at 95-99% of download for 10/20/30/... seconds, disconnect the involved thread then reconnect it ?

You already have that option.
Setup->Connection->Global connection timeout->xx seconds

After xx seconds of not receiving anything thread will perform disconnect/connect

[Deckard]

Thx. Will test. Why everytime I ask for sth, it already exists ?

[cr4zyfr4g]

I have set global connection timeout to 30sec and i still get this problem.

Sometimes i have one connection that stays on "Could not connect"/"Connecting" (don't know why, probably server prob) and so last part downloaded with this thread will stay at ~ 98%, because Alt.Binz tries to download last part with the not working thread. Only disabling that thread will help or restart Alt.Binz

To reproduce it (not exactly it, but behavior) open max connections, then one more. Open current collection and you can see download of current part will stuck at ~ 98% (one part of that file). It will never finish it until, disabling thread/restarting Alt.Binz.

[Bad_Ad84]

im getting the same (i think) some parts stay stuck decoding at 98%, have to either delete that part (and let par files repair it) or restart altbinz

[Rdl]

im getting the same (i think) some parts stay stuck decoding at 98%, have to either delete that part (and let par files repair it) or restart altbinz

Log file please in that case.

[Bad_Ad84]

i shall turn logging back on and wait till it happens again and post it.