Re: [Pan-users] Re: Some remarks about the user interface

[David Kelly]

On Sep 27, 2006, at 6:52 PM, Duncan wrote:

> > - How are the different (primary) newsservers related anyway? I can't
> >   seem to choose between them or anything.
>
> You don't choose between them in the UI.  It's a single group  
> listing and
> single post display -- unified management.  However, you can set
> priorities, so for instance pan will go to low/no-cost servers  
> first, and
> only try the premium servers if it can't find the posts on the  
> first set.
> For simplicity, the GUI config only shows two levels of priority (in
> server settings), but you can edit the servers.xml file directly to  
> any
> number of priorities.  There are a number of other no-gui config- 
> options
> available as well (including the $PAN_HOME environmental variable,
> allowing multiple sessions with different settings or simply  
> allowing you
> to place pan's working dir as desired). I list about six of them in a
> post from last week.

0.113 and 0.114 seem to behave better than prior versions I have  
tried, but just what is the expected behavior when the "backup" (via  
GUI config) server is not online? Or was online and falls off while  
the current instance of pan is running? And what happens when the  
backup was initially offline and is brought back online?

Currently if the backup is not online when I start pan, pan behaves  
as I would hope. Previously it would only run at about 2/3rds of the  
bandwidth if both primary and backup were online, and decay to  
nothing within 30 to 60 minutes. Is currently up to 95% of the  
reported download bandwidth of 0.14. Haven't benchmarked actual to- 
disk rates other than it really did run at about half the rate of 0.14.

In earlier versions it appeared to me the backup would be assigned  
tasks whether it was available or not. That a "get new headers" task  
might be assigned to the backup and never revert to the primary. Yet  
the task would stick in the Task Manager window as 100% complete.  
This appears to have been addressed. As has the lesser download  
bandwidth which eventually dropped to nothing.

I'd really like the Task Manager to indicate which tasks were on what  
server. And at the top of the window rather than just the aggregate  
data rate, break it up by servers, that one might notice backup isn't  
being used, might as well take it offline.

A problem I run across a lot with usenetserver.com is that an article  
listed in headers is not available. This broke 0.14 and prevented any  
of the other articles in a multipart posting from being assembled.  
The download stopped at the missing article yet the multipart headers  
were bound into one line item which could not be split back into its  
components. Think it was "download text" which would download all the  
articles available and then one had to yydecode externally. If the  
article was not listed in headers then pan 0.14 would assemble  
whatever it had and leave the job of repair to par2.

In this case I learned with 0.14 I could bring the backup server  
online, update headers, and select the broken article. Then pan would  
restart where it left off on the usenetserver.com account. Could stop  
the download from the alternate server, delete from queue from the  
first, re-select it on the first, and it would once again run until  
it hit a listed but missing article. Guessing this is how 0.114  
behaves with primary and backup servers.


--
David Kelly N4HHE, address@hidden
========================================================================
Whom computers would destroy, they must first drive mad.