[Pan-users] Re: So, what's left before 1.0?

[Duncan]

walt posted on Sat, 07 Nov 2009 17:14:54 -0800 as excerpted:

> On 11/07/2009 03:13 PM, Dave wrote:
> 
>> Now I need to find out how to run two separate instances of Pan such
>> that ne'er the twain shall meet.  I prefer to keep binaries and text
>> groups apart since I normally read my text groups while downloading
>> headers and then articles from binaries.
> 
> Duncan has posted the method several times to this list, so I got this
> from him :o)  You can probably google on PAN_HOME to read his old posts.
> 
> Start the second instance of pan with the PAN_HOME variable set to an
> alternative to the ~/.pan directory like this:
> 
> $PAN_HOME=~/.pan-binary  pan
> 
> This will create a new directory ~/.pan-binary, where pan will store all
> of its stuff relating to binary groups.

<voice=Michele-Tanner-aka-Olsen-twins>You got it dude!</voice>

(Talking about stuff the youngin's might not grok, as in the thread some 
weeks ago...)

That's the basic idea, but it can be made much more elegant than that.  
For instance, I have three pan instances here, each with their PAN_HOME 
set to a subdir of ~/pan/.  For config elements that are the same, the 
scorefile, the accels.txt file... I use symlinks to files in a fourth 
subdir, global.  My three pan instances/profiles/whatever are text, test, 
and bin.  Text and bin should be self-evident.  But it's also worth 
noting that when you unsubscribe from a group or visit an unsubscribed 
one, pan doesn't clean up entirely, it keeps some info around.  So the 
the test instance is for "just browsing", checking posts that someone 
mentions here, etc.  Keeping them in a test profile means I can easily 
wipe the profile and keep things nice and neat, without worrying about 
killing regular subscribed groups.

What I actually do is put starter scriptlets in ~/bin/, which is in my 
path: pan.txt, pan.bin, pan.tst, then set the PAN_HOME dir and do a 
couple other things in the scriptlets, before actually starting pan with 
an exec.

Then in my normal app launcher, I setup entries that call these 
scriptlets, one for each one, instead of having just one entry that calls 
pan directly.

It should be noted that even without pan checking the PAN_HOME 
environmental variable, this method would work.  Except in that case I'd 
have ~/.pan2 (the default data dir) be a symlink, and I'd have the 
scripts change the symlink to point at the profile I wanted.  The only 
difference is that in that case, the symlink could only point to one at a 
time, while with the environmental variable method, it's possible to run 
more than one profile at once, each pointed at a different dir, without 
them stepping on each other.

> To change subjects in mid-stream, I just noticed that K. Haley has added
> the patch to store all downloads as non-executable.  Someone asked about
> that not too long ago but I don't know which thread it was in.

That's good.  I thought he had already added it, but if not, I'm glad he 
has now! =:^)

FWIW, I setup (Gentoo users can find it at the bug I filed for it, using 
J Cuyler's subversion -> git eclass hints) a live pan-9999 ebuild that 
has a USE flag that can be set to select either K Haley's sources, or 
upstream Gnome sources.  I have it set for khaley's, of course.  So I 
should probably reinstall it, thus getting the latest updates...

-- 
Duncan - List replies preferred.   No HTML msgs.
"Every nonfree program has a lord, a master --
and if you use the program, he is your master."  Richard Stallman