Re: [Pan-users] Oh, for draft folder!

[Duncan]

Steven D'Aprano posted on Sun, 26 Jun 2011 14:20:39 +1000 as excerpted:

>> I suppose one way around that would be a simple post-slot based ring-
>> scheme.  Choose some arbitrary number of auto-save messages, and when
>> pan reaches that limit, it simply deletes the oldest to make room for
>> the next one.  If the number is say 100 messages, then message 101 will
>> replace message 1.
> 
> /me tries to stay calm.
> 
> Please. Do. Not. Silently. Delete. My. Data.
> 
> EVER.

That's why I suggested making the size of that ring-buffer a config 
option, possibly exposed only in the config file itself (not in the GUI), 
thus keeping the user settings simple, but exposing the option at least 
in the config file, for those who want to set it.  Set it to something 
like a million (or 4.294+ billion, the 32-bit unsigned long limit), and 
it never deletes.  Set it to 100 or so, and it automatically controls the 
space it requires by ring-buffering.  Set it to 0 and it disables that 
functionality.

I proposed something like 100 be the default, simply to keep pan's 
storage needs from ever expanding in much the same way the 10 MB cache 
size default does, however unreasonable I find it for my own use, but at 
least the value is available to tweak in the config file for those who 
wish to. =:^)

OTOH, as a KDE user I'd be more than happy to have all those config-file-
only settings exposed in the GUI, but with gnome-3 going even more config-
nazi than gnome-2, and with pan being at least in name a gnome app even 
if it's really only gtk dependent...

Anyway, I don't disagree with anything in your post.  It's just that I 
was looking at it from the perspective of least change from current to 
get the named feature.  Yes, having a proper sent-posts feature that 
didn't interfere with downloads (two separate concepts, I never did 
understand why Charles couldn't simply implement them that way, but then 
it's always easy to call the shots when you're not doing the coding) 
would be quite useful, even arguably a basic feature, that pan's missing 
now.  And similarly with auto-save.  But I was just thinking in terms of 
least change from current to get the auto-save, and how it might be 
implemented, with reference to the past implementation and the problems 
it had.

-- 
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