A local web cache in fact is an archived backup from the web already.I would like for WWWOFFLE to do in this case is that after realising that the file is not available in /var/spool/wwwoffle/http/wherever.org/ to look on the CD-ROM and only after this fails as well to allow me to mark for download the next time.
Why making one more copy ? Why not extend the expiry timespan (maybe fine-
tuned on URLs) instead ?
At a first glance, i only can imagine a) reducing lookup time, and
b) limited HD space. It's hard to imagine though, with recent costs nowadays;
and you can clean a cache frequently anyway.
The question to me appears to be, how to reduce lookup time ?
There was a discussion before ( see: Re: Directory sizes [was: WWWOFFLE Version 2.8 Beta],
around 8.9.03).
imho, from the UI point of view, making backups available could be done via the
request-page menu: Adding a choice 'search archive', where archive paths are defined
in the config. Links on that archived page should be looked for in the actual cache, i think,
not in the archive; eventually provoking the same circumstances again.
One could specify an 'first search archive before show request-page' in the config.
(For cd-rom's, wouldn't it have to ask you to insert it, when not already mounted ?)
Probably it would be more convenient to not mix archives with caches anyway,
so archives will be really seperate and static. This means increased redundancy, however.
From the internal point of view, it could be a new directory 'archive' in /var/cache/wwwoffle,
containing symlinks to archive drives. Working only for http would be enough imho.
Sure it's possible to code, but always there's a lot of stuff to do, and the question is,
would this feature be as important as others ? (See 'Web Archive' thread around 14.12.02)
I mean, why not ask on vger.kernel.org to implement it in the kernel ;-)
--
mi.