Automatic Saving

[Derek Atkins]

I'm not convinced it should be changed.  I like the fact that I can
"exit without saving" and not have any data disrupted.  OTOH, I can
understand that people DO want to be able to have "interim saves" in
case of a crash, in which case the user should probably be prompted
whether to use the 'last-saved' or 'auto-saved' data.

-derek

linas@linas.org (Linas Vepstas) writes:

> On Fri, Jan 03, 2003 at 09:51:37AM -0800, W Pennington was heard to remark:
> > GnuCash 1.6.6
> > 
> > Is there a feature for saving changes every ___ minutes?
> 
> Not currently.
> 
> > If not, is there an appropriate forum for feature requests?
> 
> gnucash-devel mailing list is better.
> 
> ---------
> I suppose the time has come for something like this.  I just implemented
> something like this for GnoTime (gttr.sourceforge.net)
> 
> It saves once a minute for the first 4 minutes, then once every 4
> minutes for the first 16 minutes, then once every 16 minutes for 
> the first 64 minutes, etc., forever.  There is nothing to adjust,
> there is nothing to configure, and the old copies are pruned
> automatically.  Because of the logarithmic distribution, it means that
> even if you run for years, you won't have more than 20 or so backup
> copies (viz. every 4 months for the first 16 months, etc).
> 
> The algorithm seems safe: the code never actually deletes files,
> it only renames them: The pruning works by copying a newer file on top
> of an older file.  Thus, even if the app goes totally wacko, it can't
> wack much: there's no chance of an 'rm *' happening by accident.
> 
> I think something like this would defacto server the (vast?) majority
> of GnuCash users.   What I don't know is if there is anybody who thinks
> the current scheme is better?
> 
> BTW the current scheme was designed so that it did the following:
> a) Avoid accidental 'rm -r' by never actually rm'ing
> b) backup filename has datestamp, pid and ipaddr in it, to avoid 
>    accidental over-write by same user, by other users logged onto
>    same machine, or same/other users working over NFS.
> 
>    (By contrast, the new proposed scheme intentionally over-writes
>    as a means of pruning back old copies.)
> c) Saves are done by user, so that if user really screws something up
>    (e.g. deleting an account tree) they can get out of it by not saving.
> 
> 
> So -- should the old scheme be replaced by the new scheme?
> 
> --linas
> 
> 
> -- 
> pub  1024D/01045933 2001-02-01 Linas Vepstas (Labas!) <linas@linas.org>
> PGP Key fingerprint = 8305 2521 6000 0B5E 8984  3F54 64A9 9A82 0104 5933
> _______________________________________________
> gnucash-devel mailing list
> gnucash-devel@lists.gnucash.org
> https://lists.gnucash.org/mailman/listinfo/gnucash-devel

-- 
       Derek Atkins, SB '93 MIT EE, SM '95 MIT Media Laboratory
       Member, MIT Student Information Processing Board  (SIPB)
       URL: http://web.mit.edu/warlord/    PP-ASEL-IA     N1NWH
       warlord@MIT.EDU                        PGP key available