On Sun, Nov 24, 2013 at 5:06 PM, Genghis Khan <genghiskhan@gmx.ca> wrote:
On Sat, 23 Nov 2013 23:08:38 +0100
Lars Windolf <lars.lindner@gmail.com> wrote:

> Am 23.11.2013 14:53, schrieb Genghis Khan:
> > Hello,
> >
> >
> > After release 1.8.2 - "Be lazy with cleaning...", Liferea has
> > stated up for me after ~5 seconds[0]. This has been changed between
> > releases 1.8.2 to 1.8.15 - I am unable to be more precise due to
> > this error:
> >
> > ERROR **: Fatal: The cache database was created by a newer
> > version of Liferea than this one! Trace/breakpoint trap
> >
> > If acceptable and reasonable, please add a command line argument to
> > skip on preforming VACUUM on startup. Would such feature cause to
> > WAL file growing issue[1] which was fixed in release 1.8.6[2]?
> >
> > [0]: http://sourceforge.net/mailarchive/forum.php?forum_name=liferea-devel&max_rows=25&style=ultimate&viewmonth=201203&viewday=19
> > [1]: http://liferea.blogspot.com/2012/06/every-growing-db-files.html
> > [2]: http://sourceforge.net/mailarchive/forum.php?forum_name=liferea-devel&max_rows=25&style=nested&viewmonth=201206&viewday=15
> >
> > *
> > http://liferea.blogspot.com/2012/03/182-vacuum-only-when-needed.html
>
> Hi Genghis!
>
> If you want to contribute a patch for a command line option
> please do so, I'll merge it. I do not think it is a common problem
> because you shouldn't have to VACUUM on each startup. That's
> strange. It might be worth to have a look at the output of
>
> liferea --debug-db | grep VACUUM
>
> To see the computed fragmentation rate.
>

Dear Lars,

I wish I would know how to write a patch for relevant issue.

Running Liferea 1.8.15 (GTK+2).

$ liferea --debug-db | grep VACUUM
DB: No VACUUM as freelist count/page count ratio 1.69 <= 10


Given this output no vacuum will happen. So I'd say the slowness you perceive is caused by something else...

A run with "--debug-performance" might help identifying would takes long.

Cheers,
Lars