Re: [Evolution] POP3 mbox performance and auto-expunging?

[Patrick O'Callaghan <poc usb ve>]

On Fri, 2008-08-22 at 15:51 +0200, Milan Crha wrote:
> On Fri, 2008-08-22 at 08:49 -0430, Patrick O'Callaghan wrote:
> > So the stored summary implicitly assumes the off_t size? I have to say
> > that I had no problems moving from a 32-bit to a 64-bit system and
> > preserving all my local mail, so I'm not sure there's really an issue
> > here (going the other way could be a different story of course).
>
> Yes. As far as I can tell. Maybe I'm wrong, it really depends on the
> time of the move (whether it was also between versions and some
> migration has been involved automatically, then it can sort of work
> without noticing).

It was from Fedora 8 to Fedora 9, so there would have been a version
change as well.

> The only impact, the different size of off_t, is that
> the summary will not match the mbox file, which could result in summary
> mismatch and regeneration of the summary (with "native" off_t size for
> the actual system).

I suppose that's possible. So in fact the safe way is just to delete the
summary files and let Evo re-create them :-)

> I hope it makes sense a bit what I want to tell, and that it could make
> trouble, but if you've other real experience with different behavior,
> then it's possible I'm wrong here. Hmm, I'm really not sure now... :)

It's hard to be sure of these things without looking at the source. The
summary files are binary data and I don't know if the format is
documented outside the source. In fact I don't even know if the
summaries are portable between big-endian and little-endian machines.

poc