Re: [Evolution-hackers] local addressbook format ...
- From: JP Rosevear <jpr ximian com>
- To: Michael Meeks <michael ximian com>
- Cc: toshok ximian com, evolution <evolution-hackers ximian com>
- Subject: Re: [Evolution-hackers] local addressbook format ...
- Date: Mon, 01 Mar 2004 09:29:53 -0500
On Mon, 2004-02-23 at 15:24, Michael Meeks wrote:
> Hi there,
>
> So - I've recently been the victim of multiple loss of addressbook
> data; over the last ~year evolution has corrupted my data several times
> for no readily comprehensible reason.
>
> Is this something other people have experienced ? if so - is it thought
> to be an artefact of libdb ? and/or would we fare better with a plain
> text / VCARD type format for the address data ? (and/or does the switch
> to 2.0 migrate to a different format here). It seems the e-d-s
> transition gives us a fairly clean break for this sort of thing - but I
> guess the search logic might be too painful to write and using libdb
> makes this substantially easier ?
There is a plain text file backend in CVS but it was decided it couldn't
be finished and tested in time.
> Then again - is it possible it's due to multiple concurrent wombats
> running hitting my address data ? this my happen to me more regularly
> since I tend to killall -9 bonobo-activation-server - which might result
> in new children being spawned - what sort of locking is in-place to
> handle that case ? [ we can work harder in b-a-s to stop that being
> possible if this is the sum of it ].
There is no locking in place to handle this other than the de-facto
locking of using b-a-s, so this is a possible source of the problem.
-JP
--
JP Rosevear <jpr ximian com>
Ximian, Inc.
[
Date Prev][
Date Next] [
Thread Prev][
Thread Next]
[
Thread Index]
[
Date Index]
[
Author Index]