On 14.11.2007, at 17.20, Adam McDougall wrote:
Thanks for this list of steps, I've been intending to test it and
was just about
getting ready to ask the same question. Your email would be mice
content to
throw on the dovecot wiki under fts (currently empty)
I wrote something there now. Al
On Wed, 14 Nov 2007, Adam McDougall wrote:
Thanks for this list of steps, I've been intending to test it and was
just about getting ready to ask the same question. Your email would be
mice content to throw on the dovecot wiki under fts (currently empty)
Good point, but you do it, it's bedtim
On Thu, Nov 15, 2007 at 12:07:15AM +0900, Asheesh Laroia wrote:
On Wed, 14 Nov 2007, Daniel Watts wrote:
>> Timo - we were just having a conversation about how we might be able to
>> provide full body indexed search for our clients and I realised it might
>> be worth checking the Dovecot li
On Wed, 14 Nov 2007, Daniel Watts wrote:
Timo - we were just having a conversation about how we might be able to
provide full body indexed search for our clients and I realised it might
be worth checking the Dovecot list to see if this has been done
already...
And then I find this thread!
W
Timo Sirainen wrote:
On Fri, 2007-04-06 at 00:34 +0300, Timo Sirainen wrote:
Squat-like 4 byte substrings (but can answer 1-3 char queries also):
Indexing a 1,4GB Linux kernel mailing list mbox with 367919 messages:
UID count: 367919
Index time: 129.86 CPU seconds (10.43MB/CPUs), 132.47 secon
On Fri, 13 Apr 2007 14:07:58 +0300
Timo Sirainen <[EMAIL PROTECTED]> wrote:
> gzip compression makes the uidlist still 25% smaller (total space
> 19,50%). It'd have to be used to compress the file in smaller blocks
> because zlib doesn't support quickly seeking inside the file. That would
> probab
On Fri, 2007-04-06 at 00:34 +0300, Timo Sirainen wrote:
> Squat-like 4 byte substrings (but can answer 1-3 char queries also):
Indexing a 1,4GB Linux kernel mailing list mbox with 367919 messages:
UID count: 367919
Index time: 129.86 CPU seconds (10.43MB/CPUs), 132.47 seconds
(10.23MB/s)
Memory:
Timo Sirainen wrote:
As described earlier
(http://dovecot.org/list/dovecot/2006-December/018055.html), Dovecot
nowadays has full text search indexing support in CVS HEAD.
So it takes somewhat more space, but definitely less than having both
Squat + Lucene.
No substring indexing, words up
On Fri, 2007-04-06 at 10:42 +0200, DINH Viêt Hoà wrote:
> An other problem with squat is that we can't remove items from the
> index. (the version of Cyrus). Is that still the case ?
No. Dovecot's Squat is almost completely different from Cyrus. I just
kept the name because the basic ideas are the
On 4/5/07, Timo Sirainen <[EMAIL PROTECTED]> wrote:
As described earlier
(http://dovecot.org/list/dovecot/2006-December/018055.html), Dovecot
nowadays has full text search indexing support in CVS HEAD.
Currently there are two backends: Lucene and Squat. Lucene's problem is
that standard IMAP SEA
As described earlier
(http://dovecot.org/list/dovecot/2006-December/018055.html), Dovecot
nowadays has full text search indexing support in CVS HEAD.
Currently there are two backends: Lucene and Squat. Lucene's problem is
that standard IMAP SEARCH command can't be used with it without breaking
IMA
11 matches
Mail list logo
