lucene-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Greg Colvin <g...@colvin.org>
Subject Re: undefined primitive types
Date Mon, 25 Sep 2006 16:41:10 GMT
On Sep 24, 2006, at 11:44 PM, David Balmain wrote:
> Hi Greg,
>
> I don't know which documentation of the Lucene FileFormat you are
> looking at but you can see UInt32 (Int) UInt64 (Long) and VInt defined
> here:
>
>    http://lucene.apache.org/java/docs/fileformats.html

That's where I was looking.  Nowhere does it say that Int -> UInt32
and Long -> UInt64.  I had guessed from the names that they were
signed, and that Long was 32 bits.  I'm glad to be wrong.

> Are you at liberty to tell us what you are working on?

I'm rewriting the kernel for the Captaris Alchemy product line.
We use Lucene as the document store and index, but have found some
problems that get in the way of fast bulk operations.  The worst
problem is that the Field constructors insist that the entire
contents of a binary field be copied into a byte[], and don't
allow such fields to be indexed.  This assumption that entire
fields must fit into RAM seems to be wired pretty deep into
Lucene.  Similarly, we want to use Windows kernel-mode file
transfer to retrieve documents directly from the file system to
the wire, and not have to read them into user-mode memory first.
So we need to be able to traverse the Lucene data structures
ourselves.

> You may also
> like to take a look at Ferret:
>
>    http://ferret.davebalmain.com/trac
>
> Up to version 0.9.6 it follows the Lucene file format quite closely
> apart from the fact that Ferret can't handly modified UTF-8. Also,
> it's in C, not C++.

I'll look, thanks.

> Cheers,
> Dave
>
> On 9/25/06, Greg Colvin <greg@colvin.org> wrote:
>> Just.  I'll look there, thanks.
>>
>> On Sep 24, 2006, at 10:48 PM, Otis Gospodnetic wrote:
>> > Hi Greg,
>> >
>> > Are you aware of CLucene?
>> >
>> > Otis
>> >
>> > ----- Original Message ----
>> > From: Greg Colvin <greg@colvin.org>
>> > To: java-dev@lucene.apache.org
>> > Sent: Sunday, September 24, 2006 9:25:35 PM
>> > Subject: undefined primitive types
>> >
>> > I'm trying to write C++ code following the Lucene File Formats
>> > document, and find that the terms Int, Long, and VLong are left
>> > undefined.
>> >
>> >
>> >  
>> ---------------------------------------------------------------------
>> > To unsubscribe, e-mail: java-dev-unsubscribe@lucene.apache.org
>> > For additional commands, e-mail: java-dev-help@lucene.apache.org
>> >
>> >
>> >
>> >
>> >
>> >  
>> ---------------------------------------------------------------------
>> > To unsubscribe, e-mail: java-dev-unsubscribe@lucene.apache.org
>> > For additional commands, e-mail: java-dev-help@lucene.apache.org
>> >
>> >
>>
>>
>> ---------------------------------------------------------------------
>> To unsubscribe, e-mail: java-dev-unsubscribe@lucene.apache.org
>> For additional commands, e-mail: java-dev-help@lucene.apache.org
>>
>>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: java-dev-unsubscribe@lucene.apache.org
> For additional commands, e-mail: java-dev-help@lucene.apache.org
>


---------------------------------------------------------------------
To unsubscribe, e-mail: java-dev-unsubscribe@lucene.apache.org
For additional commands, e-mail: java-dev-help@lucene.apache.org


Mime
View raw message