The branch now passes 97% of the tests. Failures all revolve around
interactions between schema evolution (change-class, redefine-class,
etc) and indexing which hasn't been implemented yet. No tests for the
new stuff yet, of course, but it's in good enough shape to play with.
Specificall
Robert makes an excellent point. For datasets that fit in memory,
caching objects and slot values in memory makes the use of lisp as a
query language really easy.
Another (unreleased) prevalence-like facility in Elephant:
In src/contrib/eslick/snapshot-set.lisp is a simple object caching
On Thu, 2008-03-06 at 10:10 -0500, Ian Eslick wrote:
> I agree with Robert. The best way to start is to use lisp as a
> query
> language and essential do a search/match over the object graph.
>
> The rub comes when you start looking at performance. A linear scan
> of
I neglected to mention
You shouldn't be using any of the controllers in contrib/eslick, those
are in-progress experiments. Which version of the source tree are you
using? This code shouldn't be loaded unless you are doing it
explicitly. You should just be loading elephant.asd.
Ian
On Mar 6, 2008, at 12:12 PM,
Alex Mizrahi writes:
> LPP> I should have added that I'm talking about a schema evolution thing
> LPP> here, i.e. the situation concerns slots *added* to a class.
>
> LPP> The slots of instances of such a class are unbound instead of being
> LPP> assigned their initform, if any.
>
> i agree with y
I agree with Robert. The best way to start is to use lisp as a query
language and essential do a search/match over the object graph.
The rub comes when you start looking at performance. A linear scan of
alot of class instances can become prohibitive, even for small set
sizes, especially s
??>> can you please describe changes you've made so i will not occasionally
??>> omit them?
LPP> Txn changes:
LPP> 1) Deadlock handling: when a deadlock is detected, one of the txns
involved
LPP>will be aborted. Change: catch error and retry the txn instead.
LPP> 2) Honor the RETRIES
> can you please describe changes you've made so i will not occasionally omit
> them?
Txn changes:
1) Deadlock handling: when a deadlock is detected, one of the txns involved
will be aborted. Change: catch error and retry the txn instead.
2) Honor the RETRIES parameter for txn restarts.
3)
Thanks! Rock on!
I will test them this weekend.
On Thu, 2008-03-06 at 14:16 +0100, Leslie P. Polzer wrote:
> Hello everyone,
>
> I wrote two concurreny test cases, added Bordeaux Threads and revised
> the Postmodern error and transaction handling (for example, deadlocks
> are no longer a proble
LPP> The slots of instances of such a class are unbound instead of
being
LPP> assigned their initform, if any.
That was a strange phenomenon that I couldn't reproduce. However,
let's keep an eye on this and revisit it (perhaps a Trac ticket
against 0.9.2). I've messed with object init
On Wed, 2008-03-05 at 21:50 -0800, Paul Legato wrote:
> I think the thing is not that we need professional consulting, but we
> are trying to become the professional consultants. :) Could you more
> knowledgable ODBMS coders recommend some good books or papers on the
> topic of ODBMS care and feedi
LPP> I should have added that I'm talking about a schema evolution thing
LPP> here, i.e. the situation concerns slots *added* to a class.
LPP> The slots of instances of such a class are unbound instead of being
LPP> assigned their initform, if any.
i agree with you -- this unbound thing is totall
LPP> I wrote two concurreny test cases, added Bordeaux Threads and revised
LPP> the Postmodern error and transaction handling (for example, deadlocks
LPP> are no longer a problem).
ouch, i was working on transaction handling myself, so now we have a
conflict..
can you please describe changes
On Mar 4, 2008, at 10:38 PM, Ian Eslick wrote:
= A little query interpreter (Waldo?, Daniel?)
Read sets of oids from indices, implement an efficient in-memory
sort&merge over oids, and deserialize after the query. I've
specified a bunch of this so could provide a set of pointers to
star
Hello everyone,
I wrote two concurreny test cases, added Bordeaux Threads and revised
the Postmodern error and transaction handling (for example, deadlocks
are no longer a problem).
Also included is the patch to close dangling connections I sent about
two weeks ago.
Please test them if you can s
15 matches
Mail list logo
