[elephant-devel] btree slot of pset becomes unbound when re-opening store
Kevin Raison
raison at chatsubo.net
Wed Oct 31 00:19:28 UTC 2007
Has any progress been made on this SBCL / pset issue? I would like to
help resolve this, as I am suffering from this problem too. If someone
could describe the specifics of the problem between SBCL and elephant, I
will take a closer look.
Thanks.
Kevin
On Tue Jun 12 10:37:09 EDT 2007, Ian Eslick wrote:
Ok, please ignore my prior e-mails and fixes. There is a definite
problem deserializing psets that only shows up in SBCL.
I'll get back to y'all.
Ian
On Jun 12, 2007, at 10:20 AM, Ian Eslick wrote:
> Actually, I now remember that I had intended to implement psets as
> ordinary classes to avoid allocating two persistent objects as the
> patch now does. The actual bug was not remembering the oid of the
> btree, but the new fix is probably cleaner anyway.
>
> Ian
>
> On Jun 12, 2007, at 10:18 AM, Ian Eslick wrote:
>
>> Well, that was a silly bug. I just checked in a patch. Please
>> let me know if it works for you now.
>>
>> Ian
>>
>> On Jun 8, 2007, at 8:49 PM, Brian Seitz wrote:
>>
>>> Hey,
>>>
>>> I've been experimenting with Elephant, and it's been really great
>>> so far.
>>>
>>> But I've had one strange issue. When I make a pset, its btree slot
>>> always becomes unbound after I've closed and re-opened the store.
>>> I've tried this with bdb, sqlite, and postgresql, all with the same
>>> result. Using Elephant 0.9 and SBCL 1.0.
>>>
>>> Here's a self contained instance (using the "friends" example
>>> from the docs):
>>>
>>> CL-USER> (require 'elephant)
>>> NIL
>>> CL-USER> (in-package :ele)
>>> #<PACKAGE "ELEPHANT">
>>> ELE> (open-store '(:bdb "/home/bseitz/eletest/"))
>>> #<BDB-STORE-CONTROLLER /home/bseitz/eletest/>
>>> ELE> (defpclass person ()
>>> ((name :accessor person-name :initarg :name)
>>> (friends :accessor person-friends :initarg :friends
>>> :initform (make-pset))))
>>> #<PERSISTENT-METACLASS PERSON>
>>> ELE> (defmethod add-friend ((me person) (them person))
>>> (insert-item them (person-friends me)))
>>> #<STANDARD-METHOD ADD-FRIEND (PERSON PERSON) {B24FC59}>
>>> ELE> (defmethod remove-friend ((me person) (them person))
>>> (remove-item them (person-friends me)))
>>> #<STANDARD-METHOD REMOVE-FRIEND (PERSON PERSON) {B288FD9}>
>>> ELE> (defmethod map-friends (fn (me person))
>>> (map-pset fn (person-friends me)))
>>> #<STANDARD-METHOD MAP-FRIENDS (T PERSON) {B2BA3C1}>
>>> ELE> (setf me (make-instance 'person :name "Brian"))
>>> #<PERSON oid:100>
>>> ELE> (setf amigo (make-instance 'person :name "Waldo"))
>>> #<PERSON oid:102>
>>> ELE> (add-friend me amigo)
>>> #<PERSON oid:102>
>>> ELE> (map-friends #'print me)
>>>
>>> #<PERSON oid:102>
>>> #<DEFAULT-PSET oid:101>
>>> ELE> (add-to-root "me" me)
>>> #<PERSON oid:100>
>>> ELE> (add-to-root "amigo" amigo)
>>> #<PERSON oid:102>
>>> ELE> (close-store)
>>> NIL
>>> ELE> (setf me nil)
>>> NIL
>>> ELE> (setf amigo nil)
>>> NIL
>>> ELE> (open-store '(:bdb "/home/bseitz/eletest/"))
>>> #<BDB-STORE-CONTROLLER /home/bseitz/eletest/>
>>> ELE> (setf me (get-from-root "me"))
>>> #<PERSON oid:100>
>>> ELE> (setf amigo (get-from-root "amigo"))
>>> #<PERSON oid:102>
>>> ELE> (describe me)
>>> #<PERSON oid:100> is an instance of class #<PERSISTENT-METACLASS
>>> PERSON>.
>>> The following slots have :INSTANCE allocation:
>>> %OID 100
>>> DBCONNECTION-SPEC-PST (:BDB "/home/bseitz/eletest/")
>>> The following slots have allocation as shown:
>>> NAME :DATABASE "Brian"
>>> FRIENDS :DATABASE #<DEFAULT-PSET oid:101>
>>> ; No value
>>> ELE> (map-friends #'print me)
>>> The slot BTREE is unbound in the object #<DEFAULT-PSET oid:101>.
>>> ; Evaluation aborted
>>> ELE> (describe (person-friends me))
>>> #<DEFAULT-PSET oid:101>
>>> is an instance of class #<STANDARD-CLASS DEFAULT-PSET>.
>>> The following slots have :INSTANCE allocation:
>>> %OID 101
>>> DBCONNECTION-SPEC-PST (:BDB "/home/bseitz/eletest/")
>>> BTREE #<unbound slot>
>>>
>>> So the BTREE that indexes (?) the persistent set is now gone? I've
>>> tried to make the slot that holds the pset indexed, but that doesn't
>>> seem to make a difference.
>>>
>>> I greatly appreciate any help you could give.
>>>
>>> Thanks,
>>>
>>> Brian
>>> _______________________________________________
>>> elephant-devel site list
>>> elephant-devel at common-lisp.net
>>> http://common-lisp.net/mailman/listinfo/elephant-devel
>>
>> _______________________________________________
>> elephant-devel site list
>> elephant-devel at common-lisp.net
>> http://common-lisp.net/mailman/listinfo/elephant-devel
>
> _______________________________________________
> elephant-devel site list
> elephant-devel at common-lisp.net
> http://common-lisp.net/mailman/listinfo/elephant-devel
More information about the elephant-devel
mailing list