From clstoreuser at gmail.com  Mon Dec  7 17:12:53 2009
From: clstoreuser at gmail.com (Gary Zu)
Date: Mon, 7 Dec 2009 11:12:53 -0600
Subject: [cl-store-devel] CL-STORE:RESTORE-ERROR - ILLEGAL-KEYWORD-VALUE
Message-ID: <5c9c54140912070912g361fd639uaf711094e67d5f8b@mail.gmail.com>

Dear All,

I have been using cl-store for saving and restoring my objects on Lispworks
1.6.0.01. It has been working great. In my saved object, it has a hash-table
that maps an integer id to a simple object. The saved instance also has
several other modeling objects.

Recently I ran into a restore error during a restoring attempt. This
happened only once and it didn't keep any data/objects before the cl-store
serialization, which means the file saved by cl-store serialization is the
only data I got. So I have no way to check if anything in my data itself was
wrong or not.  Sorry for this incomplete information. Still, could you
please help me on this? Or could you give me some ideas/clues on this error?

The error is:

Error: NIL is an illegal value for the keyword :SIZE
  1 (abort) Return to level 0.
  2 Return to top loop level 0.

Type :b for backtrace or :c <option number> to proceed.
Type :bug-form "<subject>" for a bug report template or :? for other
options.

Call to ERROR
Call to MAKE-HASH-TABLE
Call to (METHOD CL-STORE::INTERNAL-RESTORE-OBJECT (CL-STORE:CL-STORE (EQL
(QUOTE HASH-TABLE)) T))
Call to CL-STORE::HANDLE-NORMAL
Call to CL-STORE::RESTORE-TYPE-OBJECT
Call to CL-STORE::HANDLE-NORMAL
Call to (METHOD CL-STORE:BACKEND-RESTORE (CL-STORE:RESOLVING-BACKEND
STREAM))
Call to CL-STORE::RESTORE-FROM-FILE
Call to (METHOD CL-STORE:RESTORE (T))
Call to CLOS::GENERIC-FUNCTION-NON-DISCRIMINATOR
Call to LET*
Call to EVAL
Call to CAPI::CAPI-TOP-LEVEL-FUNCTION
Call to CAPI::INTERACTIVE-PANE-TOP-LOOP
Call to MP::PROCESS-SG-FUNCTION


Thanks,
a cl-store user
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://mailman.common-lisp.net/pipermail/cl-store-devel/attachments/20091207/b1543fc7/attachment.html>

From senatorzergling at gmail.com  Mon Dec  7 22:38:09 2009
From: senatorzergling at gmail.com (szergling)
Date: Tue, 8 Dec 2009 11:38:09 +1300
Subject: [cl-store-devel] CL-STORE:RESTORE-ERROR - ILLEGAL-KEYWORD-VALUE
In-Reply-To: <5c9c54140912070912g361fd639uaf711094e67d5f8b@mail.gmail.com>
References: <5c9c54140912070912g361fd639uaf711094e67d5f8b@mail.gmail.com>
Message-ID: <8a4b88390912071438x1f761c3cy69dd08811a6321c@mail.gmail.com>

On 12/8/09, Gary Zu <clstoreuser at gmail.com> wrote:
> Dear All,
>
> I have been using cl-store for saving and restoring my objects on Lispworks
> 1.6.0.01. It has been working great. In my saved object, it has a hash-table
> that maps an integer id to a simple object. The saved instance also has
> several other modeling objects.
>
> Recently I ran into a restore error during a restoring attempt. This
> happened only once and it didn't keep any data/objects before the cl-store
> serialization, which means the file saved by cl-store serialization is the
> only data I got. So I have no way to check if anything in my data itself was
> wrong or not.  Sorry for this incomplete information. Still, could you
> please help me on this? Or could you give me some ideas/clues on this error?
>
> The error is:
>
> Error: NIL is an illegal value for the keyword :SIZE
>   1 (abort) Return to level 0.
>   2 Return to top loop level 0.
>
> Type :b for backtrace or :c <option number> to proceed.
> Type :bug-form "<subject>" for a bug report template or :? for other
> options.
>
> Call to ERROR
> Call to MAKE-HASH-TABLE
> Call to (METHOD CL-STORE::INTERNAL-RESTORE-OBJECT (CL-STORE:CL-STORE (EQL
> (QUOTE HASH-TABLE)) T))
> Call to CL-STORE::HANDLE-NORMAL
> Call to CL-STORE::RESTORE-TYPE-OBJECT
> Call to CL-STORE::HANDLE-NORMAL
> Call to (METHOD CL-STORE:BACKEND-RESTORE (CL-STORE:RESOLVING-BACKEND
> STREAM))
> Call to CL-STORE::RESTORE-FROM-FILE
> Call to (METHOD CL-STORE:RESTORE (T))
> Call to CLOS::GENERIC-FUNCTION-NON-DISCRIMINATOR
> Call to LET*
> Call to EVAL
> Call to CAPI::CAPI-TOP-LEVEL-FUNCTION
> Call to CAPI::INTERACTIVE-PANE-TOP-LOOP
> Call to MP::PROCESS-SG-FUNCTION
>
>
> Thanks,
> a cl-store user
>

Hi Gary, just a quick question.

It looks like you are trying to restore a hash-table whose size is
nil. Does (hash-table-size <any-hash-table>) ever return nil (that's
what gets stored previously, see the defstore-cl-store code for
hash-tables)?  CLHS says that SIZE is a non-negative integer. Any
ideas how this could happen?  Are you hash-tables ever empty?

If this is a Lispworks bug, perhaps you could redefine the restorer to
default to :size 0 when this happens? (just a workaround)

Yong.



From clstoreuser at gmail.com  Tue Dec  8 01:25:56 2009
From: clstoreuser at gmail.com (Gary Zu)
Date: Mon, 7 Dec 2009 19:25:56 -0600
Subject: [cl-store-devel] CL-STORE:RESTORE-ERROR - ILLEGAL-KEYWORD-VALUE
In-Reply-To: <8a4b88390912071438x1f761c3cy69dd08811a6321c@mail.gmail.com>
References: <5c9c54140912070912g361fd639uaf711094e67d5f8b@mail.gmail.com>
	<8a4b88390912071438x1f761c3cy69dd08811a6321c@mail.gmail.com>
Message-ID: <5c9c54140912071725k195b3375x917c652e1b971f34@mail.gmail.com>

On Mon, Dec 7, 2009 at 4:38 PM, szergling <senatorzergling at gmail.com> wrote:

> >Hi Gary, just a quick question.
> >
> >It looks like you are trying to restore a hash-table whose size is
> >nil. Does (hash-table-size <any-hash-table>) ever return nil (that's
> >what gets stored previously, see the defstore-cl-store code for
> >hash-tables)?  CLHS says that SIZE is a non-negative integer. Any
> >ideas how this could happen?  Are you hash-tables ever empty?
> >
> >If this is a Lispworks bug, perhaps you could redefine the restorer to
> >default to :size 0 when this happens? (just a workaround)
> >
> >Yong.
>

Thanks! I have no idea how the (hash-table-size <any-hash-table>) ever
returned nil in Lispworks (1.6.0.01). Besides size, I logged out all
parameters restored for the saved hash table. They read as follows:

[DEBUG] : rehash-size = NIL
[DEBUG] : rehash-threshold = 7
[DEBUG] : size = NIL
[DEBUG] : test = EQL
[DEBUG] : count = 601

rehash-size and rehash-threshold are both wrong, too. If I make an empty
hash table, the Lispworks gives size=37, rehash-size=2.0,
rehash-threshold=1.0 - surely they got them all right.

You can see that my hash table got 601 entries. I indeed got them all
restored using your workaround idea. Still not sure how those bad parameter
values got stored in the first place.

Pointing me to the source code is a great help. Many thanks again!

Gary
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://mailman.common-lisp.net/pipermail/cl-store-devel/attachments/20091207/11dbabcb/attachment.html>

From senatorzergling at gmail.com  Tue Dec  8 05:10:39 2009
From: senatorzergling at gmail.com (szergling)
Date: Tue, 8 Dec 2009 18:10:39 +1300
Subject: [cl-store-devel] CL-STORE:RESTORE-ERROR - ILLEGAL-KEYWORD-VALUE
In-Reply-To: <5c9c54140912071725k195b3375x917c652e1b971f34@mail.gmail.com>
References: <5c9c54140912070912g361fd639uaf711094e67d5f8b@mail.gmail.com>
	<8a4b88390912071438x1f761c3cy69dd08811a6321c@mail.gmail.com>
	<5c9c54140912071725k195b3375x917c652e1b971f34@mail.gmail.com>
Message-ID: <8a4b88390912072110j136d809ap214075d1655008a3@mail.gmail.com>

On 12/8/09, Gary Zu <clstoreuser at gmail.com> wrote:

> returned nil in Lispworks (1.6.0.01). Besides size, I logged out all
> parameters restored for the saved hash table. They read as follows:
>
> [DEBUG] : rehash-size = NIL
> [DEBUG] : rehash-threshold = 7
> [DEBUG] : size = NIL
> [DEBUG] : test = EQL
> [DEBUG] : count = 601
>
> rehash-size and rehash-threshold are both wrong, too. If I make an empty
> hash table, the Lispworks gives size=37, rehash-size=2.0,
> rehash-threshold=1.0 - surely they got them all right.
>
> You can see that my hash table got 601 entries. I indeed got them all
> restored using your workaround idea. Still not sure how those bad parameter
> values got stored in the first place.

Either the hash-table-* functions or some hash-table were acting funny,
or your store may have been corrupted. I don't know how data
could be corrupted, and I would assume that any cl-store version
incompatibilities would be picked up by asserts. This is still a little
suspicious. Does the restored hash-table's content look intact?
Not sure how else to go about this...

> Pointing me to the source code is a great help. Many thanks again!

You're welcome!

Yong.



From clstoreuser at gmail.com  Thu Dec 10 17:02:24 2009
From: clstoreuser at gmail.com (Gary Zu)
Date: Thu, 10 Dec 2009 11:02:24 -0600
Subject: [cl-store-devel] CL-STORE:RESTORE-ERROR - ILLEGAL-KEYWORD-VALUE
In-Reply-To: <8a4b88390912072110j136d809ap214075d1655008a3@mail.gmail.com>
References: <5c9c54140912070912g361fd639uaf711094e67d5f8b@mail.gmail.com>
	<8a4b88390912071438x1f761c3cy69dd08811a6321c@mail.gmail.com>
	<5c9c54140912071725k195b3375x917c652e1b971f34@mail.gmail.com>
	<8a4b88390912072110j136d809ap214075d1655008a3@mail.gmail.com>
Message-ID: <5c9c54140912100902i6def0dfy83870e29b272227b@mail.gmail.com>

I have taken 2 days off my work and jut got back.

On Mon, Dec 7, 2009 at 11:10 PM, szergling <senatorzergling at gmail.com>wrote:

> On 12/8/09, Gary Zu <clstoreuser at gmail.com> wrote:
> >> returned nil in Lispworks (1.6.0.01). Besides size, I logged out all
> >> parameters restored for the saved hash table. They read as follows:
> >>
> >> [DEBUG] : rehash-size = NIL
> >> [DEBUG] : rehash-threshold = 7
> >> [DEBUG] : size = NIL
> >> [DEBUG] : test = EQL
> >> [DEBUG] : count = 601
> >>
> >> rehash-size and rehash-threshold are both wrong, too. If I make an empty
> >> hash table, the Lispworks gives size=37, rehash-size=2.0,
> >> rehash-threshold=1.0 - surely they got them all right.
> >>
> >> You can see that my hash table got 601 entries. I indeed got them all
> >> restored using your workaround idea. Still not sure how those bad
> parameter
> >> values got stored in the first place.
> >
> >Either the hash-table-* functions or some hash-table were acting funny,
> >or your store may have been corrupted. I don't know how data
> >could be corrupted, and I would assume that any cl-store version
> >incompatibilities would be picked up by asserts. This is still a little
> >suspicious. Does the restored hash-table's content look intact?
> >Not sure how else to go about this...
>
>  Yes, the restored hash-table works just fine as I can tell.

Thanks!
-Gary
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://mailman.common-lisp.net/pipermail/cl-store-devel/attachments/20091210/c9953dbf/attachment.html>