Re: Should loading a shared object try to fetch? (was: Should ERXEC get sharedEC automagically?)
Chuck,
> On 22 Aug 2018, at 5:03 AM, Chuck Hill wrote:
> Assuming that we are looking at the same version of ERXGenericRecord
Probably we are: I have downloaded Wonder-Applications.tar.gz just yesterday,
to replace v6 which I have been using before. The ERExtension framework's
Info.plist says “7.1-SNAPSHOT”. Is there a build # or a build time somewhere in
the frameworks to check?
The main problem was that yesternight, I have been a bit at the sleepy side,
and I have pursued a wrong exception (caused, it seems, by a bug in my catch
block); terribly sorry I have bothered you with that, too :/
What really happens there is that faultWithPrimaryKeyValue — the first time it
is called, which also is the first time the DB is accessed at all — causes the
SEC to initialise, load the shared objects... and then throws
java.lang.IllegalArgumentException: The shared context recently initialized the
object which is already registered in this context.
Objects must be unique in both contexts.
(full trace for reference below, caused directly by WO code, alas, so we can't
check the source).
Far as I understand this, I sort of suspect with faultWithPrimaryKeyValue the
following might be what happens:
(i) ERXEC.faultWithPrimaryKeyValue somehow starts with creating a globalID for
entity/PK and registering it itself (even though the ERXEC has a SEC)
(ii) further processing on-demand creates the DB stack, which creates the SEC
yadda yadda...
(iii) ... and SEC dutifully loads the shared objects
(iv) and, the inconsistence betw. a shared object just loaded into SEC and the
registered gid of (i) is found and throws.
To make sure it is not caused by the fact that at the moment of the first
faultWithPrimaryKeyValue the ERXEC has not a SEC yet (for no DB operation has
been performed yet and thus
EODefaultSharedEditingContextWasInitializedNotification did not come yet), I
tried to add
===
if (!ec.sharedEditingContext)
ec.sharedEditingContext=EOSharedEditingContext.defaultSharedEditingContext()
object=EOUtilities.faultWithPrimaryKeyValue(ec,ename,pk)
===
but that did not help. Then, recalling with objectWithPrimaryKeyValue it did
work (but for the superfluous fetch), I have tried
===
if (!ec.sharedEditingContext)
object=EOUtilities.objectWithPrimaryKeyValue(ec,ename,pk)
else
object=EOUtilities.faultWithPrimaryKeyValue(ec,ename,pk)
===
and this did help: the first time, objectWithPrimaryKeyValue is called, causes
the SEC to initialise and load shared EOs without any problem (but for, of
course, the superfluous harmless SELECT being sent to DB). From the time on,
faultWithPrimaryKeyValue is being used, and it works as expected — no fetch,
shared EOs returned directly.
Well work it does, but isn't that a super-ugly contraption? :(
Incidentally, since yesternight I have added awakeFromInsertion logging, I have
found another suspicious thing: for the shared EOs (loaded as a side-effect of
the first objectWithPrimaryKeyValue call) I do not get awakeFromInsertion at
all! (Before I haven't noticed, for I did not log it.) The sharedEC loads all
the shared objects all right, but never awakeFromInsertion comes. Is this the
intended behaviour?
Aside of that, quite unimportant followup:
> the NPE is from here:
>
> EOGlobalID gid =
> editingContext.globalIDForObject(this);
> if (gid.isTemporary()) {
>
> That suggests that editingContext != this.editingContext()
Just before super.awakeFromInsertion, this.EC has been null; that caused the
problem.
> In other words, that the object is registered in an editing context that is
> different from the parameter to the awakeFromInsertion method. Can you check
> what ECs those are? I am not sure how you would get them to be different…
The mess has been caused by a bug in my exception handler, which I have alas
overlooked — before trying faultWithPrimaryKeyValue the handler has never been
used :)
> Also, I see two awakeFromInsertion methods in your code:
> at
> cz.ocs.model.OCSEnterpriseObject.super$4$awakeFromInsertion(OCSEnterpriseObject.groovy)
> and
> at
> cz.ocs.model.OCSEnterpriseObject.awakeFromInsertion(OCSEnterpriseObject.groovy:265)
> // [*]
>
> Or that just an artifact from using Groovy?
Yup, for some reason, the thing inserts its own stub for each supercall.
Thanks and all the best,
OC
Here's the full trace of the real exception:
===
java.lang.IllegalArgumentException: The shared context recently initialized the
object which is already registered in this context.
Objects must be unique in both contexts.
at
com.webobjects.eocontrol.EOEditingContext._processInitializedObjectsInSharedContext(EOEditingContext.java:2602)
at
java.base/jdk.internal.reflect.NativeMethodAccessorImpl.invoke0(Native Method)
at
java.base/jdk.internal.reflect.NativeMethodAccessorImpl.invoke(NativeMethodAccessorImpl.java:62)
at
Re: Should loading a shared object try to fetch? (was: Should ERXEC get sharedEC automagically?)
Assuming that we are looking at the same version of ERXGenericRecord, the NPE
is from here:
EOGlobalID gid =
editingContext.globalIDForObject(this);
if (gid.isTemporary()) {
That suggests that editingContext != this.editingContext()
In other words, that the object is registered in an editing context that is
different from the parameter to the awakeFromInsertion method. Can you check
what ECs those are? I am not sure how you would get them to be different…
Also, I see two awakeFromInsertion methods in your code:
at
cz.ocs.model.OCSEnterpriseObject.super$4$awakeFromInsertion(OCSEnterpriseObject.groovy)
and
at
cz.ocs.model.OCSEnterpriseObject.awakeFromInsertion(OCSEnterpriseObject.groovy:265)
// [*]
Or that just an artifact from using Groovy?
Chuck
From: "ocs@ocs"
Date: Tuesday, August 21, 2018 at 7:05 PM
To: Chuck Hill
Cc: "webobjects-dev@lists.apple.com"
Subject: Re: Should loading a shared object try to fetch? (was: Should ERXEC
get sharedEC automagically?)
Chuck,
thanks again! I did not know that (well one could write a much bigger book than
you did with just those things I do not know...)
Nevertheless, there still must be some ugly fault of mine. When using
objectWithPrimaryKeyValue, there's the superfluous fetch, but it works.
When replaced by faultWithPrimaryKeyValue, I keep getting NPEs somewhere inside
of awakeFromInsertion — at the moment, I can't really make sense of it :( It
looks like this:
===
== about to load 110 from 348ad293 (SEC ) // first time, there's yet no SEC...
4105 [main] DEBUG NSLog - Using JDBCPlugIn
'com.webobjects.jdbcadaptor.FrontbasePlugIn' for JDBCAdaptor@2076611420
... and it is being initialised here; it fetches properly all the shared EOs ...
4110 [main] DEBUG NSLog - connecting with dictionary: {username = "";
password = ""; URL =
"jdbc:FrontBase://localhost/SBERDAT3/user=FINServis/isolation=read_committed/locking=optimistic";
}
... ... ... the proper SQL to load all the shared objects here ... ... ...
4235 [main] DEBUG NSLog - === Commit Internal Transaction
... and here's a problem; my overridden awakeFromInsertion logs this just
before a super.awakeFromInsertion:
awakeFromInsertion: in
er.extensions.eof.ERXEC@348ad293
... and a log _after_ super does not happen; instead, I get this:
java.lang.Exception:
... ... ...
Caused by: java.lang.NullPointerException
at
er.extensions.eof.ERXGenericRecord.awakeFromInsertion(ERXGenericRecord.java:512)
at
cz.ocs.model.OCSEnterpriseObject.super$4$awakeFromInsertion(OCSEnterpriseObject.groovy)
at java.base/jdk.internal.reflect.NativeMethodAccessorImpl.invoke0(Native
Method)
at
java.base/jdk.internal.reflect.NativeMethodAccessorImpl.invoke(NativeMethodAccessorImpl.java:62)
at
java.base/jdk.internal.reflect.DelegatingMethodAccessorImpl.invoke(DelegatingMethodAccessorImpl.java:43)
at java.base/java.lang.reflect.Method.invoke(Method.java:564)
at org.codehaus.groovy.reflection.CachedMethod.invoke(CachedMethod.java:98)
... groovy stuff here ...
at
cz.ocs.model.OCSEnterpriseObject.awakeFromInsertion(OCSEnterpriseObject.groovy:265)
// [*]
at
com.webobjects.eocontrol.EOEditingContext.insertObjectWithGlobalID(EOEditingContext.java:2871)
at er.extensions.eof.ERXEC.insertObjectWithGlobalID(ERXEC.java:975)
at er.extensions.eof.ERXEC$insertObjectWithGlobalID$0.call(Unknown Source)
... groovy stuff here ...
at cz.ocs.model.OCSEnterpriseObject.staticEOSVM(OCSEnterpriseObject.groovy:164)
// the faultWith... here
===
The awake code looks simply like this:
===
void awakeFromInsertion(EOEditingContext ec) {
println "awakeFromInsertion: ${this} in ${ec}" // this one happens before the
NPE
super.awakeFromInsertion(ec) // [*]
println "supered awakeFromInsertion: ${this} in ${ec}" // we do not get this far
... ... ...
===
Alas, at the moment I have no idea at all what in my code might be the culprit
:(
Thanks and all the best,
OC
On 22 Aug 2018, at 3:14 AM, Chuck Hill
mailto:ch...@gevityinc.com>> wrote:
Yes, that makes complete sense. EOUtilities.objectWithPrimaryKeyValue is
documented to “Fetches the Enterprise Object identified by the specified
primary key value”. It will fetch. Always. It might not use the fetched
result, but you asked it to fetch so it will. The method you want is
faultWithPrimaryKeyValue. That will either resolve the object reference from
an EO already in the EC, in the shared EC, or in the snapshot cache. If all of
those fail, only then will it fetch.
And yes, the JavaDocs for faultWithPrimaryKeyValue are wrong (copied from
objectWithPrimaryKeyValue).
Chuck
From: "ocs@ocs" mailto:o...@ocs.cz>>
Date: Tuesday, August 21, 2018 at 5:15 PM
To: Chuck Hill mailto:ch...@gevityinc.com>>
Cc: "webobjects-dev@lists.apple.com"
mailto:webobjects-dev@lists.apple.com>>
Subject: Should loading a shared object try to fetch? (was: Should ERXEC get
sharedEC
Re: Should loading a shared object try to fetch? (was: Should ERXEC get sharedEC automagically?)
Chuck,
thanks again! I did not know that (well one could write a much bigger book than
you did with just those things I do not know...)
Nevertheless, there still must be some ugly fault of mine. When using
objectWithPrimaryKeyValue, there's the superfluous fetch, but it works.
When replaced by faultWithPrimaryKeyValue, I keep getting NPEs somewhere inside
of awakeFromInsertion — at the moment, I can't really make sense of it :( It
looks like this:
===
== about to load 110 from 348ad293 (SEC ) // first time, there's yet no SEC...
4105 [main] DEBUG NSLog - Using JDBCPlugIn
'com.webobjects.jdbcadaptor.FrontbasePlugIn' for JDBCAdaptor@2076611420
... and it is being initialised here; it fetches properly all the shared EOs ...
4110 [main] DEBUG NSLog - connecting with dictionary: {username = "";
password = ""; URL =
"jdbc:FrontBase://localhost/SBERDAT3/user=FINServis/isolation=read_committed/locking=optimistic";
}
... ... ... the proper SQL to load all the shared objects here ... ... ...
4235 [main] DEBUG NSLog - === Commit Internal Transaction
... and here's a problem; my overridden awakeFromInsertion logs this just
before a super.awakeFromInsertion:
awakeFromInsertion: in
er.extensions.eof.ERXEC@348ad293
... and a log _after_ super does not happen; instead, I get this:
java.lang.Exception:
... ... ...
Caused by: java.lang.NullPointerException
at
er.extensions.eof.ERXGenericRecord.awakeFromInsertion(ERXGenericRecord.java:512)
at
cz.ocs.model.OCSEnterpriseObject.super$4$awakeFromInsertion(OCSEnterpriseObject.groovy)
at
java.base/jdk.internal.reflect.NativeMethodAccessorImpl.invoke0(Native Method)
at
java.base/jdk.internal.reflect.NativeMethodAccessorImpl.invoke(NativeMethodAccessorImpl.java:62)
at
java.base/jdk.internal.reflect.DelegatingMethodAccessorImpl.invoke(DelegatingMethodAccessorImpl.java:43)
at java.base/java.lang.reflect.Method.invoke(Method.java:564)
at
org.codehaus.groovy.reflection.CachedMethod.invoke(CachedMethod.java:98)
... groovy stuff here ...
at
cz.ocs.model.OCSEnterpriseObject.awakeFromInsertion(OCSEnterpriseObject.groovy:265)
// [*]
at
com.webobjects.eocontrol.EOEditingContext.insertObjectWithGlobalID(EOEditingContext.java:2871)
at er.extensions.eof.ERXEC.insertObjectWithGlobalID(ERXEC.java:975)
at er.extensions.eof.ERXEC$insertObjectWithGlobalID$0.call(Unknown
Source)
... groovy stuff here ...
at
cz.ocs.model.OCSEnterpriseObject.staticEOSVM(OCSEnterpriseObject.groovy:164) //
the faultWith... here
===
The awake code looks simply like this:
===
void awakeFromInsertion(EOEditingContext ec) {
println "awakeFromInsertion: ${this} in ${ec}" // this one happens before the
NPE
super.awakeFromInsertion(ec) // [*]
println "supered awakeFromInsertion: ${this} in ${ec}" // we do not get this far
... ... ...
===
Alas, at the moment I have no idea at all what in my code might be the culprit
:(
Thanks and all the best,
OC
> On 22 Aug 2018, at 3:14 AM, Chuck Hill wrote:
>
> Yes, that makes complete sense. EOUtilities.objectWithPrimaryKeyValue is
> documented to “Fetches the Enterprise Object identified by the specified
> primary key value”. It will fetch. Always. It might not use the fetched
> result, but you asked it to fetch so it will. The method you want is
> faultWithPrimaryKeyValue. That will either resolve the object reference from
> an EO already in the EC, in the shared EC, or in the snapshot cache. If all
> of those fail, only then will it fetch.
>
> And yes, the JavaDocs for faultWithPrimaryKeyValue are wrong (copied from
> objectWithPrimaryKeyValue).
>
>
> Chuck
>
>
> From: "ocs@ocs" mailto:o...@ocs.cz>>
> Date: Tuesday, August 21, 2018 at 5:15 PM
> To: Chuck Hill mailto:ch...@gevityinc.com>>
> Cc: "webobjects-dev@lists.apple.com "
> mailto:webobjects-dev@lists.apple.com>>
> Subject: Should loading a shared object try to fetch? (was: Should ERXEC get
> sharedEC automagically?)
>
> Even without ERXObjectStoreCoordinatorPool, there's another strange
> behaviour. I happen to be loading my objects (many of which happen to be in
> the SEC) using “EOUtilities.objectWithPrimaryKeyValue”.
>
> If the object happens to be in the SEC, I get it all right, but when I switch
> on SQL log, it looks like it is fetched anyway?!? I.e., something like this:
>
> ===
> println "== about to load $pk from $ec.identityHashString (SEC
> $ec.sharedEditingContext.identityHashString)"
> def foo=ec.sharedEditingContext.objectsByEntityName[ename].find {
> it.rawPrimaryKey==pk }
> if (foo!=nil) println " it IS pre-loaded in SEC: $foo"
> //object=EOUtilities.objectWithPrimaryKeyValue(ec,ename,pk)
>
> object=ERXEOControlUtilities.objectWithPrimaryKeyValue(ec,ename,pk,NSArray.EmptyArray,NO,NO)
> println " got $object"
> ===
>
> logs out something like
>
> ===
> ==
Re: Should loading a shared object try to fetch? (was: Should ERXEC get sharedEC automagically?)
Yes, that makes complete sense. EOUtilities.objectWithPrimaryKeyValue is
documented to “Fetches the Enterprise Object identified by the specified
primary key value”. It will fetch. Always. It might not use the fetched
result, but you asked it to fetch so it will. The method you want is
faultWithPrimaryKeyValue. That will either resolve the object reference from
an EO already in the EC, in the shared EC, or in the snapshot cache. If all of
those fail, only then will it fetch.
And yes, the JavaDocs for faultWithPrimaryKeyValue are wrong (copied from
objectWithPrimaryKeyValue).
Chuck
From: "ocs@ocs"
Date: Tuesday, August 21, 2018 at 5:15 PM
To: Chuck Hill
Cc: "webobjects-dev@lists.apple.com"
Subject: Should loading a shared object try to fetch? (was: Should ERXEC get
sharedEC automagically?)
Even without ERXObjectStoreCoordinatorPool, there's another strange behaviour.
I happen to be loading my objects (many of which happen to be in the SEC) using
“EOUtilities.objectWithPrimaryKeyValue”.
If the object happens to be in the SEC, I get it all right, but when I switch
on SQL log, it looks like it is fetched anyway?!? I.e., something like this:
===
println "== about to load $pk from $ec.identityHashString (SEC
$ec.sharedEditingContext.identityHashString)"
def foo=ec.sharedEditingContext.objectsByEntityName[ename].find {
it.rawPrimaryKey==pk }
if (foo!=nil) println " it IS pre-loaded in SEC: $foo"
//object=EOUtilities.objectWithPrimaryKeyValue(ec,ename,pk)
object=ERXEOControlUtilities.objectWithPrimaryKeyValue(ec,ename,pk,NSArray.EmptyArray,NO,NO)
println " got $object"
===
logs out something like
===
== about to load 101 from 25673087 (SEC a7cf42f)
it IS pre-loaded in SEC:
4663 [main] DEBUG NSLog - === Begin Internal Transaction
4663 [main] DEBUG NSLog - evaluateExpression:
4664 [main] DEBUG NSLog - 1 row(s) processed
4665 [main] DEBUG NSLog - === Commit Internal Transaction
got
===
Happens even with a
“ERXEOControlUtilities.objectWithPrimaryKeyValue(...,NSArray.EmptyArray,false,false)”
instead.
Does it make any sense? I would presume that with the desired object in the
SEC, no database roundtrip should be needed (and done) at all?
Thanks and all the best,
OC
On 21 Aug 2018, at 10:02 PM, ocs@ocs mailto:o...@ocs.cz>> wrote:
Indeed! If I switch off the OSCPool, it starts to work properly.
Thanks just again!
Nevertheless, I still must be missing something of grave importance, for with
OCSPool (I use ), I would presume the SEC for the pool being currently used by
the ERXEC would load the shared objects?
It does not: the global one does automatically load the shared objects, but the
SEC-based one of ERXEC remains empty.
Note: the code in question does not run in a session context; it is performed
at launch, before the first session is created. Might that be important perhaps?
All the best,
OC
On 21 Aug 2018, at 9:42 PM, Chuck Hill
mailto:ch...@gevityinc.com>> wrote:
Are you using the ERXObjectStoreCoordinatorPool? It keeps one SEC per pool,
not one shared globally. EOSharedEditingContext.defaultSharedEditingContext()
is the global one.
Chuck
From: "ocs@ocs" mailto:o...@ocs.cz>>
Date: Tuesday, August 21, 2018 at 12:23 PM
To: Chuck Hill mailto:ch...@gevityinc.com>>
Cc: "webobjects-dev@lists.apple.com"
mailto:webobjects-dev@lists.apple.com>>
Subject: Re: Should ERXEC get sharedEC automagically?
P.S. It seems ERX completely ignores the default shared EC, using its own one.
If I try e.g., this:
===
println "The default sharedEC is
${EOSharedEditingContext.defaultSharedEditingContext()}"
6.times {
def e=ERXEC.newEditingContext()
println "EC $e gets sec $e.sharedEditingContext"
}
println "The default sharedEC still is
${EOSharedEditingContext.defaultSharedEditingContext()}"
===
it looks like this:
===
The default sharedEC is com.webobjects.eocontrol.EOSharedEditingContext@26bbe604
2005 [main] INFO er.extensions.eof.ERXObjectStoreCoordinatorPool -
initializing Pool...
2008 [main] INFO er.extensions.eof.ERXObjectStoreCoordinatorPool -
initializing Pool finished
EC er.extensions.eof.ERXEC@40e32762 gets sec
com.webobjects.eocontrol.EOSharedEditingContext@5875de6a
EC er.extensions.eof.ERXEC@7d78f3d5 gets sec
com.webobjects.eocontrol.EOSharedEditingContext@5875de6a
EC er.extensions.eof.ERXEC@f5b6e78 gets sec
com.webobjects.eocontrol.EOSharedEditingContext@5875de6a
EC er.extensions.eof.ERXEC@71926a36 gets sec
com.webobjects.eocontrol.EOSharedEditingContext@5875de6a
EC er.extensions.eof.ERXEC@48976e6d gets sec
com.webobjects.eocontrol.EOSharedEditingContext@5875de6a
EC er.extensions.eof.ERXEC@7f6874f2 gets sec
com.webobjects.eocontrol.EOSharedEditingContext@5875de6a
The default sharedEC still is
com.webobjects.eocontrol.EOSharedEditingContext@26bbe604
===
Thanks and all the best,
OC
On 21 Aug 2018, at 9:07 PM, ocs@ocs mailto:o...@ocs.cz>> wrote:
Chuck,
