patsystem-users — PAT user discussions

[Patsystem-users] Re: using PAT with RDBMS

[Tomasz N. <nt...@ir...>]

Michael Frieß <MichaelFriess <at> gmx.net> writes:

> Hi,
> you wrote on your 
> Homepage "Though, it still allows integration with relational databases." (on
http://patsystem.sourceforge.net/about.html at 
> the very bottom).
> 
> How can one use PAT 
> with MySQL (or Oracle)? I would appreciate it if you could give me 
> some more information on that topic. 
>  
> Thanks,
> Michael

Well. There are several options.

First of all external storage like RDBMS allows one to store a _lot_ of data
outside the main memory. That's advantage when there is really huge amount of
data. It is advicable in this situation to simply store/load part of your
business objects outside of main-memory/PAT/Prevayler.

Let's take historical data as an example: an information about when an user
makes a transfer of money. There could be lot of simple Record objects in the
database. (How you load/store/map them is up to you)

So the User class might look like this:
User
{
    String name;
    //...
    List<Record> records;
}

Or better (the example uses Spring's JDBCTemplate, google: equinox,
https://equinox.dev.java.net/)

UserManager
{
    //..
    public List getHistoricalRecordsOfUser(User user)
    {
        return new UsersQuery(getDataSource()).executeQuery();
    }
}
//adapt this
    class RecordsQuery extends MappingSqlQuery {
        public RecordsQuery(DataSource ds) {
            super(ds, "SELECT * FROM user_historical_ops");
            compile();
        }   
        protected Object mapRow(ResultSet rs, int rowNum) throws SQLException {
            Record record = new Record();
            record.setId(new Long(rs.getLong("id")));
            record.setWhat(rs.getString("type"));
            record.setWhen(rs.getString("when"));
            //etc...
            return record;
        }   
    }           


Every single developer may introduce RDBMS to projects using PAT without much
problems (except the RDBM maintenance). Just a custom developing needed to
load/store data in a RDBMS.
And that gives you 2 things: simplicity and speed of data layer using PAT and
flexibility and interoperability of RDBMS.

The next scenario is to integrate with an external RDBM system. That'd be also
not a problem. One can use some sort of network protocols to exchange data
(Hessian, RMI or whatever exists there).

So everything is about customisation and a common sense. Use PAT for as the
primary data layer (because simplifies development up) and integrate it with a
RDBM when the need comes.

I hope that helps. Sorry for a lag with the reply.

Tomasz

Re: [Patsystem-users] AOP-Loadtime, log4j and PAT

[Tomasz N. <Tom...@ii...>]

On the next beautiful day: Mon, 19 September, 2005, Michael Frie=DF wrote=
:

> Hi,
> =20
> I'm trying to build my project with the AOP Classloader. I'm getting th=
e
> following error message:
> [..]
> =20
> =20
> The funny things are, that 1) log4j is definitely loaded as you can see=
 on
> the first three lines (my app uses log4j, too) and 2) it works fine if =
I use
> the AOP-compiler... (without the AOP Classloader)
> =20
> Any ideas what might be the reason for the error?
> =20
> Thanks,
> Michael


If you send/re-check the part of your build.xml (or whatever else)
responsible for setting the classpath and invoking JBossAOP then it
will be more helpfull to find the place.

From what we see, there is a problem with the absence of the log4j.jar.

Send the file please.

Tomasz


--=20
  _i______'simplicity_is_the_key'__________tomasz_nazar
  _ii____'i_am_concern_oriented'________________PJKM.pl
  _iii__'patsystem.sf.net'___________________linux_user
  _Heaven_&_Fellows,_PPP______________________prevayler

[Patsystem-users] AOP-Loadtime, log4j and PAT

[<Mic...@gm...>]

Hi,
 
I'm trying to build my project with the AOP Classloader. I'm getting the
following error message:
 
  _____  

runClient:
     [java] [12:09:06      ConsoleClient] INFO    ConsoleClient is being
executed!
     [java] [12:09:06 ofilingInterceptor] DEBUG   profiling
eshop.test.ConsoleClient->test
     [java] [12:09:06 ofilingInterceptor] DEBUG   done (360ms)
     [java] java.lang.NoClassDefFoundError: org/apache/log4j/Logger
     [java]     at
org.nthx.pat.RootCreationInterceptor.<clinit>(RootCreationInterceptor.java:5
0)
     [java]     at
sun.reflect.NativeConstructorAccessorImpl.newInstance0(Native Method)
     [java]     at
sun.reflect.NativeConstructorAccessorImpl.newInstance(NativeConstructorAcces
sorImpl.java:39)
     [java]     at
sun.reflect.DelegatingConstructorAccessorImpl.newInstance(DelegatingConstruc
torAccessorImpl.java:27)
     [java]     at
java.lang.reflect.Constructor.newInstance(Constructor.java:274)
     [java]     at java.lang.Class.newInstance0(Class.java:308)
     [java]     at java.lang.Class.newInstance(Class.java:261)
     [java]     at
org.jboss.aop.advice.GenericAspectFactory.createPerClass(GenericAspectFactor
y.java:146)
     [java]     at org.jboss.aop.Advisor.addPerClassAspect(Advisor.java:487)
     [java]     at
org.jboss.aop.advice.ScopedInterceptorFactory.create(ScopedInterceptorFactor
y.java:52)
     [java]     at
org.jboss.aop.Advisor.createInterceptorChain(Advisor.java:531)
     [java]     at
org.jboss.aop.Advisor.resolveConstructorPointcut(Advisor.java:696)
     [java]     at
org.jboss.aop.ClassAdvisor.createInterceptorChains(ClassAdvisor.java:454)
     [java]     at org.jboss.aop.ClassAdvisor$1.run(ClassAdvisor.java:233)
     [java]     at java.security.AccessController.doPrivileged(Native
Method)
     [java]     at
org.jboss.aop.ClassAdvisor.attachClass(ClassAdvisor.java:205)
     [java]     at
org.jboss.aop.AspectManager.getAdvisor(AspectManager.java:439)
     [java]     at eshop.pojo.biz.EShopApp.<clinit>(EShopApp.java)
     [java]     at
eshop.test.ConsoleClient.eshop$test$ConsoleClient$test$aop(ConsoleClient.jav
a:106)
     [java]     at
eshop.test.ConsoleClient$test_N7986479998679492778.invokeNext(ConsoleClient$
test_N7986479998679492778.java)
     [java]     at
eshop.aop.SimpleProfilingInterceptor.invoke(SimpleProfilingInterceptor.java:
91)
     [java]     at
eshop.test.ConsoleClient$test_N7986479998679492778.invokeNext(ConsoleClient$
test_N7986479998679492778.java)
     [java]     at eshop.test.ConsoleClient.test(ConsoleClient.java)
     [java]     at eshop.test.ConsoleClient.main(ConsoleClient.java:88)
     [java] Exception in thread "main" 

  _____  

 
The funny things are, that 1) log4j is definitely loaded as you can see on
the first three lines (my app uses log4j, too) and 2) it works fine if I use
the AOP-compiler... (without the AOP Classloader)
 
Any ideas what might be the reason for the error?
 
Thanks,
Michael

[Patsystem-users] using PAT with RDBMS

[<Mic...@gm...>]

Hi,
 
you wrote on your Homepage "Though, it still allows integration with
relational databases." (on http://patsystem.sourceforge.net/about.html at
the very bottom).
 
How can one use PAT with MySQL (or Oracle)? I would appreciate it if you
could give me some more information on that topic. 
 
Thanks,
Michael

[Patsystem-users] Re: Difficulties with PAT

[Tomasz N. <Tom...@ii...>]

> From: Michael
> The problem occures when I am trying to create objects. If I create them
> instantly (= assigning their values within their constructor):
>     HelpMePlz help = new HelpMePlz("pat");
> and restart my application, then their are correctly reloaded
> However if I create them with an empty constructor:
>     HelpMePlz help = new HelpMePlz();
>     help.setAbout("pat");
> and then assigning their values with setter methods and restarting my
> application, they are loaded as "null" objects


Yep!

You missed out '@pat.transaction' on setter method ('setAbout(..)'). :)
I will try to emphasize this better on the webpage:

Even setters change state of your objects, so you also have to annotate
them.

So in your code:
/**
 * @@pat.bo
 */
public class Bonbon
{
    //..

    public void setNumber(int number) //<-- has to have @pat.transaction
    {
        this.number = number;
    }
}

As you know JBossAOP, you can use regular expressions instead having
annotation inside the code:

<aop>
<annotation-introduction expr="execution(void @pat.bo->set*(..))">
    @pat.transaction
</annotation-introduction>
</aop>

(put above inside jboss-aop.xml, put the file into META-INF dir, and
packa as my-configuration.jar, put into classpath).

This saves some amount of code, but on the other hand ... you have to
manage external file instead of having annotations inside the source.
You choose what's the best for you :)



Take care, and good luck with your essay - Tomasz


-- 
  _i______'simplicity_is_the_key'__________tomasz_nazar
  _ii____'i_am_concern_oriented'________________PJKM.pl
  _iii__'patsystem.sf.net'___________________linux_user
  _Heaven_&_Fellows,_PPP______________________prevayler

[Patsystem-users] M1-08: Deploy PAT on Tomcat/Struts

[Tomasz N. <Tom...@ii...>]

Hi there!

I've just released a template of a web appliation which uses PAT.
If anyone uses Struts then he is happy with that as it is based on it.
Validation and struts-el, language properties also included.

The basic functionality includes creating an object and presenting a
list of all objects existing in the system. Not too much but a piece of
good code to start with..


Just copy downloaded 'pat-struts-template.war' into you favourite TOMCAT
installation and you're almost there. One also has to setup
'jboss.aop.path' for Tomcat - but the application is so smart ;) it
informs if it's done. It isn't so smart to setup this property,
though. :(

Details:
http://patsystem.sourceforge.net/struts_template.html


Cheers!

t.

-- 
  _i______'simplicity_is_the_key'__________tomasz_nazar
  _ii____'i_am_concern_oriented'__________________iiuwr
  _iii__'patsystem.sf.net'___________________linux_user
  _Heaven_&_Fellows,_PPP______________________prevayler

Re: [Patsystem-users] what about object deletion?

[Tomasz N. <Tom...@ii...>]

On the next beautiful day: Fri, 24 June, 2005, Vladimir Atehort=FAa wrote=
:

> Hi
>=20
>=20
> >PAT doesn't. You can.
> >((org.nthx.pat.Persisteble)root).removeBO(oid);
>=20
> I reaally wish I dont have my business objects (where logic belongs) co=
upled/aware_of  org.nthx.pat.Persisteble.

Of course :)

> >Map consists of two 4 keys-values:
> >1 -> root                (we need to keep that object)
> >2 -> root.child
> >3 -> other               (this should be garbage collected soon)
> >4 -> other.andHisChild
>=20
> >happen? How can any weak hashMap recognize to remove 'other' (among wi=
th
> >his children) and not 'root' (with children)?=20
>=20
> >There are no references to OID: 1 and OID: 3.
> >I think there might be some tricky hack to that. I just didn't have ti=
me
>=20
> If i'm not wrong, it should be rather easy:
>=20
>  public IdentityMapMixin()
>     {
>         uniqueOID =3D 1L;
>         map =3D new org.apache.commons.collections.map.ReferenceMap;   =
    // should use a map with strong keys and soft values.


Can you handle writing a JUnit test case for that:

Steps:
1) create sample data for PAT (u may use org.nthx.pat.demo or add
anything you need to src/tests/)

2) take snapshot

3) Pat.unload() or whatever it takes - maybe even shutdown of the whole V=
M=20
will be necessary .. yeap, I think so.

4) restore data by creating root again and asserting:
    a) are objects on the place
    b) there are no garbage in IdentityMap

5) Of course change to IdentityMapMixin is necessary.

Can you? :)
If it will work I will include _your_ contribution to PAT.

t.


--=20
  _i______'simplicity_is_the_key'__________tomasz_nazar
  _ii____'i_am_concern_oriented'__________________iiuwr
  _iii__'patsystem.sf.net'___________________linux_user
  _Heaven_&_Fellows,_PPP______________________prevayler

Re: [Patsystem-users] what about object deletion?

[<vla...@ip...>]

Hi


>> =BFhow does PAT handle the "old objects eat all ram" issue?

>PAT doesn't. You can.
>Just write on OID down, and use
>((org.nthx.pat.Persisteble)root).removeBO(oid);
>in the appropriate moment..

I reaally wish I dont have my business objects (where logic belongs) =
coupled/aware_of  org.nthx.pat.Persisteble.


>I was looking at WeakHashMap, but it isn't serializable (I don't know =
if
>making it Serializable would cause sth extra..), and second: I'm not
>sure how exactly does it work when (de)serializing such IdentityMap.
>
>If you have some experience on that, share it here.

weakHashMap definitelly won't cut it.
weakhashmap holds soft references for the KEYS, which in PAT case's are =
the OIDs.
As soon as no one holds a strong refference to any OID, the OID key will =
be removed from the map and thus the BO (map value). this is not what we =
want because most likely OIDs are not used, pure pojo references are. =
composite objects instead of DB foreign keys.

jakarta commons collections offers Maps that hold soft or weak =
references on the Values rather than the keys.
this means that if a BO is not reachable from another BO (meaning, it =
can't be reached indirectly from the root), only then could it be GC'd. =
I think this is what we want.

Using jakarta common's collections ReferenceMap, the difference between =
weakreference and softreference lies only in timing:
* an object being weak referenced is erased as soon as the object is not =
reachable (by other "real" strong refferences).
* an object being soft referenced is not erased immediatly when it =
becomes not reachable, but later, when (if) memory gets tight.

Dont even consider using sun's default weakhashmap.

>Here is example on the problem I just can't solve..
>I'm not sure how all weak hashMaps solve following problem:

>Map consists of two 4 keys-values:
>1 -> root                (we need to keep that object)
>2 -> root.child
>3 -> other               (this should be garbage collected soon)
>4 -> other.andHisChild

>'root' is our ROOT. Object 2 is referenced by root and stored in IM
>(IdentityMap). Object 3 however, is stored in IM (cause it was created)
>but is not referenced by anyone and should be removed in the next
>session.

>So we close application, take snapshot, open application,  and what =
will
>happen? How can any weak hashMap recognize to remove 'other' (among =
with
>his children) and not 'root' (with children)?=20

>There are no references to OID: 1 and OID: 3.

Only sun's weakhashmap needs references to OID's.

If you use jakarta commons, all you need is to somewhere hold areference =
to root. root itself holds references to its children, and they do the =
grandchildren. If any normal business logic makes an object unreachable, =
then it should be garbagecollected (BO's dont want to go to identitymap =
and look for it).
No object reachable from root will ever be garbage collected. Maybe  =
RootCreationInterceptor can arrange for there being a strong reference =
to root somewhere, or maybe we need to turn root into a singleton object =
(referenced by its class, in a static field).


>I think there might be some tricky hack to that. I just didn't have =
time
>and didn't care. If you really need this we can think of it. I would be
>happy to see the solution and apply it to PAT. Want to contribute?

>:)

If i'm not wrong, it should be rather easy:
changing IdentityMapMixin's constructor to not use:

 public IdentityMapMixin()
    {
        uniqueOID =3D 1L;
        map =3D new HashMap();
    }

but instead:

 public IdentityMapMixin()
    {
        uniqueOID =3D 1L;
        map =3D new org.apache.commons.collections.map.ReferenceMap;     =
  // should use a map with strong keys and soft values.
    }


... of course, then add a new library, include it's license text etc.

Re: [Patsystem-users] what about object deletion?

[Tomasz N. <Tom...@ii...>]

On the next beautiful day: Fri, 24 June, 2005, Vladimir Atehort=FAa wrote=
:

> Hi.
>=20
> I've been re-checking the source, and it seems business objects never g=
et garbage collected, mainly because the identity map holds strong refere=
nces to them.

For my primary app I just don't care about the problem.

Persisteble interface of the root object provides the funcionality. I've
never used it however. (where was YAGNI at the moment? ;) )

> =BFhow does PAT handle the "old objects eat all ram" issue?

PAT doesn't. You can.
Just write on OID down, and use
((org.nthx.pat.Persisteble)root).removeBO(oid);
in the appropriate moment..


> I can only think of the obvious solution, which is using a Soft HashMap=
 instead of a regular hashmap.
> Sun's provided weakhashmap wouldnt work because it holds weak refferenc=
es to keys, and PAT needs weak refferences to values.

I was looking at WeakHashMap, but it isn't serializable (I don't know if
making it Serializable would cause sth extra..), and second: I'm not
sure how exactly does it work when (de)serializing such IdentityMap.

If you have some experience on that, share it here.

Here is example on the problem I just can't solve..
I'm not sure how all weak hashMaps solve following problem:

Map consists of two 4 keys-values:
1 -> root                (we need to keep that object)
2 -> root.child
3 -> other               (this should be garbage collected soon)
4 -> other.andHisChild

'root' is our ROOT. Object 2 is referenced by root and stored in IM
(IdentityMap). Object 3 however, is stored in IM (cause it was created)
but is not referenced by anyone and should be removed in the next
session.

So we close application, take snapshot, open application,  and what will
happen? How can any weak hashMap recognize to remove 'other' (among with
his children) and not 'root' (with children)?=20

There are no references to OID: 1 and OID: 3.


I think there might be some tricky hack to that. I just didn't have time
and didn't care. If you really need this we can think of it. I would be
happy to see the solution and apply it to PAT. Want to contribute?

:)

> Regards,
> Vladimir

Best regards,
Tomasz

--=20
  _i______'simplicity_is_the_key'__________tomasz_nazar
  _ii____'i_am_concern_oriented'__________________iiuwr
  _iii__'patsystem.sf.net'___________________linux_user
  _Heaven_&_Fellows,_PPP______________________prevayler

[Patsystem-users] what about object deletion?

[<vla...@ip...>]

Hi.

I've been re-checking the source, and it seems business objects never =
get garbage collected, mainly because the identity map holds strong =
references to them.

=BFhow does PAT handle the "old objects eat all ram" issue?

I can only think of the obvious solution, which is using a Soft HashMap =
instead of a regular hashmap.
Sun's provided weakhashmap wouldnt work because it holds weak =
refferences to keys, and PAT needs weak refferences to values.

I'd say the best candidate woul dbe jakarta commons collection's =
"Reference Map" configured to use weak refferences:

http://jakarta.apache.org/commons/collections/apidocs-COLLECTIONS_3_1/org=
/apache/commons/collections/map/ReferenceMap.html

Regards,

Vladimir

Re: [Patsystem-users] Is this possible?

[Tomasz N. <Tom...@ii...>]

On the next beautiful day: Wed, 22 June, 2005, Vladimir Atehort=FAa wrote=
:

> Hello.
>=20

Hi Vladimir!

> This idea has been haunting me, so I want to share it to see what you t=
hink:
>=20
> Imagine one has a PAT system, where "business logic" resides where it s=
hould: in the methods of the business objects. I call my set of business =
objects my "domain model"=20
>=20
> Then I have some objects outside from the domain model (and outside the=
 pat system), that are interested in the objects inside the domain model.=
 like:
>=20
>=20
> Unfortunately I don't know much about AOP, except from a few articles I=
 read online and the PAT code itself.
>=20
> =BFwhat do you think of this?

Well, I think you're becoming "aspectized" :)

Before continuing with the topic have a look at JBoss' page about
'Observable" pattern.
I think this is what you need.

They've implemented the pattern using AOP. I haven't tried it in real
life, but was looking at it before.=20


"Observing" concern is orthogonal to persistence (which PAT provides).
Hence, I'm not so eager on mering it into PAT (for now).
But for you this isn't any problem at all.

You may use both together: PAT and Observabble aspect. Any help on that
you need - just ask.

PAT comes with default <PAT>/etc/your-aspects-here-aop.xml
Put your custom aspect configurations (mixins, annotations, ..) there.

You will probably have to copy contents of
http://wiki.jboss.org/wiki/Wiki.jsp?page=3DGOFObserverableJBossAopXML
into your-aspects-here-aop.xml file.


The whole page about AOP'ed Observable pattern  is here:
http://wiki.jboss.org/wiki/Wiki.jsp?page=3DGOFObservable

Get back here and tell the story how it worked out.
Or if that's not what you need, then I'm open for discussion..

:)

Tomasz


PS. Few tips about JBossAOP:
Tip 1: any annotation you see in the code (@@pat.bo for example) may be
moved to configuration file (XML) (and vice-versa)
Tip 2: there can be many AOP configuration files
Tip 3: 'jboss.aop.path' system variable specifies path to your AOP XML
files

--=20
  _i______'simplicity_is_the_key'__________tomasz_nazar
  _ii____'i_am_concern_oriented'__________________iiuwr
  _iii__'patsystem.sf.net'___________________linux_user
  _Heaven_&_Fellows,_PPP______________________prevayler

[Patsystem-users] Is this possible?

[<vla...@ip...>]

Hello.

This idea has been haunting me, so I want to share it to see what you =
think:

Imagine one has a PAT system, where "business logic" resides where it =
should: in the methods of the business objects. I call my set of =
business objects my "domain model"=20

Then I have some objects outside from the domain model (and outside the =
pat system), that are interested in the objects inside the domain model. =
like:

* A swing GUI responsible for visualizing some of the business objects. =
It wants to know when a visualized object has changed (as a result of a =
PAT/prevayler transaction/command) in order to refresh the windows, =
instead of -for example- repeated polling. It does not want to modify =
them or anything. just display.

I have some objects like this, with responsabilities that don't belong =
in the domain model (like visualization), and those objects want to be =
notified of changes is the domain objects (instead of for example, =
repeated polling).

The obvious solution is to do it "by hand" using the observer pattern: =
to have my domain objects implement some "eventsource" interface, and my =
external (GUI) objects implement some "eventlistener" interface, then =
register the listeners to the selected sources and voil=E8

However, this not only becomes tedious (repeating trough every single =
object the methods that implement "eventsource" (like =
"registerNotificationListener", "notifyListeners", etc ), and is error =
prone / pollutes the bussines objects, it also seems to be a natural =
place to put AOP in action.

In the same way that the PAT code merges clean POJO "business objects" =
with those "mixin" implementations of PAT interfaces, I think a default =
implementation could be made of "eventsource" interface and have the AOP =
magic "mix" any business object with that class, thus having free =
"beans/swing style event notification" for any business object existing =
or future.

Unfortunately I don't know much about AOP, except from a few articles I =
read online and the PAT code itself.

=BFwhat do you think of this?

[Patsystem-users] Re: PAT's roadmap

[Tomasz N. <nt...@ir...>]

Vladimir Atehortúa <vladimir.atehortua <at> iptotal.com> writes:

> Hello all.
>  
> I'd like to know what is PAT's 
> roadmap.

Hi! First of all sorry for the delay. Your email haven't arrived to me.. very
strange. It is on the list, but not in my mailbox.. 

  
> Where is it going, I mean it could be going towards 
> improving the prevalence layer to get improved performance (prevayler's site 
> says its transctions are slower than MySQL and SQLServer, altough queries are 
> much faster), or it could be coing towards new features in the AOP 
> arena.

Forget about performance. I'm not on it. (Though nothing is faster than reading
your memory :) - write speed is much more worse.)

It's about simpler application model. AOP - as you've probably know - helps a
lot here.

Where is it going to? Who knows. Project is driven by a need. There isn't any
roadmap here. When need arises...


> Just a very small comment:
>  
> in org.nthx.pat.IdentityMapMixin, in the constructor:
>  
>     public IdentityMapMixin()    
> {        uniqueOID = 1L;        map = new HashMap();    }
>  
>  
> The hashmap is initializaed with no capacity. Be 
> aware that resizing a map always involve either rehashing or 
> copying the entire 
> contents of the map. Thus, map resizing should be avoided. If you 
> know how many 
> objects are expected to fit into the map (maybe from the highest OID) you can 
> avoid a ton of map resizing at startup (when the identity map is being 
> populated) and thus improve boot time.

Thanks for looking into the source! :)
If a great demand on the feature exists I could provide some sort of
configuration for that...

  
> Regards,
>  
> Vladimir Atehortúa


Take care,
Tomasz Nazar

[Patsystem-users] testing - ignore this

[Tomasz N. <Tom...@ii...>]

ignore this..

-- 
  _i______'simplicity_is_the_key'__________tomasz_nazar
  _ii____'i_am_concern_oriented'__________________iiuwr
  _iii__'patsystem.sf.net'___________________linux_user
  _Heaven_&_Fellows,_PPP______________________prevayler

[Patsystem-users] PAT's roadmap

[<vla...@ip...>]

Hello all.

I'd like to know what is PAT's roadmap.

Where is it going, I mean it could be going towards improving the =
prevalence layer to get improved performance (prevayler's site says its =
transctions are slower than MySQL and SQLServer, altough queries are =
much faster), or it could be coing towards new features in the AOP =
arena.

Just a very small comment:

in org.nthx.pat.IdentityMapMixin, in the constructor:

    public IdentityMapMixin()
    {
        uniqueOID =3D 1L;
        map =3D new HashMap();
    }



The hashmap is initializaed with no capacity. Be aware that resizing a =
map always involve either rehashing or copying the entire contents of =
the map. Thus, map resizing should be avoided. If you know how many =
objects are expected to fit into the map (maybe from the highest OID) =
you can avoid a ton of map resizing at startup (when the identity map is =
being populated) and thus improve boot time.

Regards,

Vladimir Atehort=FAa

Re: [Patsystem-users] strange pat.transaction doesn't but snapshot does

[Tomasz N. <Tom...@ii...>]

On the next beautiful day: Tue, 14 June, 2005, Chris wrote:

> Hi!
> 
> I have following situation:
> 
> /** @@pat.transaction */
> public void updateEntry(Entry entry, int id) {
> 	remove(id); // pat.transaction
> 	add(entry); // pat.transaction
> }


The similar examples are in the distro:
http://patsystem.sourceforge.net/api/org/nthx/pat/tests/RecursiveTransactionsTest.java.html
Test: 'testTransferMoney'

It should work the way you've described. Are you shure your test case
tests the right thing? ;)


T.

-- 
  _i______'simplicity_is_the_key'__________tomasz_nazar
  _ii____'i_am_concern_oriented'__________________iiuwr
  _iii__'patsystem.sf.net'___________________linux_user
  _Heaven_&_Fellows,_PPP______________________prevayler

[Patsystem-users] strange pat.transaction doesn't but snapshot does

[Chris <ci...@ya...>]

Hi!

I have following situation:

/** @@pat.transaction */
public void updateEntry(Entry entry, int id) {
	remove(id); // pat.transaction
	add(entry); // pat.transaction
}

the above method updates entry by removing the old one
and then inserting the new one.

After running this method I get what I expect, the
fields in entry are updated. I checked it no problem.
But it should be acomplished as a pat.transaction so
the changes should be visible after reloading the
application. But they're not updated.

But in the same situation after running updateEntry
method when I run takeSnapshot(), after reloading
everything is fine. The entry is updated.

Where is the problem, what am I forgetting?

thanks for help
ciapecki

__________________________________________________
Do You Yahoo!?
Tired of spam?  Yahoo! Mail has the best spam protection around 
http://mail.yahoo.com 

[Patsystem-users] M1-06 released

[Tomasz N. <Tom...@ii...>]

Hi there!


PAT-M1-06 was released few days ago.

I suggest you to add: http://patsystem.sourceforge.net/rss.xml
to your favourite RSS reader if you haven't done it already.

What has changed:

Only minor changes:

Docs updated - it is worth to have a look at:
docs/htdocs/framework.gtml#transactionsArgs at least. 

Fixed since M1-05:
a) Pat.unload() stays 
b) when instantiating ROOT second time exception
with a warning is thrown


Also, I blogged a little about Ilybra (app which uses PAT) and aspects
here:
http://www.ii.uni.wroc.pl/~nthx/blog/


Take care,
Tomasz Nazar



-- 
  _i______'simplicity_is_the_key'__________tomasz_nazar
  _ii____'i_am_concern_oriented'__________________iiuwr
  _iii__'patsystem.sf.net'___________________linux_user
  _Heaven_&_Fellows,_PPP______________________prevayler

Re: [Patsystem-users] constructors as pat.transactions

[Tomasz N. <Tom...@ii...>]

On Thu, May 12, 2005 at 06:08:22AM -0700, Chris wrote:
> > build.xml: (it isn't necessary, but for safety I've
> > changed it)
> > <target name="test" description="run tests"
> > depends="clean, compile">
> >                                                     
> > ^^^^^
> 
> Now it works (at least under Linux) ;)
> The clean target helped, maybe there were some tedious
> classes from the previous builds.
> 
> later will try under Windows.
> 
> thanks a lot
> ciapek
> 
> Krzysztof Cierpisz

Thanks a lot to you ;)
Check in on Windows in spare time and let me know..

see you,
t.

-- 
  _i______'simplicity_is_the_key'__________tomasz_nazar
  _ii____'i_am_concern_oriented'__________________iiuwr
  _iii__'patsystem.sf.net'___________________linux_user
  _Heaven_&_Fellows,_PPP______________________prevayler

Re: [Patsystem-users] constructors as pat.transactions

[Chris <ci...@ya...>]

> build.xml: (it isn't necessary, but for safety I've
> changed it)
> <target name="test" description="run tests"
> depends="clean, compile">
>                                                     
> ^^^^^

Now it works (at least under Linux) ;)
The clean target helped, maybe there were some tedious
classes from the previous builds.

later will try under Windows.

thanks a lot
ciapek

Krzysztof Cierpisz

..........................
gg: 2349371 ..............
mobile: (+49 176) 23329677
phone:   (+49 89) 95482147


		
Yahoo! Mail
Stay connected, organized, and protected. Take the tour:
http://tour.mail.yahoo.com/mailtour.html

Re: [Patsystem-users] constructors as pat.transactions

[Tomasz N. <Tom...@ii...>]

On Wed, May 11, 2005 at 10:42:20PM -0700, Chris wrote:
> > Ciapek, please re-run your test cases with M1-05 on
> > Windows and report
> > results here, if you may. Thanks!
> 
> I just did that (Windows), but it looks worse than
> before:
> Even the simple constructor Register() doesn't go
> through:
> 
> java.lang.NoSuchMethodException: getSize

  ^^^
So this above was due to old libs, right? ;)

> 
> ciapek

build.xml: (it isn't necessary, but for safety I've changed it)
<target name="test" description="run tests" depends="clean, compile">
                                                     ^^^^^


sources/register/tests/RegisterTest.java:

    public void testConstructorGivenEntries() {
        entries = new ArrayList();
        entries.add(sampleEntry);
        register=new Register((ArrayList)entries);
        assertEquals(1,register.getNumberOfEntries());
    }

    public void testConstructorGivenEmptyEntries()
    {
        register = new Register(new ArrayList());
        assertEquals(0,register.getNumberOfEntries());
    }


One has to execute Pat.unload(), before every ROOT constructor.
So in this case 'Pat.unload()' exists inside 'setUp()', what makes it
the same.
So, you have to split your test into two. They test different things
anyway, right? 

Take care,
t.

PS. I'll make 'new ROOT(..) triggers Pat.unload()' default behaviour 
in next major release.



-- 
  _i______'simplicity_is_the_key'__________tomasz_nazar
  _ii____'i_am_concern_oriented'__________________iiuwr
  _iii__'patsystem.sf.net'___________________linux_user
  _Heaven_&_Fellows,_PPP______________________prevayler

Re: [Patsystem-users] constructors as pat.transactions

[Tomasz N. <Tom...@ii...>]

On Wed, May 11, 2005 at 10:42:20PM -0700, Chris wrote:
> > Ciapek, please re-run your test cases with M1-05 on
> > Windows and report
> > results here, if you may. Thanks!
> 
> I just did that (Windows), but it looks worse than
> before:
> Even the simple constructor Register() doesn't go
> through:
> 
> Caused by: java.lang.RuntimeException:
> java.lang.NoSuchMethodException: getSize
> 
> The same Failure I get running Register(List)
> testCase.
> 
> maybe you change something in
> prevayler1-for_pat-M1-04.jar and I should update it?
> 
> ciapek
> 
> Krzysztof Cierpisz


Remove lib/*04*.jar
cp <PAT>/dist/*.jar <REGISTER>/lib/

If this won't help, may you send me the whole test case once more?
thnx

t.

-- 
  _i______'simplicity_is_the_key'__________tomasz_nazar
  _ii____'i_am_concern_oriented'__________________iiuwr
  _iii__'patsystem.sf.net'___________________linux_user
  _Heaven_&_Fellows,_PPP______________________prevayler

Re: [Patsystem-users] constructors as pat.transactions

[Chris <ci...@ya...>]

> Ciapek, please re-run your test cases with M1-05 on
> Windows and report
> results here, if you may. Thanks!

I just did that (Windows), but it looks worse than
before:
Even the simple constructor Register() doesn't go
through:

Testcase:
testConstructor(register.tests.RegisterTest):	Caused
an ERROR
null
java.lang.ExceptionInInitializerError
	at
register.tests.RegisterTest.testConstructor(Unknown
Source)
	at
sun.reflect.NativeMethodAccessorImpl.invoke0(Native
Method)
	at
sun.reflect.NativeMethodAccessorImpl.invoke(NativeMethodAccessorImpl.java:39)
	at
sun.reflect.DelegatingMethodAccessorImpl.invoke(DelegatingMethodAccessorImpl.java:25)
Caused by: java.lang.RuntimeException:
java.lang.NoSuchMethodException: getSize
	at
org.jboss.aop.ClassAdvisor.attachClass(ClassAdvisor.java:250)
	at
org.jboss.aop.AspectManager.getAdvisor(AspectManager.java:321)
	at register.Register.<clinit>(Unknown Source)
	... 16 more
Caused by: java.lang.NoSuchMethodException: getSize
	at
org.jboss.aop.ClassAdvisor.getMethod(ClassAdvisor.java:261)
	at
org.jboss.aop.ClassAdvisor.getMethod(ClassAdvisor.java:269)
	at
org.jboss.aop.ClassAdvisor.populateMixinMethods(ClassAdvisor.java:332)
	at
org.jboss.aop.ClassAdvisor$1.run(ClassAdvisor.java:228)
	at java.security.AccessController.doPrivileged(Native
Method)
	at
org.jboss.aop.ClassAdvisor.attachClass(ClassAdvisor.java:209)
	... 18 more


The same Failure I get running Register(List)
testCase.

maybe you change something in
prevayler1-for_pat-M1-04.jar and I should update it?

ciapek

Krzysztof Cierpisz

..........................
gg: 2349371 ..............
mobile: (+49 176) 23329677
phone:   (+49 89) 95482147


		
__________________________________ 
Yahoo! Mail Mobile 
Take Yahoo! Mail with you! Check email on your mobile phone. 
http://mobile.yahoo.com/learn/mail 

Re: [Patsystem-users] constructors as pat.transactions

[Tomasz N. <Tom...@ii...>]

On Tue, May 10, 2005 at 03:22:42AM -0700, Chris wrote:
> > Watch this.. 
> > 
> >     public void testConstructorGivenEntries()
> >     {
> >         register = new Register(new ArrayList());
> >        
> > assertEquals(0,register.getNumberOfEntries());
> >         entries = new ArrayList();
> >         entries.add(sampleEntry);
> >         register.setEntries(entries);
> >        
> > assertEquals(1,register.getNumberOfEntries());
> >     }
> > 
> 
> when I run exactly the same configuration you wrote, I
> got Failure:
> 
> Maybe that's the problem of Windows, I will try later
> at home on Linux box. It looks like setUp() didn't
> reset Pat and there's still the sampleEntry added in
> the first Test.

Ciapek, please re-run your test cases with M1-05 on Windows and report
results here, if you may. Thanks!

http://prdownloads.sourceforge.net/patsystem/pat-M1-05.tgz?download

Take care,
Tomasz

-- 
  _i______'simplicity_is_the_key'__________tomasz_nazar
  _ii____'i_am_concern_oriented'__________________iiuwr
  _iii__'patsystem.sf.net'___________________linux_user
  _Heaven_&_Fellows,_PPP______________________prevayler

[Patsystem-users] New release - M1-05

[Tomasz N. <Tom...@ii...>]

Hi guys!

New release ready to download.
http://prdownloads.sourceforge.net/patsystem/pat-M1-05.tgz?download

Please download and run tests on Windows:
$ ant tests

And report to me if that pass or fails. Thanks.


====
This is from RSS feed: http://patsystem.sourceforge.net/rss.xml

 1.  M1-05 release
 Docs simplified. Fixed since M1-04: a) dist/*.jar are now synchronized
 with src b) a try to fix Windows-tests failures I encourage you to at
 least look through web pages. Please test for yourselves if tests run
 on Windows (and report).
====
 

Take care,
Tomasz Nazar

-- 
  _i______'simplicity_is_the_key'__________tomasz_nazar
  _ii____'i_am_concern_oriented'__________________iiuwr
  _iii__'patsystem.sf.net'___________________linux_user
  _Heaven_&_Fellows,_PPP______________________prevayler