parentBehaviouralEntity is nil for Java local variable declared in static block (FAMIXLocalVariable)

List overview All Threads
Download

newer

older

TRCanvas canvas/surface

Java interfaces and clientTypes

Fuhrmanator

12 Sep 2016 12 Sep '16

6 p.m.

Here's another anomaly I found in my reasoning about Java interfaces through MSE files. In the example with ArgoUML 0.28.1 there's a class called ConfigurationFactory.java that declares a local variable in a static block: static { ... IConfigurationFactory newFactory = null; ... } Within MOOSE, the FAMIXLocalVariable "newFactory" shows up in the structuresWithDeclaredTypes for the FAMIXClass "IConfigurationFactory". However, its parentBehaviouralEntity is nil. Is this because static blocks in Java don't map to behavioural entities in FAMIX? I'm trying to understand the nil case, as it doesn't happen with all FAMIXLocalVariables. The problem is obviously that I can't find the dependency on the interface in this case. In the stated example, there's also a private static final Attribute, so the MOOSE dependency comes through that in the end (it's a redundant dependency). But I'm guessing there could be some cases where a dependency via local variable used in a static block might not always be stored in an Attribute. Any ideas on how to work around this? Possibly related: https://github.com/moosetechnology/moose/issues/785 -- View this message in context: http://forum.world.st/parentBehaviouralEntity-is-nil-for-Java-local-variabl… Sent from the Moose mailing list archive at Nabble.com.

Show replies by date

Tudor Girba

12 Sep 12 Sep

6:46 p.m.

New subject: parentBehaviouralEntity is nil for Java local variable declared in static block (FAMIXLocalVariable)

Hi, The static blocks should map on an <init> stub method belonging to the parent class. The latest jdt2famix has this semantics. Could it be that the MSE you are using is an old one? If so, could you redo the MSE with jdt2famix? If no, could you point me more closely to the source code? Also, please only use this address: moose-dev(a)list.inf.unibe.ch :) Cheers, Doru

...

On Sep 12, 2016, at 6:00 PM, Fuhrmanator <fuhrmanator(a)gmail.com> wrote: Here's another anomaly I found in my reasoning about Java interfaces through MSE files. In the example with ArgoUML 0.28.1 there's a class called ConfigurationFactory.java that declares a local variable in a static block: static { ... IConfigurationFactory newFactory = null; ... } Within MOOSE, the FAMIXLocalVariable "newFactory" shows up in the structuresWithDeclaredTypes for the FAMIXClass "IConfigurationFactory". However, its parentBehaviouralEntity is nil. Is this because static blocks in Java don't map to behavioural entities in FAMIX? I'm trying to understand the nil case, as it doesn't happen with all FAMIXLocalVariables. The problem is obviously that I can't find the dependency on the interface in this case. In the stated example, there's also a private static final Attribute, so the MOOSE dependency comes through that in the end (it's a redundant dependency). But I'm guessing there could be some cases where a dependency via local variable used in a static block might not always be stored in an Attribute. Any ideas on how to work around this? Possibly related: https://github.com/moosetechnology/moose/issues/785 -- View this message in context: http://forum.world.st/parentBehaviouralEntity-is-nil-for-Java-local-variabl… Sent from the Moose mailing list archive at Nabble.com. _______________________________________________ Moose-dev mailing list Moose-dev(a)list.inf.unibe.ch https://www.list.inf.unibe.ch/listinfo/moose-dev

-- www.tudorgirba.com www.feenk.com "The coherence of a trip is given by the clearness of the goal."

Cris Fuhrman

13 Sep 13 Sep

7:15 p.m.

OK - trying a reply via e-mail as opposed to the web-based interface (hope this helps). Yes, I was using the MSE files from the moose book reference, which are old. So, I regenerated the MSE using the latest jdt2famix, but I've run into another problem (being a smalltalk/pharo newbie). Before I could find all interfaces limited to the argouml namespace with the following expression: interfaces := (MooseModel root allModels third allClasses select: [ :each | (each container name beginsWith:'org::argouml') and: (each isInterface)]) asSet. However, the latest jdt2famix model has a namespaces such that a container name is just e.g. "core" and then one has to traverse to the parent namespace, etc. The "org::argouml' part I'm looking for is at the end of such a traversal. I used the Inspector on the instances and couldn't understand how to get a fully qualified name as before. I'm not enough of a Pharo/SmallTalk coder to figure out how to concatenate it (build it myself). Is there some magic to make this simple as before? Cheers! On Mon, Sep 12, 2016 at 12:46 PM, Tudor Girba <tudor(a)tudorgirba.com> wrote:

...

On Sep 12, 2016, at 6:00 PM, Fuhrmanator <fuhrmanator(a)gmail.com> wrote: Here's another anomaly I found in my reasoning about Java interfaces

through

MSE files. In the example with ArgoUML 0.28.1 there's a class called ConfigurationFactory.java that declares a local variable in a static

block:

static { ... IConfigurationFactory newFactory = null; ... } Within MOOSE, the FAMIXLocalVariable "newFactory" shows up in the structuresWithDeclaredTypes for the FAMIXClass "IConfigurationFactory". However, its parentBehaviouralEntity is nil. Is this because static blocks in Java don't map to behavioural entities

FAMIX? I'm trying to understand the nil case, as it doesn't happen with

all

FAMIXLocalVariables. The problem is obviously that I can't find the dependency on the

interface

in this case. In the stated example, there's also a private static final Attribute, so

the

MOOSE dependency comes through that in the end (it's a redundant dependency). But I'm guessing there could be some cases where a

dependency

via local variable used in a static block might not always be stored in

Attribute. Any ideas on how to work around this? Possibly related: https://github.com/moosetechnology/moose/issues/785 -- View this message in context: http://forum.world.st/

parentBehaviouralEntity-is-nil-for-Java-local-variable- declared-in-static-block-FAMIXLocalVariable-tp4915235.html

Sent from the Moose mailing list archive at Nabble.com. _______________________________________________ Moose-dev mailing list Moose-dev(a)list.inf.unibe.ch https://www.list.inf.unibe.ch/listinfo/moose-dev

Cris Fuhrman

8:37 p.m.

OK - I went message surfing in the inspector after a good lunch and found the new way -- it's mooseName instead of name: interfaces := (MooseModel root allModels third allClasses select: [ :each | (each container mooseName beginsWith:'org::argouml') and: (each isInterface)]) asSet. I also confirm that the init block variables have a good parentBehaviouralEntity. Thanks again! On Tue, Sep 13, 2016 at 1:15 PM, Cris Fuhrman <fuhrmanator(a)gmail.com> wrote:

...

On Sep 12, 2016, at 6:00 PM, Fuhrmanator <fuhrmanator(a)gmail.com> wrote: Here's another anomaly I found in my reasoning about Java interfaces

through

MSE files. In the example with ArgoUML 0.28.1 there's a class called ConfigurationFactory.java that declares a local variable in a static

block:

FAMIX? I'm trying to understand the nil case, as it doesn't happen with

all

FAMIXLocalVariables. The problem is obviously that I can't find the dependency on the

interface

in this case. In the stated example, there's also a private static final Attribute,

so the

MOOSE dependency comes through that in the end (it's a redundant dependency). But I'm guessing there could be some cases where a

dependency

via local variable used in a static block might not always be stored in

Attribute. Any ideas on how to work around this? Possibly related: https://github.com/moosetechnology/moose/issues/785 -- View this message in context: http://forum.world.st/parentBe

haviouralEntity-is-nil-for-Java-local-variable-declared- in-static-block-FAMIXLocalVariable-tp4915235.html

Tudor Girba

9:08 p.m.

New subject: parentBehaviouralEntity is nil for Java local variable declared in static block (FAMIXLocalVariable)

Hi, Great. You found it! :) Cheers, Doru

...

On Sep 13, 2016, at 8:37 PM, Cris Fuhrman <fuhrmanator(a)gmail.com> wrote: OK - I went message surfing in the inspector after a good lunch and found the new way -- it's mooseName instead of name: interfaces := (MooseModel root allModels third allClasses select: [ :each | (each container mooseName beginsWith:'org::argouml') and: (each isInterface)]) asSet. I also confirm that the init block variables have a good parentBehaviouralEntity. Thanks again! On Tue, Sep 13, 2016 at 1:15 PM, Cris Fuhrman <fuhrmanator(a)gmail.com> wrote: OK - trying a reply via e-mail as opposed to the web-based interface (hope this helps). Yes, I was using the MSE files from the moose book reference, which are old. So, I regenerated the MSE using the latest jdt2famix, but I've run into another problem (being a smalltalk/pharo newbie). Before I could find all interfaces limited to the argouml namespace with the following expression: interfaces := (MooseModel root allModels third allClasses select: [ :each | (each container name beginsWith:'org::argouml') and: (each isInterface)]) asSet. However, the latest jdt2famix model has a namespaces such that a container name is just e.g. "core" and then one has to traverse to the parent namespace, etc. The "org::argouml' part I'm looking for is at the end of such a traversal. I used the Inspector on the instances and couldn't understand how to get a fully qualified name as before. I'm not enough of a Pharo/SmallTalk coder to figure out how to concatenate it (build it myself). Is there some magic to make this simple as before? Cheers! On Mon, Sep 12, 2016 at 12:46 PM, Tudor Girba <tudor(a)tudorgirba.com> wrote: Hi, The static blocks should map on an <init> stub method belonging to the parent class. The latest jdt2famix has this semantics. Could it be that the MSE you are using is an old one? If so, could you redo the MSE with jdt2famix? If no, could you point me more closely to the source code? Also, please only use this address: moose-dev(a)list.inf.unibe.ch :) Cheers, Doru

-- www.tudorgirba.com www.feenk.com "The coherence of a trip is given by the clearness of the goal." _______________________________________________ Moose-dev mailing list Moose-dev(a)list.inf.unibe.ch https://www.list.inf.unibe.ch/listinfo/moose-dev _______________________________________________ Moose-dev mailing list Moose-dev(a)list.inf.unibe.ch https://www.list.inf.unibe.ch/listinfo/moose-dev

-- www.tudorgirba.com www.feenk.com "We can create beautiful models in a vacuum. But, to get them effective we have to deal with the inconvenience of reality."

2782

days inactive

2783

days old

moose-dev@list.inf.unibe.ch

Manage subscription

4 comments

3 participants

tags (0)

participants (3)

Cris Fuhrman
Fuhrmanator
Tudor Girba