moosetechnology/moose

MOOSE - Platform for software and data analysis.

moose's Issues

FAMIXNameResolver fixing uniqueNameForMethodWithSmalltalkSelector:

Originally reported on Google Code with ID 19

FAMIXNameResolver 
uniqueNameForMethodWithSmalltalkSelector: 
                hasClassScope:
                class:.

does not use at all the hasClassScope: argument

Reported by stephane.ducasse on 2008-09-28 19:55:43

Put author and timestamp in entity property

Originally reported on Google Code with ID 21

In squeak this is easy to know when and who changed a method 
so we should add this information in Moose

Reported by stephane.ducasse on 2008-12-02 08:45:58

FamixTest>>testMSEImport does not work

Originally reported on Google Code with ID 1

 FamixTest>>testMSEImport does not work as it should

Reported by stephane.ducasse on 2008-08-18 20:29:32

Task: Improve the Moose Finder GUI interface

Originally reported on Google Code with ID 5

Classes need to be commented. Especially MooseFinder, which seems to be the main class.
Few 
example how to open such finder are necessary

MooseFinder can be loaded from the Moose SqueakSource repository.

Reported by alexandre.bergel on 2008-08-21 08:49:03

Many EyeSee examples walkback because #isColor not defined on Object

Originally reported on Google Code with ID 22

Folks, I was playing with the EyeSeeDevelopment from scg repository.  Very
nice!

When I evaluate "EyeSee.examples new historyDiagram" with doit it fails
with a message not understood.

The fix is to define the method #isColor on Object.

-Alan Wostenberg
Key Technology
PS: if this isn't the right place to submit bug report, please let me know.

Reported by awostenberg on 2009-01-09 21:27:55

accesses to replace FAMIXBehaviouralEntity>>outgoingAccesses

Originally reported on Google Code with ID 32

just helping the intuition.

Sometimes it helps to have the direction of the relationship in the name, sometimes
you dont care.

Reported by simon.denier on 2009-01-26 09:56:39

Describe MooseModel in Fame

Originally reported on Google Code with ID 36

MooseModel needs to be described for Fame to handle import/export of MSE projects.
Basically, a Model in package Moose has a name (String) and a list of Famix.Entity

Reported by simon.denier on 2009-02-11 16:16:06

clean the VW visitor to remove doXXX

Originally reported on Google Code with ID 24

The VWvisitor is not that good and there are a lot of old methods.

Reported by stephane.ducasse on 2009-01-20 16:29:59

fix isStub in VW and in Sq

Originally reported on Google Code with ID 27

and use the incompleteProperty instead for the current recursive isStub

Reported by stephane.ducasse on 2009-01-22 15:43:18

[MOOSE] port SCG algorithms to SqMoose

Originally reported on Google Code with ID 41

See SCG-Algorithms in Bern store.

Lots of valuable code in there, which we should port to make a library (or even better,
a 
"platform") of generic algorithms.

Reported by simon.denier on 2009-02-26 18:11:36

Fix reference to MetaDescriptionInspector

Originally reported on Google Code with ID 14

MooseElement>>inspectorClasses 

    ^super inspectorClasses copyWith: MetaDescriptionInspector

Reported by stephane.ducasse on 2008-08-22 21:22:47

Push properties pragma detection/cache into Fame

Originally reported on Google Code with ID 29

Moose entities properties are declared using the<property: #LOC longName: description:>

pragma.
This allows to retrieve a property using its symbolic name (#LOC), as well as to cache
the result 
of property computation.

Right now when a property is accessed with its symbolic name, MooseEntity>>propertyNamed:

launches the MoosePragmaProcessor to parse all classes from the hierarchy and select
the 
corresponding method. It is done each time.

The idea would be to at least do the pragma processing one time, when importing a 
(meta)model, so that the system would directly retrieve the right method if necessary.
So it 
should be done in Fame for genericity.

Note that it was the previous behavior with Meta, although there was too complex (redundant)

things with a cache at both the symbolic access and at the method level.

Reported by simon.denier on 2009-01-22 18:13:19

fix resolve: in >>FAMIX2SmalltalkMethodVisitor

Originally reported on Google Code with ID 7

We should use the compiler to resolve the bindings and create adequate Famix20 models.
Need marcus knowledge

Reported by stephane.ducasse on 2008-08-21 09:59:04

Task: Migration Map

Originally reported on Google Code with ID 6

Check what has been imported in Squeak. Moose has originally been developed on VisualWorks,

and only a part of the Moose suit-tools has been imported. We lost track on what has
been 
migrated to Squeak. 
The goal of migrating Moose is _not_ to port on Squeak everything that is Moose related,
but a 
minimal core big enough to make it useable. 

We need a list of packages that are loadable in Squeak, the ones that are missing (i.e.,
exist in VW 
but not in Squeak).

Please, provide this list as a comment to this issue.

Reported by alexandre.bergel on 2008-08-21 09:51:56

[FAMIX3] NamedEntity#packagedIn should be derived and renamed to parentPackage

Originally reported on Google Code with ID 45

The declaration (and subsequent implementation) is wrong as it is not derived and there
is an 
instance variable at the level of NamedEntity.

Instead, packagedIn is derived and should be defined on a per-subclass basis.

Also rename to parentPackage for homogeneity with other attribute names.

Reported by simon.denier on 2009-02-26 18:42:21

Specialize from/to in FamixAssociation subclasses

Originally reported on Google Code with ID 31

from/to is a couple of generic accessors for FamixAssociation. Like belongsTo for NamedEntity,
it is 
an abstract property in FamixAssociation which is derived in subclasses.

from/to mapping:
Access from a method to a structure - Access: accessor -> variable
Invocation from a method to a structure - Invocation: sender -> receiver
Reference from a method to a structure- Reference: from -> to
Inheritance from a class to its superclass- Inheritance: subclass-> superclass

Reported by simon.denier on 2009-01-22 18:32:45

[FAME] Handling of unknown attributes when importing MSE files (mismatched models)

Originally reported on Google Code with ID 38

Currently the importer throws a DNU when it tries to initialize an unknown property.
It would be 
better if the importer showed an error in the Transcript, then continued with the process,
ignoring 
unknown properties. It would accomodate for differences between platform-specificities
of Famix 
for example.

See FMFutureAttribute>>maybeEnd
 (...) property setOn: owner instance values: values

FM3MetaDescription>>attributeNamed:
(returns nil if absent)
FM3MetaDescription>>attributeNamed: aString ifAbsent: aBlock
(actually the block is not executed in all absent cases)

Reported by simon.denier on 2009-02-16 09:56:33

Smalltalk importer: fix stub

Originally reported on Google Code with ID 48

Stef fix stub

Reported by simon.denier on 2009-02-26 18:53:12

ModelTest>>testReferenceModel

Originally reported on Google Code with ID 17

Please make this test green :)

Reported by stephane.ducasse on 2008-08-25 15:44:22

outgoingInheritances into superclassInheritances

Originally reported on Google Code with ID 23

We should change the famix 30 interface

Reported by stephane.ducasse on 2009-01-20 14:43:04

Package extension importer for FAMIX20

Originally reported on Google Code with ID 9

Famix 20 st code to famix importer should work for package extension

Reported by stephane.ducasse on 2008-08-21 16:10:58

self/super/.... for Famix 20

Originally reported on Google Code with ID 10

Famix 20 st code to famix importer should work for self super....

Reported by stephane.ducasse on 2008-08-21 16:11:36

[FAME] add helper methods for multivalued attributes

Originally reported on Google Code with ID 43

The generator could take care of compiling addXxxx helper methods for multivalued attribute.

Reported by simon.denier on 2009-02-26 18:25:46

Generate Famix30 from fame

Originally reported on Google Code with ID 11

Get the famix30 model generated from fame

Get the Group associated

Reported by stephane.ducasse on 2008-08-22 09:50:51

fix sourcetext to use the equivalent of Filename

Originally reported on Google Code with ID 12

sourceText

    self flag: #broken.
    "fix Filename use below"
    self mooseModel isJava
        ifTrue:
            [| filename separator |
            separator := "Filename" self separator asSymbol asString.
            separator := Filename separator asSymbol asString.
            self sourceAnchor isNil
                ifFalse:
                    [filename := ('src' , separator , self mooseModel name , separator
                        , self sourceAnchor) asFilename]
                ifTrue:
                    [filename := ('src' , separator , self mooseModel name , separator
                        ,
                            ((self mooseName
                                copyReplaceAll: '::'
                                with: separator) copyUpTo: $$) , '.java') asFilename].
            ^filename exists
                ifTrue: [filename readStream contents]
                ifFalse: ['']].
    self mooseModel isSmalltalk
        ifTrue:
            [^self mooseName , Character cr asSymbol asString
                ,
                    ((self attributes collect: #name) join: Character cr asSymbol asString)
                , Character cr asSymbol asString
                ,
                    ((self methods collect: #sourceText) join: Character cr asSymbol asString)].
    self mooseModel isPython ifTrue: [^self pythonASTNode sourceText].
    ^''

Reported by stephane.ducasse on 2008-08-22 21:20:11

Fix references to ModelHistory

Originally reported on Google Code with ID 15

allModelHistories

    <navigation: 'All model histories'>
    ^self allWithType: ModelHistory asMetaDescription

Reported by stephane.ducasse on 2008-08-22 21:23:25

Implement a Visitor for Famix3

Originally reported on Google Code with ID 39

Implement a Visitor pattern for the elements of Famix3.
- define accept: protocol in Famix classes
- define a default Visitor which clients could subclass for mininal specialization

The default Visitor should:
- implement a default traversal of a Famix model, guaranteeing that each element would
be 
visited only once (done using the hierarchical structure of package-class-method-attribute...

namespace?)
- be parameterized by ImportingContext so as to easily configure the default traversal

Part of Famix-Extensions

Reported by simon.denier on 2009-02-26 17:53:07

Adapt #property:longname:description pragma to #MSEProperty:type:

Originally reported on Google Code with ID 37

See Issue 29 for the motivation.

Create  #MSEProperty:type: pragma frm the old ones #property:longname:description:
Keep the old ones for future use by Moose Finder?
Move those extended properties to a specific *Moose-Famix3-extensions

Reported by simon.denier on 2009-02-13 17:58:31

FamixAttribute Object.DependentsFields is duplicated by importer (famix2/famix3)

Originally reported on Google Code with ID 28

Class variable DependentsFields from Object appears two times in a Famix model created
by the 
Smalltalk importer.

Appers in both Famix2 and Famix3.

See FamixPropertiesTest>>testClassAttributes: test inherited attributes

Reported by simon.denier on 2009-01-22 15:47:44

[FAMIX3] implement select utilities in Package to select classes in childNamedEntities

Originally reported on Google Code with ID 46

Package#childNamedEntities should be reserved for elements other than classes like subpackages

and namespaces

Reported by simon.denier on 2009-02-26 18:50:00

SqMoose - FAMIX3.0 ImplicitVariable name should be a symbol and not a string

Originally reported on Google Code with ID 20

FAMIX3.0 ImplicitVariable name should be a symbol and not a string

Reported by stephane.ducasse on 2008-11-19 15:06:41

Smalltalk importer: remove duplicate registration

Originally reported on Google Code with ID 47

See how classes are registrered in namespace, package.

Also check how the extended attribute is handled by the importer

Reported by simon.denier on 2009-02-26 18:51:52

Loading package alone could make sense

Originally reported on Google Code with ID 3

It would be nice that we could load package alone (without been forced to load package)
For this it would be good to extend the MoosePundleImporterTask or MooseSqueakClassPackageImporterTask.




testImportLANPackage
    "self debug: #testImportLANPackage"

    | model |
    model := MooseModel2 new.   
    MooseSqueakClassPackageImporterTask new
        importerClass: FAMIX2SmalltalkImporter;
        doNotRunCandidateOperator; 
        importingContext: ImportingContext new importPackage; yourself;
        model: model;
        addFromPackageNamed: #'Moose-LAN';
        run;
        yourself.
    self deny: (model allPackages isEmpty).
        self assert: (model allPackages size = 1)

Reported by stephane.ducasse on 2008-08-19 14:10:55

mooseModel is not working and I cruelly need help for this one.

Originally reported on Google Code with ID 2

testReferenceModelInEntities 
    "self debug: #testReferenceModelInEntities"

    self model entities 
        do: [:each | self assert: each mooseModel == self model]

Reported by stephane.ducasse on 2008-08-18 21:05:44

Famix2.0 classes described with FAME

Originally reported on Google Code with ID 8

I would really like to have FAMIX20 described in FAME
This is important that FAMIX20 and FAMIX30 are in sync at this level

Reported by stephane.ducasse on 2008-08-21 16:09:53

Dummy issue for GitHub Export

Originally reported on Google Code with ID 34

To avoid confusion with the generic accessors in FamixAssociation (which will do the
redirection 
anyway)

Reported by simon.denier on 2009-01-29 13:44:55

Fix resolve: in >>FAMIX2SmalltalkMethodVisitor

Originally reported on Google Code with ID 13

Famix2AbstractFile>>systemFolderSeparatorString
    ^Filename separator asSymbol asString

Reported by stephane.ducasse on 2008-08-22 21:20:52

Implement previous/next in Famix3 SqMoose

Originally reported on Google Code with ID 30

previous/next (was above/below in previous Famix3 model) are a couple of FamixAssociation

properties which should represent a sequence in the order of association (it is an
abstract 
representation of the sequence of statements).

Currently the Smalltalk importer for SqMoose do nothing about it.

Reported by simon.denier on 2009-01-22 18:20:34

[FAMIX] Compability visitor for Famix2/famix3

Originally reported on Google Code with ID 44

Would be cool to have a visitor which will build the corresponding Famix3 model from
a Famix2 
model. And would ease the transition.

Reported by simon.denier on 2009-02-26 18:28:58

We should check if all the tests have been imported in Squeak

Originally reported on Google Code with ID 4

Please load the latest version in VW and in Squeak and check if the tests are all imported
in 
Squeak.

- If not please load the fileOut30 from the scg store and save the code (open the file
setting and 
choose Squeak). 

- Import it in Squeak, run the FixUnderscore or FixLF goodies that you can find in
the 
FixUnderscore package.

- add the package to the Moose-all
- and publish the code
- You can also check the Undeclared by running Smalltalk cleanOutUndeclared. 
and checking the value of the Undeclared global variable

Reported by stephane.ducasse on 2008-08-19 14:29:39

FAMIX3: review method organization to clean Famix-Core

Originally reported on Google Code with ID 40

Famix-Core should only contain the reference implementation of Famix3 as described and

generated by Fame.

Other methods should be moved to Famix-Implementation (general purpose helper methods),

Famix-Extensions (more specific extensions), Famix-Smalltalk (language specific extension).

Reported by simon.denier on 2009-02-26 18:03:22

[FAME] Add annotation/comment support in Fame metametamodel

Originally reported on Google Code with ID 42

Annotation/comment support in the metametamodel would allow self-described metamodel
to be 
exported and generated.

Doru takes care of this.

Reported by simon.denier on 2009-02-26 18:19:25

[FAMIX3] Check attribute declarations are correct and synchronized

Originally reported on Google Code with ID 50

Check attribute declarations are correct and synchronized with code and between them.
Some type errors have appeared and some attributes have changed (name, relationship).

For example:
- ContainerEntity#incomingReferences and #outgoingReferences should point to source/target

instead of from/to.

Reported by simon.denier on 2009-02-26 19:01:57

Update Famix-Specifications package

Originally reported on Google Code with ID 35

Update Famix3 specs in the Famix-Specifications package to mirror the changes in the
reference 
implementation of Famix3 in SqMoose.

This is important to get a cross-platform Famix3 model

Reported by simon.denier on 2009-02-09 14:07:51

implicit importer is wrong: an implicit variable should be attached to a method not a class

Originally reported on Google Code with ID 16

mseImplicitVar := self model entityNamed: #'Root::Smalltalk::LAN::Node.self'.
    self assert: mseImplicitVar name = #self.
    self assert: mseImplicitVar mooseName = #'Root::Smalltalk::LAN::Node.self'.
    "self assert: (mseImplicitVar interfaceSignatures includes: #send:)."
    mseImplicitVar := self model entityNamed: #'Root::Smalltalk::LAN::Node.super'.
    self assert: mseImplicitVar name = #super.
    self assert: mseImplicitVar mooseName = #'Root::Smalltalk::LAN::Node.super'.

Reported by stephane.ducasse on 2008-08-24 09:12:48

Moose-LAN test model is slightly outdated

Originally reported on Google Code with ID 26

To get a running example of the Moose-LAN model (and playing around), type the following
in a 
workspace
(TODO: make it a class method in Node...)

n1 := WorkStation new
    name: (SingleDestinationAddress  new id: 'tto').
n2  := PrintServer new
    name: (SingleDestinationAddress  new id: 'ffo').

n1 nextNode: n2.
n2 nextNode: n1.

n1 originate: (Packet new addressee: (SingleDestinationAddress  new id: 'ffo'); contents:
'hello').


It will display strings in Transcript telling what the simulator does, like:
a SingleDestinationAddress sending the packet to a SingleDestinationAddress
------------------------------------------------------------
Printer a SingleDestinationAddress prints hello

Only problem (as it appears) is that SingleDestinationAddress does not print its id/name
(or, that SingleDestinationAddress has to be tested against another SingleDestinationAddress
when we 
would like to just have a node name)

Reported by simon.denier on 2009-01-22 14:05:38

Rename FamixReference from/to properties to source/target

Originally reported on Google Code with ID 33

To avoid confusion with the generic accessors in FamixAssociation (which will do the
redirection 
anyway)

Reported by simon.denier on 2009-01-29 13:44:55

fix the charts from the VW moose browser

Originally reported on Google Code with ID 18

What steps will reproduce the problem?
1. Open the browser
2. ...
3. click on the buttons on the right hand side of a group view

Reported by tudor.girba on 2008-08-28 13:30:31

[FAMIX3] Smalltalk Importer: implement candidate invocations

Originally reported on Google Code with ID 49

[FAMIX3] Smalltalk Importer: fix candidate invocations

Reported by simon.denier on 2009-02-26 18:56:46

Rename VisualWorksParseTree in VisualWorksParseTreeVisitor

Originally reported on Google Code with ID 25

Misleading name are killing

Reported by stephane.ducasse on 2009-01-20 16:56:50