Eclipse Community Forums: OCL » Environment and name resolving

Help

Home

Home » Modeling » OCL » Environment and name resolving

Show: Today's Messages :: Show Polls :: Message Navigator

Environment and name resolving [message #57578]

Mon, 02 June 2008 12:22

Felix Dorner

Messages: 676
Registered: July 2009

Senior Member

Hi,

I have troubles understanding how name resolving exactly works.

Given the following hierarchy, (Packages small letters, Classes capitalized)

model::first
::Person
::second
::Car

In the context of Person, which of the following path names would
resolve the "Car" element?

a) "second::Car"

b) "first::second::Car"

c) "model::first::second::Car"

Thanks,
Felix

Report message to a moderator

Re: Environment and name resolving [message #57602 is a reply to message #57578]

Mon, 02 June 2008 12:48

Felix Dorner

Messages: 676
Registered: July 2009

Senior Member

As OCL has already implemented this complex mechanism, how difficult
would it be to get its service to resolve an element by a path name,
given a class or an operation where the lookup should start?

Felix

Report message to a moderator

Re: Environment and name resolving [message #57648 is a reply to message #57578]

Tue, 03 June 2008 10:51

Eclipse User

Originally posted by: cdamus.zeligsoft.com

--=-yfgvPrt33Mqbi2ZRbFEK
Content-Type: text/plain
Content-Transfer-Encoding: 7bit

Hi, Felix,

All of these (partially) qualified names will find the Car type, as all
of "second" "first" and "model" are unique names within the namespaces
containing Person.

Cheers,

Christian

On Mon, 2008-06-02 at 14:22 +0200, Felix Dorner wrote:

> Hi,
>
> I have troubles understanding how name resolving exactly works.
>
> Given the following hierarchy, (Packages small letters, Classes capitalized)
>
> model::first
> ::Person
> ::second
> ::Car
>
> In the context of Person, which of the following path names would
> resolve the "Car" element?
>
> a) "second::Car"
>
> b) "first::second::Car"
>
> c) "model::first::second::Car"
>
>
>
> Thanks,
> Felix
>
>
>

--=-yfgvPrt33Mqbi2ZRbFEK
Content-Type: text/html; charset=utf-8

<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.0 TRANSITIONAL//EN">
<HTML>
<HEAD>
<META HTTP-EQUIV="Content-Type" CONTENT="text/html; CHARSET=UTF-8">
<META NAME="GENERATOR" CONTENT="GtkHTML/3.16.0">
</HEAD>
<BODY>
Hi, Felix, 
 
All of these (partially) qualified names will find the Car type, as all of "second" "first" and "model" are unique names within the namespaces containing Person. 
 
Cheers, 
 
Christian 
 
 
On Mon, 2008-06-02 at 14:22 +0200, Felix Dorner wrote:
<BLOCKQUOTE TYPE=CITE>
<PRE>
Hi,

I have troubles understanding how name resolving exactly works.

Given the following hierarchy, (Packages small letters, Classes capitalized)

model::first
 ::Person
 ::second
 ::Car

In the context of Person, which of the following path names would 
resolve the "Car" element?

a) "second::Car"

b) "first::second::Car"

c) "model::first::second::Car"

Thanks,
Felix

</PRE>
</BLOCKQUOTE>
</BODY>
</HTML>

--=-yfgvPrt33Mqbi2ZRbFEK--

Report message to a moderator

Re: Environment and name resolving [message #57674 is a reply to message #57602]

Tue, 03 June 2008 10:53

Eclipse User

Originally posted by: cdamus.zeligsoft.com

--=-qH4tvBgMOOkLhbmOyzmG
Content-Type: text/plain
Content-Transfer-Encoding: 7bit

Hi, Felix,

I'm not sure ... it may be different for Ecore and for UML. Have a look
in the debugger at what utilities the
Environment::lookupPathName(List<String>) or whatever it's called (don't
have it in front of me) uses.

cW

On Mon, 2008-06-02 at 14:48 +0200, Felix Dorner wrote:

> As OCL has already implemented this complex mechanism, how difficult
> would it be to get its service to resolve an element by a path name,
> given a class or an operation where the lookup should start?
>
> Felix

--=-qH4tvBgMOOkLhbmOyzmG
Content-Type: text/html; charset=utf-8

<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.0 TRANSITIONAL//EN">
<HTML>
<HEAD>
<META HTTP-EQUIV="Content-Type" CONTENT="text/html; CHARSET=UTF-8">
<META NAME="GENERATOR" CONTENT="GtkHTML/3.16.0">
</HEAD>
<BODY>
Hi, Felix, 
 
I'm not sure ... it may be different for Ecore and for UML.  Have a look in the debugger at what utilities the Environment::lookupPathName(List<String>) or whatever it's called (don't have it in front of me) uses. 
 
cW 
 
On Mon, 2008-06-02 at 14:48 +0200, Felix Dorner wrote:
<BLOCKQUOTE TYPE=CITE>
<PRE>
As OCL has already implemented this complex mechanism, how difficult 
would it be to get its service to resolve an element by a path name, 
given a class or an operation where the lookup should start?

Felix
</PRE>
</BLOCKQUOTE>
</BODY>
</HTML>

--=-qH4tvBgMOOkLhbmOyzmG--

Report message to a moderator

Re: Environment and name resolving [message #57695 is a reply to message #57648]

Tue, 03 June 2008 12:25

Felix Dorner

Messages: 676
Registered: July 2009

Senior Member

Christian W. Damus wrote:
> Hi, Felix,
>
> All of these (partially) qualified names will find the Car type, as all
> of "second" "first" and "model" are unique names within the namespaces
> containing Person.

Strike! Thats what I thought. I guess that there are no "style"
guidelines though...?

Thanks,
Felix

Report message to a moderator

Re: Environment and name resolving [message #57719 is a reply to message #57674]

Tue, 03 June 2008 12:27

Felix Dorner

Messages: 676
Registered: July 2009

Senior Member

Christian W. Damus wrote:
> Hi, Felix,
>
> I'm not sure ... it may be different for Ecore and for UML. Have a look
> in the debugger at what utilities the
> Environment::lookupPathName(List<String>) or whatever it's called (don't
> have it in front of me) uses.

Ok I'll check it out. Sad that the good ideas to reduce work always show
up _after_ the work has ben done :-(

Felix

Report message to a moderator

Re: Environment and name resolving [message #57744 is a reply to message #57695]

Tue, 03 June 2008 13:01

Eclipse User

Originally posted by: cdamus.zeligsoft.com

--=-vXE1v7EUBZeVSDhuO0F4
Content-Type: text/plain
Content-Transfer-Encoding: 7bit

Hi, Felix,

No, not really. In the fact of refactoring of the model structure,
you'd have to change any of these qualifications anyway.

cW

On Tue, 2008-06-03 at 14:25 +0200, Felix Dorner wrote:

> Christian W. Damus wrote:
> > Hi, Felix,
> >
> > All of these (partially) qualified names will find the Car type, as all
> > of "second" "first" and "model" are unique names within the namespaces
> > containing Person.
>
> Strike! Thats what I thought. I guess that there are no "style"
> guidelines though...?
>
> Thanks,
> Felix

--=-vXE1v7EUBZeVSDhuO0F4
Content-Type: text/html; charset=utf-8

<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.0 TRANSITIONAL//EN">
<HTML>
<HEAD>
<META HTTP-EQUIV="Content-Type" CONTENT="text/html; CHARSET=UTF-8">
<META NAME="GENERATOR" CONTENT="GtkHTML/3.16.0">
</HEAD>
<BODY>
Hi, Felix, 
 
No, not really.  In the fact of refactoring of the model structure, you'd have to change any of these qualifications anyway. 
 
cW 
 
On Tue, 2008-06-03 at 14:25 +0200, Felix Dorner wrote:
<BLOCKQUOTE TYPE=CITE>
<PRE>
Christian W. Damus wrote:
> Hi, Felix,
> 
> All of these (partially) qualified names will find the Car type, as all 
> of "second" "first" and "model" are unique names within the namespaces 
> containing Person.

Strike! Thats what I thought. I guess that there are no "style" 
guidelines though...?

Thanks,
Felix
</PRE>
</BLOCKQUOTE>
</BODY>
</HTML>

--=-vXE1v7EUBZeVSDhuO0F4--

Report message to a moderator

Re: Environment and name resolving [message #57769 is a reply to message #57578]

Tue, 03 June 2008 18:46

Felix Dorner

Messages: 676
Registered: July 2009

Senior Member

Hi Christian,

the whole issue is that I don't understand how instance specifications
are handled. For example Enumeration literals. These are instances of
their Enum type and according to the OCL spec accessed, e.g. via
Gender::male or Gender::female.

Now what if that Enumeration defines additional features, say a query
operation "hasBrain()". How would I call that operation:

Gender::male::hasBrain()

or

Gender::male.hasBrain()

?

Thanks again for your help.

Felix

Report message to a moderator

Re: Environment and name resolving [message #57794 is a reply to message #57769]

Tue, 03 June 2008 18:58

Eclipse User

Originally posted by: cdamus.zeligsoft.com

--=-Z047iOkjuDjRUyYtjfWK
Content-Type: text/plain
Content-Transfer-Encoding: 7bit

Hi, Felix,

If hasBrain is a static operation, then you would do:

Gender::hasBrain()

or (though discouraged because it looks like Gender might be a variable;
the OCL spec allows both forms)

Gender.hasBrain()

If hasBrain is not static, then you must do:

Gender::male.hasBrain()

HTH,

Christian

On Tue, 2008-06-03 at 20:46 +0200, Felix Dorner wrote:

> Hi Christian,
>
> the whole issue is that I don't understand how instance specifications
> are handled. For example Enumeration literals. These are instances of
> their Enum type and according to the OCL spec accessed, e.g. via
> Gender::male or Gender::female.
>
> Now what if that Enumeration defines additional features, say a query
> operation "hasBrain()". How would I call that operation:
>
> Gender::male::hasBrain()
>
> or
>
> Gender::male.hasBrain()
>
> ?
>
>
> Thanks again for your help.
>
> Felix

--=-Z047iOkjuDjRUyYtjfWK
Content-Type: text/html; charset=utf-8

<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.0 TRANSITIONAL//EN">
<HTML>
<HEAD>
<META HTTP-EQUIV="Content-Type" CONTENT="text/html; CHARSET=UTF-8">
<META NAME="GENERATOR" CONTENT="GtkHTML/3.16.0">
</HEAD>
<BODY>
Hi, Felix, 
 
If hasBrain is a static operation, then you would do: 
 
  Gender::hasBrain() 
 
or (though discouraged because it looks like Gender might be a variable; the OCL spec allows both forms) 
 
  Gender.hasBrain() 
 
If hasBrain is not static, then you must do: 
 
  Gender::male.hasBrain() 
 
HTH, 
 
Christian 
 
On Tue, 2008-06-03 at 20:46 +0200, Felix Dorner wrote:
<BLOCKQUOTE TYPE=CITE>
<PRE>
Hi Christian,

the whole issue is that I don't understand how instance specifications 
are handled. For example Enumeration literals. These are instances of 
their Enum type and according to the OCL spec accessed, e.g. via 
Gender::male or Gender::female.

Now what if that Enumeration defines additional features, say a query 
operation "hasBrain()". How would I call that operation:

Gender::male::hasBrain()

or

Gender::male.hasBrain()

?

Thanks again for your help.

Felix
</PRE>
</BLOCKQUOTE>
</BODY>
</HTML>

--=-Z047iOkjuDjRUyYtjfWK--

Report message to a moderator

Re: Environment and name resolving [message #57818 is a reply to message #57794]

Tue, 03 June 2008 19:11

Felix Dorner

Messages: 676
Registered: July 2009

Senior Member

Christian W. Damus wrote:
> Hi, Felix,
>
> If hasBrain is a static operation, then you would do:
>
> Gender::hasBrain()
>
> or (though discouraged because it looks like Gender might be a variable;
> the OCL spec allows both forms)
>
> Gender.hasBrain()
>
> If hasBrain is not static, then you must do:
>
> Gender::male.hasBrain()

Ok. I just found out how to use the OCL console and verified this
myself. I guess the selected element provides the context for the
expressions I enter? I have a class and an Enum both in the same (root)
package. If I select the class, then then "Gender::male.hasBrain()"
doesnt resolve. When I do it fully qualified it does
("root::Gender::male.hasBrain()"): So do those "partially qualified
paths" only work if the path elements (without the tail) are all packages?

Felix

Report message to a moderator

Re: Environment and name resolving [message #57842 is a reply to message #57818]

Tue, 03 June 2008 22:35

Eclipse User

Originally posted by: cdamus.zeligsoft.com

--=-DSm0WKGDGLsoYq499CUK
Content-Type: text/plain
Content-Transfer-Encoding: 7bit

Hi, Felix,

Does your OCL Console show M2 as the current modeling level?

You want M1 for OCL expressions in the context of classifiers in your
model. Otherwise, your OCL expressions are in the context of
classifiers in the UML metamodel. The fully qualified reference to the
Gender type worked because OCL was able to jump from the UML scope to
"root" scope by searching the resource set for top-level namespaces.

Note that, for parsing expressions at M1 level, you will also need to
have the target metamodel set to UML (not Ecore) if you are working with
a UML model.

Cheers,

Christian

On Tue, 2008-06-03 at 21:11 +0200, Felix Dorner wrote:

> Christian W. Damus wrote:
> > Hi, Felix,
> >
> > If hasBrain is a static operation, then you would do:
> >
> > Gender::hasBrain()
> >
> > or (though discouraged because it looks like Gender might be a variable;
> > the OCL spec allows both forms)
> >
> > Gender.hasBrain()
> >
> > If hasBrain is not static, then you must do:
> >
> > Gender::male.hasBrain()
>
> Ok. I just found out how to use the OCL console and verified this
> myself. I guess the selected element provides the context for the
> expressions I enter? I have a class and an Enum both in the same (root)
> package. If I select the class, then then "Gender::male.hasBrain()"
> doesnt resolve. When I do it fully qualified it does
> ("root::Gender::male.hasBrain()"): So do those "partially qualified
> paths" only work if the path elements (without the tail) are all packages?
>
> Felix

--=-DSm0WKGDGLsoYq499CUK
Content-Type: text/html; charset=utf-8

<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.0 TRANSITIONAL//EN">
<HTML>
<HEAD>
<META HTTP-EQUIV="Content-Type" CONTENT="text/html; CHARSET=UTF-8">
<META NAME="GENERATOR" CONTENT="GtkHTML/3.16.0">
</HEAD>
<BODY>
Hi, Felix, 
 
Does your OCL Console show M2 as the current modeling level? 
 
You want M1 for OCL expressions in the context of classifiers in your model.  Otherwise, your OCL expressions are in the context of classifiers in the UML metamodel.  The fully qualified reference to the Gender type worked because OCL was able to jump from the UML scope to "root" scope by searching the resource set for top-level namespaces. 
 
Note that, for parsing expressions at M1 level, you will also need to have the target metamodel set to UML (not Ecore) if you are working with a UML model. 
 
Cheers, 
 
Christian 
 
 
On Tue, 2008-06-03 at 21:11 +0200, Felix Dorner wrote:
<BLOCKQUOTE TYPE=CITE>
<PRE>
Christian W. Damus wrote:
> Hi, Felix,
> 
> If hasBrain is a static operation, then you would do:
> 
> Gender::hasBrain()
> 
> or (though discouraged because it looks like Gender might be a variable; 
> the OCL spec allows both forms)
> 
> Gender.hasBrain()
> 
> If hasBrain is not static, then you must do:
> 
> Gender::male.hasBrain()

Ok. I just found out how to use the OCL console and verified this 
myself. I guess the selected element provides the context for the 
expressions I enter? I have a class and an Enum both in the same (root) 
package. If I select the class, then then "Gender::male.hasBrain()" 
doesnt resolve. When I do it fully qualified it does 
("root::Gender::male.hasBrain()"): So do those "partially qualified 
paths" only work if the path elements (without the tail) are all packages?

Felix
</PRE>
</BLOCKQUOTE>
</BODY>
</HTML>

--=-DSm0WKGDGLsoYq499CUK--

Report message to a moderator

Previous Topic:	Re: Is it possible to alias types
Next Topic:	[Announce] MDT OCL 1.2.0RC3 is available

Goto Forum:

-=] Back to Top [=-

[ Syndicate this forum (XML) ] [

]

Current Time: Thu Apr 18 01:46:20 GMT 2024

.:: Contact :: Home ::.

Breadcrumbs

Sign up to our Newsletter