Show Posts

This section allows you to view all posts made by this member. Note that you can only see posts made in areas you currently have access to.


Messages - tanja

Pages: 1 [2] 3 4 5
16
General Board / Baseline diff export?
« on: September 06, 2010, 06:32:56 pm »
Is there anything like this? For instance, an XML file (with tags like added, deleted, modified, moved, etc.)? Or at least a csv file?

I know there is the export of the whole baseline as XMI, but then I would have to compare two XMIs...

17
General Board / Re: How to exclude fragment from RTF document?
« on: March 12, 2010, 04:33:04 am »
Perfect - it works.

Fragment is called InteractionFragment.

Thanks.

18
General Board / How to exclude fragment from RTF document?
« on: March 11, 2010, 08:11:14 pm »
I have not found a solution on the forum and would not like to go manual way (with old-style rtf).

I use loop fragments on sequence diagrams to convey the semantics of something repetitive, but the fragment in itself has no other meaning, so I'd be happy to not have it printed as child element (in a header, with empty doc).

Is there a way?

19
General Board / Re: Diagram: Character length of inital/default Va
« on: March 22, 2009, 10:48:59 pm »
This seems to be a pretty serious limitation. We need to display version name and cannot fit it to 20 characters, so the very important info gets lost on the diagram.
I hope that this limitation gets fixed soon...

20
General Board / Project-specific settings for format
« on: October 11, 2008, 05:26:05 am »
Is there a way to define and store within the model file project specific fonts and colours? What I want to achieve is that whoever on the team opens the shared .eap file will see the diagrams exactly the same way (and not with own preferences) and any of us can produce report that will look exactly the same.

Currently, when I generate report (rtf), the diagrams are with the styles (colours/fonts) I have customised on my installed version of EA (e.g., white fill, no shadow, black text and lines for all). Another colleague didn't modify any of the settings related to diagrams look, so the reports he generates are with default EA settings (light rose fill, red/green text for class attributes/operations, Arial 10 fonts, etc.) You can then imagine that his diagrams get deformed when I open the file (and mines get deformed when he opens the file) - and all the work one of us does (to fit stuff on one page, the distance and aligning, etc.) is lost...

Please, don't point me to project colours get/set, or status colors settings - I know those but they don't fulfill our requirements.

Thanks in advance.

21
General Board / Need to uninstall EA5 before installing EA6?
« on: January 18, 2006, 02:29:15 am »
On the registered user page, there are 2 'use cases':
- if you use evaluation version
- if you have an installed registered version and want to patch it

However, it is not stated whether I must uninstall an old release version (e.g., EA5) and reboot in order to correctly install the new release version (e.g., EA6).

Many applications do not require this and thus say nothing about uninstalling. But, from some search on this forum, I see that some people uninstall EA5, reboot, then install EA6.

If the sequence 'uninstall old->reboot->install new' is the requirement, could somebody please provide an explicit statement on that in the installation guide section of download site (and to this post :-).

Thanx!

22
General Board / Re: java forward/reverse/synch
« on: November 01, 2005, 01:51:06 am »
I know it would work with bi-directional, but then I'm modifying the original model (an international standard that shouldn't be modified), which has *unspecified* directions.

The people in the standards committee were considering moving from Rose to some other tool, but this might be a stopper for them to use EA ...

Thanx anyway for replies !

23
General Board / Re: java forward/reverse/synch
« on: October 31, 2005, 01:06:09 am »
Thanx for so quick feedback!

I've retried yesterday all three cases after I sent out the posting, by first deleting completely an existing association in the model.

Worth mentioning is that in Synch options, I keep existing associations, if I remember well, I did so because I've seen some associations disappear in the very beginning.

So, here the results when doing synch from code -> model, where I first delete the association in question from the model and let EA generate them from the code:

1) with a java collection, EA creates ONE (not two) directed association. If I make it, by hand, of unspecified direction, and I add the other (missing) association end name and cardinalities, and then repeat the synch, the association stays as I modified it, i.e., as I would like to have them all (this is probably because of the synch local option to not remove existing associations).

2, 3) with a java array or a single valued field, EA creates 2 uni-directional associations. If I try, by hand, to remove one and adjust the remaining one as above (make it unspecified direction, add cardinalities, add the name of the other association end), then synch again, EA keeps what I modified but also adds a single uni-directional association that I've deleted by hand.

What else to try?

24
General Board / Re: java forward/reverse/synch
« on: October 30, 2005, 03:22:58 am »
Sorry for my late reply, but I was away for a few weeks.

My problem is related to associations handling when forward and reverse engineering.

Background:
-----------------
The original model was in Rational Rose. All the associations were with unspecified direction, but both ends of each and every association have a name. This means that, in the diagram there are no arrows at any association end, but the associations are in fact bi-directional (as both ends are named). If class A has a member of class B, then B also has a member of A (you can extrapolate this to any cardinality).

Yes, I know that the 'unspecified' direction is inherently wrong if bi-directional is meant, but I cannot change the modellers who are in a standards' committee :-), while I have to deal with the model they produced. Since they've been using Rose for modelling, and in Rose associations are by default with unspecified direction, all the associations are so. The model itself is an ontology, so it is important to have names at both ends of each association. Also, the standard does not want to explicitly state all the directions, in order to allow for implementation flexibility.

From Rose to EA:
---------------------
Now, I've exported Rose model into XMI and imported into EA - so far so good.

In EA - code generation and manual fixes:
-----------------------------------------------------
I started forward engineering from UML to java and stumbled over several problems. Some of them I could solve by manually fixing problems in each and every of about 500 files (now you understand why I was a bit irritated in my first posting):
- imports were missing in .java files
- for each association, EA generated member variables for only one class; e.g., for association between A and B, A had a member of type, but B had not a member of type A.

In EA - reverse engineering/synch and what I couldn't fix:
------------------------------------------------------------------------
After the above issues have been fixed, and the whole code could compile, I've started reverse engineering from the code into the model: In project browser, I select a package, and from its context menu (right mouse click) I select Code Engineering -> Synchronise package contents -> Reverse engineer (source->model). This is now my only mode of operation, since I have my code as master source.

What I would like to preserve in the UML model is the ***single association of unspecified direction*** in the case of each two classes A and B pointing to each other, for every possible cardinality implementation (collection, array, single value). This is what I mean by "normal".

My problem is that EA does synchronisation differently as a function of what I use in my code to implement the cardinalities. Here 3 examples:

1) The case of a java collection at any or both sides produces (or better said: preserves) the single association of unspecified direction in the model. Example:
public class A {
 private ArrayList<B> myBs; // card. [1..*] or [0..*]
}
public class B {
 private A myA;  // card. [1] or [0..1]
}

2) The case of a java array DOES NOT preserve the original association in the model.  It creates two uni-directional associations. Example:
public class A {
 private B[] myBs; // card. [1..*] or [0..*]
}
public class B {
 private A myA;  // card. [1] or [0..1]
}

3) The case of a single value at both sides DOES NOT preserve the original association in the model. It creates two uni-directional associations. Example:
public class A {
 private B myB; // card. [1] or [0..1]
}
public class B {
 private A myA; // card. [1] or [0..1]
}

So, please tell me how I could make the cases 2) and 3) behave as case 1)? In particular, the non-coherent result produced by EA for the 3 cases above is somewhat disturbing, and limits me in the implementation choice if I want to preserve the model.

Hope this one was not "Houston, we have a problem" :-)

Thanx in advance!

25
General Board / java forward/reverse/synch
« on: October 05, 2005, 06:47:23 am »
Hello,

I've posted a topic at <http://www.sparxsystems.com.au/cgi-bin/yabb/YaBB.pl?board=general;action=display;num=1124737672;start=0#0>, and for the first time since visiting this Forum, got no answers. I concluded that it must be because I was not polite enough :-), but I might be wrong. So, while presenting my excuses for somewhat iritated mood for the previous post, I give it another try, in hope that some kind soul would help me with a workaround, or simply confirm there is an issue.

I've played more with synchronisation from java code into model. In all what follows, the associations have roles explicitly defined on both sides, with cardinalities (i.e., the two classes reference each other). Here the conclusions:
- an array in the code produces 2 uni-directional association in model- I found no way to make them a single, direction agnostic one;
- any java member that is not a Collection (e.g., ArrayList) produces a single uni-directional link (the second one is lost somewhere). This boils down to associations with UML cardinalities 1 and 0..1, i.e., anything that is not considered as a collection in general;
-  the only correct handling of associations between two classes (I mean: you already have a non-directed association with both roles specified) after reverse synch is where at least one role has multiplicity 0..* or 1..*, i.e., where I have java ArrayList in my code (this doesn't work for arrays, as stated above).

It would be more than useful to have an explicit list of mappings for code engineering, which would save some time for developers/modellers.

With hope somebody hears me ...


26
General Board / java 5 enum, rev. engineer (regression for 772)
« on: November 28, 2005, 03:03:00 am »
Hello,

This piece of code:

public enum Operator {
  AND((short) 1), // <<<====
  // rest omitted

at synchronisation from code into model generates the error:

There was an error parsing C:\test\Test.java on line 2.
Unexpected Symbol: )
Would you like to continue the import?

I guess the problem is the cast (short). This very same code has been successfully reverse engineered with some of the previous builds (prior to 772), as shown below (I'm not sure whether this pic will show, but beleive me it was correct, generated with EA):



I summarise the cases from some 30 enum classes where the same kind of error appears:

AND((short) 1),
QUALITY_MASK((int) 0x000000C0L),

The java source code for these enums has been generated from IDL with idl2j compiler, so blame them for casts :-)

Is this going to be fixed soon? In the meantime, is there a workaround?

Thanks in advance,
Tanja

ps: note to Sparx - how about putting strings in error messages surrounded by something? I would have preferred the above error message look so:
Unexpected Symbol: ')'

27
General Board / Re: Actor multiplicity
« on: November 22, 2005, 02:32:21 am »
Thomas, thanx again.

My intent is absolutely NOT to point out the EA is wrong - I just wanted to have the multiplicity displayed for actors. If the UML Bible is against it, I must not complain :-)

It seems that I have no choice, and that I should make all the actors change to stereotyped classes.

Reg's,
Tanja

28
General Board / Re: Actor multiplicity
« on: November 22, 2005, 01:56:49 am »
Thomas,

Thanx for your reply.

What do you mean by "to read through Superstructures"? I do not intend to use automation or similar, I'd just like that EA displays the multiplicity (which is defined on actor element) on a diagram.

Should I conclude that this is impossible?

29
General Board / Actor multiplicity
« on: November 21, 2005, 06:00:09 am »
Hello,

It is possible to define multiplicity for an actor (the same way as for, e.g., class), but I find no way to display that multiplicity on a diagram in the way the multiplicity is displayed for a class.

Any hint would be highly apreciated!

30
General Board / java FE/RE associations
« on: August 22, 2005, 12:07:52 pm »
Hello,

I have a huge model that I've imported from Rose into EA 5. I've managed to generate java code into directories as appropriate (e.g., org.iec.something) and to have EA name the packages appropriately. However, I found several problems:
a) there are no imports generated for the classes from another packages, so I had to do it manually (with eclipse 'Organise imports', but still file by file)
b) for associations (no diamond) and for aggregations (white diamond), only one role is generated, and there is no rule which one with respect to source or target, cardinalities, diamond position - and this is VERY annoying. Again, I had to do this by hand. In Rose, all the associations had unspecified direction and have been kept so in EA. Each and every association has both role names specified, and both side cardinalities. I've also set the option "Generate when direction unspecified".

I've discovered a strange thing with EA: when there is an association (any kind), EA creates an attribute in the class in the model that seems to be linked to the association and thus on a diagram, I have both the association and the attribute - some other tools I know of show either association or attribute, but never both.
(1) Is there a way to show only one? For code engineering, this would mean there are two fields with the same name/type in my class...

What I've been doing now all the time is to synchronise (i.e., reverse engineer) from the code into the model, as I'm keeping the code as master.

Now, two more things I don't know how to solve, and I beg for your help:
(2) How can I get javadoc comments for java class fields (which are those "linked" to associations) in the EA model notes for association roles? Comments get into the EA model only for attributes of the class (those (strangely) fabricated from association roles)...

(3) Among many associations (I estimate some 300), most are kept correctly in the model, after I've added missing attribute in one of the two classes, in java source code. After synchronising (from code into the model), the original cardinalities are still there, one association with unspecified direction is still linking the two classes, the diamond (if originally present) is still there. However, for some 20 associations, I get systematically two uni-directional links - I do not want this, because the direction is not specified and that is the semantics in the model - bad or good, but that's it. This "twinning" happens always when I have cardinality of max 1 (0..1, 1, any combination of these), and for a few other cases (0..n, 1..n). What to do to "merge" these into a single, direction-unspecified association/aggregation?

Any help will be highly apreciated!
Tanja

ps: This post sounds very negative, but I've really spent a lot of time doing things by hand, and the above 3 problems I can't do at all (at least, I don't know how). And I'm only at the begining of the port of Rose models... I've read on other posts that the models imported from Rose XMI when exported into XMI generate huge files (some bug on GUID repeting), so it might be that the strange behaviour in (3) above is due to something similar.

Pages: 1 [2] 3 4 5