Eclipse Community Forums
Forum Search:

Search      Help    Register    Login    Home
Today's Messages (on)  | Unanswered Messages (off)

Forum: C / C++ IDE (CDT)
 Topic: the automatically generated makefile
Re: the automatically generated makefile [message #1413046 is a reply to message #1413009] Wed, 27 August 2014 12:28
David Hall is currently offline David Hall
Messages: 7
Registered: August 2014
Junior Member
Now I find that the only way to have a generated makefile visible is to create a "Makefile Project"

All this time I have been creating an "Executable" project.

So Executable projects don't use makefiles. Is that the way this is??
Forum: Graphiti
 Topic: org.eclipse.graphiti.features.impl.AbstractFeatureProvider.getPictogramElementForBusinessObject(Obje
Re: org.eclipse.graphiti.features.impl.AbstractFeatureProvider.getPictogramElementForBusinessObject( [message #1412941 is a reply to message #1410371] Wed, 27 August 2014 07:56
Michael Wenz is currently offline Michael Wenz
Messages: 1587
Registered: July 2009
Location: Walldorf, Germany
Senior Member
Overriding getAllPictogramElementsForBusinessObject() would be an option,
but maybe also implementing your own logic for that object comparision in
equalsBusinessObjects() in your tool behavior provider would b e sufficient
as well.

Michael
 Topic: Connection Decorator with Text
Re: Connection Decorator with Text [message #1412945 is a reply to message #1412600] Wed, 27 August 2014 08:04
Michael Wenz is currently offline Michael Wenz
Messages: 1587
Registered: July 2009
Location: Walldorf, Germany
Senior Member
Yes, it is. Simply provide a direct editing feature for the connection
decorator with Text. The framework asks the feature provider
(getDirectEditingFeature) for such a feature when the Text of the decorator
is clicked after selection as for any other shape.

Michael
 Topic: Multiple create features
Multiple create features [message #1412921] Wed, 27 August 2014 07:09
John Cole is currently offline John Cole
Messages: 27
Registered: June 2013
Junior Member
Hello everyone,

I've created a ToolBehaviorProvider as described in the tutorial and now I want to add several create features to the palette. More precisely, the user should get the possibility to choose one of many graphical primitives from the palette and draw them in the editor.

What is the way to create several create features so that they are recognized in the ToolBehaviorProvider and shown on the palette?

Many thanks in advance!
Re: Multiple create features [message #1412947 is a reply to message #1412921] Wed, 27 August 2014 08:06
Michael Wenz is currently offline Michael Wenz
Messages: 1587
Registered: July 2009
Location: Walldorf, Germany
Senior Member
Simply provide all the create features in the getCreateFeatures() method of
your feature provider. By default all create features are added to the
palette.

Michael
 Topic: How to set the selection programmatically?
Re: How to set the selection programmatically? [message #1412942 is a reply to message #1411170] Wed, 27 August 2014 07:58
Michael Wenz is currently offline Michael Wenz
Messages: 1587
Registered: July 2009
Location: Walldorf, Germany
Senior Member
Hm, the passed container shape instance must exist on the diagram,
especially it must live in the same editing domain/resource set as the
editor uses. Maybe that is causing trouble here?

Michael
Re: How to set the selection programmatically? [message #1413050 is a reply to message #1412942] Wed, 27 August 2014 12:39
Al B is currently offline Al B
Messages: 41
Registered: June 2012
Member
Yes, all live in the same editing domain and resource set

ContainerShape containerShape = (ContainerShape) context.getPictogramElement();
EditorAssist.getActiveEditor().getDiagramBehavior().setPictogramElementForSelection(containerShape);
Goal goal = (Goal) getBusinessObjectForPictogramElement(containerShape);
		
Object a = EditorAssist.getActiveEditor().getEditingDomain();
Object c = TransactionUtils.getDomain(containerShape);
Object d = TransactionUtils.getDomain(goal);
		
if (a == c && c == d && a == d) {
   System.out.println("YES, they all have the same editing domain and resourse set");

[Updated on: Wed, 27 August 2014 12:39]

Report message to a moderator

Forum: e(fx)clipse
 Topic: CleanUpAddon and DndAddon
CleanUpAddon and DndAddon [message #1413045] Wed, 27 August 2014 12:28
Bastien Bart is currently offline Bastien Bart
Messages: 6
Registered: May 2014
Junior Member
Hi,

I have noticed a strange (?) comportment with the CleanUpAddon :

When you drag the last part from a container and drop it to an other container, the source container disappears because it doesn't contain anymore part. That's OK and works fine for me.

When you close the last part of a container, this last one becomes empty but doesn't disappear and you can't drop anymore part to it. Is it the expected behavior? Or is there something i did wrong?

And I have a question about the DndAddon too :

When you drag a part from a container and move mouse over an other container, the 4 zones (top, bottom, left, right) where you can drop the part appear. Works amazing for me. But the display of the 4 zones happens on menu-bar and tool-bar too (see attachment menu-bar.png). Is it normal?

Thx
Bastien
  • Attachment: menu-bar.png
    (Size: 4.20KB, Downloaded 2 times)
Re: CleanUpAddon and DndAddon [message #1413052 is a reply to message #1413045] Wed, 27 August 2014 12:41
Thomas Schindl is currently offline Thomas Schindl
Messages: 5292
Registered: July 2009
Senior Member
On 27.08.14 18:28, Bastien Bart wrote:
> Hi,
>
> I have noticed a strange (?) comportment with the CleanUpAddon :
>
> When you drag the last part from a container and drop it to an other container, the source container disappears because it doesn't contain anymore part. That's OK and works fine for me.
>
> When you close the last part of a container, this last one becomes empty but doesn't disappear and you can't drop anymore part to it. Is it the expected behavior? Or is there something i did wrong?
>

Not normal behavior! Sounds like a bug.

> And I have a question about the DndAddon too :
>
> When you drag a part from a container and move mouse over an other container, the 4 zones (top, bottom, left, right) where you can drop the part appear. Works amazing for me. But the display of the 4 zones happens on menu-bar and tool-bar too (see attachment menu-bar.png). Is it normal?

Sounds like a bug!

Please file 2 bugzillas. I guess for the 1st one I need a reproduceable
sample. You can start with the sample in
org.eclipse.fx.testcases.dnd.app in our git-repo.

Please note the center stack IS tagged with NoAutoCollapse. One thing I
see missing is a feedback when one drags to an empty stack (feel free to
file a 3rd bugzilla).

Tom
Forum: EGit
 Topic: SymlinksNotSupportedException when resetting to HEAD
SymlinksNotSupportedException when resetting to HEAD [message #1412932] Wed, 27 August 2014 07:44
Guenther Grau is currently offline Guenther Grau
Messages: 2
Registered: August 2014
Junior Member
Hi,

I just downloaded and installed luna on Win 7/Java 8u20. I have imported projects from an existing local git repository. When I try to reset the repository to head, I get an error message. The appended exception is logged in .log

Any idea what could be going wrong here?

Thanx and best regards,

Guenther


!ENTRY org.eclipse.core.jobs 4 2 2014-08-27 13:09:47.396
!MESSAGE An internal error occurred during: "Resetting to HEAD".
!STACK 0
org.eclipse.jgit.api.errors.JGitInternalException: Exception caught during execution of reset command. {0}
at org.eclipse.jgit.api.ResetCommand.call(ResetCommand.java:227)
at org.eclipse.egit.core.op.ResetOperation.reset(ResetOperation.java:97)
at org.eclipse.egit.core.op.ResetOperation.access$0(ResetOperation.java:85)
at org.eclipse.egit.core.op.ResetOperation$1.run(ResetOperation.java:74)
at org.eclipse.core.internal.resources.Workspace.run(Workspace.java:2313)
at org.eclipse.egit.core.op.ResetOperation.execute(ResetOperation.java:78)
at org.eclipse.egit.core.internal.job.JobUtil$2.runInWorkspace(JobUtil.java:105)
at org.eclipse.core.internal.resources.InternalWorkspaceJob.run(InternalWorkspaceJob.java:38)
at org.eclipse.core.internal.jobs.Worker.run(Worker.java:54)
Caused by: org.eclipse.jgit.errors.SymlinksNotSupportedException: Symlinks are not supported with this OS/JRE
at org.eclipse.jgit.util.FS.createSymLink(FS.java:612)
at org.eclipse.jgit.dircache.DirCacheCheckout.checkoutEntry(DirCacheCheckout.java:1203)
at org.eclipse.jgit.dircache.DirCacheCheckout.doCheckout(DirCacheCheckout.java:459)
at org.eclipse.jgit.dircache.DirCacheCheckout.checkout(DirCacheCheckout.java:394)
at org.eclipse.jgit.api.ResetCommand.checkoutIndex(ResetCommand.java:393)
at org.eclipse.jgit.api.ResetCommand.call(ResetCommand.java:200)
... 8 more
Re: SymlinksNotSupportedException when resetting to HEAD [message #1412994 is a reply to message #1412932] Wed, 27 August 2014 10:15
Christian Halstrick is currently offline Christian Halstrick
Messages: 100
Registered: July 2009
Senior Member
It seems you are trying to work on a repository which contains symbolic links on a Windows machine. In such a case JGit should honor core.symlinks and should create text-files (containing the path to the link target) instead of symlinks. This would allow you to at least checkout a project.

In general EGit/JGit seem to work correctly. I can clone https://github.com/chalstrick/differentFileModes.git and switch to branch symLinks on windows. Something most be different on your machine.

Can you tell exactly which repo you cloned and which steps you did to reach the bug?


Ciao
Chris
Re: SymlinksNotSupportedException when resetting to HEAD [message #1413005 is a reply to message #1412994] Wed, 27 August 2014 10:46
Guenther Grau is currently offline Guenther Grau
Messages: 2
Registered: August 2014
Junior Member
Hi Chris,

I cloned https ://github.com/graug/zxing.git sometime March or April 2013 to develop a feature for ZXing. I haven't been working on the project since May 2013. Now that I have some time on my hands, I downloaded Luna and Java 8 and imported the projects from my existing local git repo using the eclipse repository view.
I then open the context menu of the master branch and select rebase. I select the zxing upstream master project https ://github.com/zxing/zxing.git. Eclipse then tells me that I have local changes and asks me what to do with them. It offers me to reset or stash (or commit the changes). Neither of reset or stash works. When I select reset to discard my local changes I get above mentioned error. Stashing my changes also doesn't work. It stashes the changes but opens the same dialog informing me about changes again.
As a workaround I'm creating a new local clone of https ://github.com/graug/zxing.git but this takes forever to complete Sad

Thanx for looking into this and best regards,

Guenther

P.S.: Had to cripple the links because the forum wouldn't let me post URLs except to eclipse.org as I have posted less than 5 messages.
 Topic: JGit version of 'git merge-tree'
JGit version of 'git merge-tree' [message #1413057] Wed, 27 August 2014 13:01
Sneakers Please is currently offline Sneakers Please
Messages: 1
Registered: August 2014
Junior Member
Hi,

I'm trying to get JGit to output the results of a three-way merge but without changing anything in the working directory. The second answer here:

stackoverflow.com/questions/501407/is-there-a-git-merge-dry-run-option

does what I want using git. Essentially:

git merge-tree `git merge-base branch_one branch_two` branch_one branch_two

Is there any way to get JGit to produce equivalent output?

I've got as far as creating a ResolveMerger with inCore = true. This more or less seems to produce the results that I need although I'd need to format the objects into the git-style output myself. Also, it doesn't expose the toBeDeleted member.

Is there an easier way to do this or is this feature missing from JGit?

Thanks!
Forum: M2T (model-to-text transformation)
 Topic: [Acceleo] Problems with Builder in Kepler
Re: [Acceleo] Problems with Builder in Kepler [message #1413035 is a reply to message #1090768] Wed, 27 August 2014 12:01
Marc-Florian Wendland is currently offline Marc-Florian Wendland
Messages: 58
Registered: January 2013
Member
Hi all,

I eventually figured out the reason why the first character gets lost. It
has nothing to do whatsoever with Acceleo compiler settings, encoding
scheme, Eclipse or Java version or XMI parsing. The very reason for this is
a technical deficiency of Acceleo to load/work with metamodels that are
located in the workspace and not deployed in the Eclipse instance. After
removing each and every reference to a locally available metamodel (I simply
deployed the metamodels in my development environment) everything is fine
with Acceleo.

I debugged the location were the first character (in my case, instead of
TestPlan Acceleo looked for estPlan) gets lost. It happens in this method:

org.eclipse.emf.ecore.impl.BasicEObjectImpl.eObjectForURIFragmentSegment(String
uriFragmentSegment), lines 507 and 511

In line 551, the first character is actually cut off.

Again, this problem does not occur if one only worked with deployed models
in the Acceleo modules. Any reference to a locally deployed model is
problematic.

Is this a known issue? Shall I add a critical (I really think such
fundamental issue is critical to Acceleo) Bugzilla for this problem.

Best regards,
Marc-Florian


"Ed Willink" wrote in message news:kv0472$u3i$1@xxxxxxxxe.org...

Hi

This is the missing first letter problem. "Set(Parameter)" would have
been sensible.

Something somewhere has got a byte offset and it causes chaos. The
problem eventually goes away once you rebuild everything enough times.
I've not managed to work out where the inconsistency is.

Try cleaning, restarting, wiping the bin folders with your OS, switch
backwards and forwards between binary/ASCII emtl files, changing JRE to
some other version and back again, make sure your Kepler compiler
preferences are the same as on Indigo ...

Regards

Ed Willink


On 20/08/2013 15:29, Marc-Florian Wendland wrote:
> Dear all,
>
> I am migrating our software from Indigo to Kepler and I encountered the
> following error when trying to build the Acceleo project:
>
> Errors occurred during the build.
> Errors running builder 'Acceleo Builder' on project 'xxx'.
> The feature 'et(Parameter)' is not a valid feature
>
> As a result I do not have an output files generated.
>
> I cannot interpret this error message at all. Again, the very same project
> worked in Indigo.
>
> Thanks for any help in advance
> Marc-Florian
Re: [Acceleo] Problems with Builder in Kepler [message #1413036 is a reply to message #1090768] Wed, 27 August 2014 12:01
Marc-Florian Wendland is currently offline Marc-Florian Wendland
Messages: 58
Registered: January 2013
Member
Hi all,

I eventually figured out the reason why the first character gets lost. It
has nothing to do whatsoever with Acceleo compiler settings, encoding
scheme, Eclipse or Java version or XMI parsing. The very reason for this is
a technical deficiency of Acceleo to load/work with metamodels that are
located in the workspace and not deployed in the Eclipse instance. After
removing each and every reference to a locally available metamodel (I simply
deployed the metamodels in my development environment) everything is fine
with Acceleo.

I debugged the location were the first character (in my case, instead of
TestPlan Acceleo looked for estPlan) gets lost. It happens in this method:

org.eclipse.emf.ecore.impl.BasicEObjectImpl.eObjectForURIFragmentSegment(String
uriFragmentSegment), lines 507 and 511

In line 551, the first character is actually cut off.

Again, this problem does not occur if one only worked with deployed models
in the Acceleo modules. Any reference to a locally deployed model is
problematic.

Is this a known issue? Shall I add a critical (I really think such
fundamental issue is critical to Acceleo) Bugzilla for this problem.

Best regards,
Marc-Florian


"Ed Willink" wrote in message news:kv0472$u3i$1@xxxxxxxxe.org...

Hi

This is the missing first letter problem. "Set(Parameter)" would have
been sensible.

Something somewhere has got a byte offset and it causes chaos. The
problem eventually goes away once you rebuild everything enough times.
I've not managed to work out where the inconsistency is.

Try cleaning, restarting, wiping the bin folders with your OS, switch
backwards and forwards between binary/ASCII emtl files, changing JRE to
some other version and back again, make sure your Kepler compiler
preferences are the same as on Indigo ...

Regards

Ed Willink


On 20/08/2013 15:29, Marc-Florian Wendland wrote:
> Dear all,
>
> I am migrating our software from Indigo to Kepler and I encountered the
> following error when trying to build the Acceleo project:
>
> Errors occurred during the build.
> Errors running builder 'Acceleo Builder' on project 'xxx'.
> The feature 'et(Parameter)' is not a valid feature
>
> As a result I do not have an output files generated.
>
> I cannot interpret this error message at all. Again, the very same project
> worked in Indigo.
>
> Thanks for any help in advance
> Marc-Florian
Re: [Acceleo] Problems with Builder in Kepler [message #1413037 is a reply to message #1090768] Wed, 27 August 2014 12:02
Marc-Florian Wendland is currently offline Marc-Florian Wendland
Messages: 58
Registered: January 2013
Member
Hi all,

I eventually figured out the reason why the first character gets lost. It
has nothing to do whatsoever with Acceleo compiler settings, encoding
scheme, Eclipse or Java version or XMI parsing. The very reason for this is
a technical deficiency of Acceleo to load/work with metamodels that are
located in the workspace and not deployed in the Eclipse instance. After
removing each and every reference to a locally available metamodel (I simply
deployed the metamodels in my development environment) everything is fine
with Acceleo.

I debugged the location were the first character (in my case, instead of
TestPlan Acceleo looked for estPlan) gets lost. It happens in this method:

org.eclipse.emf.ecore.impl.BasicEObjectImpl.eObjectForURIFragmentSegment(String
uriFragmentSegment), lines 507 and 511

In line 551, the first character is actually cut off.

Again, this problem does not occur if one only worked with deployed models
in the Acceleo modules. Any reference to a locally deployed model is
problematic.

Is this a known issue? Shall I add a critical (I really think such
fundamental issue is critical to Acceleo) Bugzilla for this problem.

Best regards,
Marc-Florian


"Ed Willink" wrote in message news:kv0472$u3i$1@xxxxxxxxe.org...

Hi

This is the missing first letter problem. "Set(Parameter)" would have
been sensible.

Something somewhere has got a byte offset and it causes chaos. The
problem eventually goes away once you rebuild everything enough times.
I've not managed to work out where the inconsistency is.

Try cleaning, restarting, wiping the bin folders with your OS, switch
backwards and forwards between binary/ASCII emtl files, changing JRE to
some other version and back again, make sure your Kepler compiler
preferences are the same as on Indigo ...

Regards

Ed Willink


On 20/08/2013 15:29, Marc-Florian Wendland wrote:
> Dear all,
>
> I am migrating our software from Indigo to Kepler and I encountered the
> following error when trying to build the Acceleo project:
>
> Errors occurred during the build.
> Errors running builder 'Acceleo Builder' on project 'xxx'.
> The feature 'et(Parameter)' is not a valid feature
>
> As a result I do not have an output files generated.
>
> I cannot interpret this error message at all. Again, the very same project
> worked in Indigo.
>
> Thanks for any help in advance
> Marc-Florian
Re: [Acceleo] Problems with Builder in Kepler [message #1413038 is a reply to message #1090768] Wed, 27 August 2014 12:02
Marc-Florian Wendland is currently offline Marc-Florian Wendland
Messages: 58
Registered: January 2013
Member
Hi all,

I eventually figured out the reason why the first character gets lost. It
has nothing to do whatsoever with Acceleo compiler settings, encoding
scheme, Eclipse or Java version or XMI parsing. The very reason for this is
a technical deficiency of Acceleo to load/work with metamodels that are
located in the workspace and not deployed in the Eclipse instance. After
removing each and every reference to a locally available metamodel (I simply
deployed the metamodels in my development environment) everything is fine
with Acceleo.

I debugged the location were the first character (in my case, instead of
TestPlan Acceleo looked for estPlan) gets lost. It happens in this method:

org.eclipse.emf.ecore.impl.BasicEObjectImpl.eObjectForURIFragmentSegment(String
uriFragmentSegment), lines 507 and 511

In line 551, the first character is actually cut off.

Again, this problem does not occur if one only worked with deployed models
in the Acceleo modules. Any reference to a locally deployed model is
problematic.

Is this a known issue? Shall I add a critical (I really think such
fundamental issue is critical to Acceleo) Bugzilla for this problem.

Best regards,
Marc-Florian


"Ed Willink" wrote in message news:kv0472$u3i$1@xxxxxxxxe.org...

Hi

This is the missing first letter problem. "Set(Parameter)" would have
been sensible.

Something somewhere has got a byte offset and it causes chaos. The
problem eventually goes away once you rebuild everything enough times.
I've not managed to work out where the inconsistency is.

Try cleaning, restarting, wiping the bin folders with your OS, switch
backwards and forwards between binary/ASCII emtl files, changing JRE to
some other version and back again, make sure your Kepler compiler
preferences are the same as on Indigo ...

Regards

Ed Willink


On 20/08/2013 15:29, Marc-Florian Wendland wrote:
> Dear all,
>
> I am migrating our software from Indigo to Kepler and I encountered the
> following error when trying to build the Acceleo project:
>
> Errors occurred during the build.
> Errors running builder 'Acceleo Builder' on project 'xxx'.
> The feature 'et(Parameter)' is not a valid feature
>
> As a result I do not have an output files generated.
>
> I cannot interpret this error message at all. Again, the very same project
> worked in Indigo.
>
> Thanks for any help in advance
> Marc-Florian
Re: [Acceleo] Problems with Builder in Kepler [message #1413064 is a reply to message #1413038] Wed, 27 August 2014 13:34
Ed Willink is currently offline Ed Willink
Messages: 4033
Registered: July 2009
Senior Member
Hi

Well-debugged.

A few users have encountered this.

It's certainly really irritating when it happens, but it does seem to go
away after enough rebuild/re-install cycles, so perhaps only major.

Absolutely worth a Bugzilla.

Regards

Ed Willink



On 27/08/2014 17:02, Marc-Florian Wendland wrote:
> Hi all,
>
> I eventually figured out the reason why the first character gets lost. It
> has nothing to do whatsoever with Acceleo compiler settings, encoding
> scheme, Eclipse or Java version or XMI parsing. The very reason for this is
> a technical deficiency of Acceleo to load/work with metamodels that are
> located in the workspace and not deployed in the Eclipse instance. After
> removing each and every reference to a locally available metamodel (I
> simply
> deployed the metamodels in my development environment) everything is fine
> with Acceleo.
>
> I debugged the location were the first character (in my case, instead of
> TestPlan Acceleo looked for estPlan) gets lost. It happens in this method:
>
> org.eclipse.emf.ecore.impl.BasicEObjectImpl.eObjectForURIFragmentSegment(String
>
> uriFragmentSegment), lines 507 and 511
>
> In line 551, the first character is actually cut off.
>
> Again, this problem does not occur if one only worked with deployed models
> in the Acceleo modules. Any reference to a locally deployed model is
> problematic.
>
> Is this a known issue? Shall I add a critical (I really think such
> fundamental issue is critical to Acceleo) Bugzilla for this problem.
>
> Best regards,
> Marc-Florian
>
>
> "Ed Willink" wrote in message news:kv0472$u3i$1@xxxxxxxxe.org...
>
> Hi
>
> This is the missing first letter problem. "Set(Parameter)" would have
> been sensible.
>
> Something somewhere has got a byte offset and it causes chaos. The
> problem eventually goes away once you rebuild everything enough times.
> I've not managed to work out where the inconsistency is.
>
> Try cleaning, restarting, wiping the bin folders with your OS, switch
> backwards and forwards between binary/ASCII emtl files, changing JRE to
> some other version and back again, make sure your Kepler compiler
> preferences are the same as on Indigo ...
>
> Regards
>
> Ed Willink
>
>
> On 20/08/2013 15:29, Marc-Florian Wendland wrote:
>> Dear all,
>>
>> I am migrating our software from Indigo to Kepler and I encountered the
>> following error when trying to build the Acceleo project:
>>
>> Errors occurred during the build.
>> Errors running builder 'Acceleo Builder' on project 'xxx'.
>> The feature 'et(Parameter)' is not a valid feature
>>
>> As a result I do not have an output files generated.
>>
>> I cannot interpret this error message at all. Again, the very same
>> project
>> worked in Indigo.
>>
>> Thanks for any help in advance
>> Marc-Florian
>
Forum: EclipseLink
 Topic: ElipseLink configuration problem in E4 RCP app
Re: ElipseLink configuration problem in E4 RCP app [message #1412834 is a reply to message #1412711] Wed, 27 August 2014 03:26
Don Smyth is currently offline Don Smyth
Messages: 33
Registered: April 2013
Member
Doh! I changed the log4jdbc.DriverSpy to the hsqldb jdbc driver and all works fine.
 Topic: Right outer join
Right outer join [message #1413076] Wed, 27 August 2014 14:27
Gowthami Shanmugam is currently offline Gowthami Shanmugam
Messages: 2
Registered: August 2014
Junior Member
Hi,

I read few forum and found that eclipse link 2.4 implementation of PA will not right outer join.

Can some one confirm
 Topic: Left join with where condition
Left join with where condition [message #1413080] Wed, 27 August 2014 14:38
Gowthami Shanmugam is currently offline Gowthami Shanmugam
Messages: 2
Registered: August 2014
Junior Member
Hi,

I'm trying to use left join as below in eclipse link 2.4.

Query:

select mt.id from MemberTransaction as mt left join mt.pointsTransaction as ptr where ptr.memberTransaction.id IS NULL


and the generated SQL query is:

SELECT * FROM (SELECT /*+ FIRST_ROWS */ a.*, ROWNUM rnum FROM (
SELECT t0.MEMBER_TRANSACTION_ID AS a1
FROM RS.MEMBER_TRANSACTION t0
LEFT OUTER JOIN RS.POINTS_TRANSACTION t1 ON (t1.MEMBER_TRANSACTION_ID = t0.MEMBER_TRANSACTION_ID),
RS.MEMBER_TRANSACTION t2
WHERE ((t2.MEMBER_TRANSACTION_ID IS NULL) AND (t2.MEMBER_TRANSACTION_ID = t1.MEMBER_TRANSACTION_ID))) a WHERE ROWNUM <= ?) WHERE rnum > ?

If you look into the query in where condition "t2.MEMBER_TRANSACTION_ID = t1.MEMBER_TRANSACTION_ID" was added but it was not in my actual query.

When i did research and found that "ptr.memberTransaction.id IS NULL" will lead to add condition "t2.MEMBER_TRANSACTION_ID = t1.MEMBER_TRANSACTION_ID".

I'm not sure why eclipse link is behaving different. i tried in hibernate and got the actual result without any additional condition.

Is it a bug or the expected behavior.
Forum: NatTable
 Topic: TreeGrid question
TreeGrid question [message #1412876] Wed, 27 August 2014 05:16
Arjan Kok is currently offline Arjan Kok
Messages: 31
Registered: July 2009
Member
Hi,

I am trying to use the TreeGrid based on the provided TreeGridExample.

However, the table does not show the tree as expected : the printTree function prints the tree correct, but the table does not.

Any idea what could be the problem?

Thanks,
Arjan.
Re: TreeGrid question [message #1412883 is a reply to message #1412876] Wed, 27 August 2014 05:31
Dirk Fauth is currently offline Dirk Fauth
Messages: 1262
Registered: July 2012
Senior Member
How should I know what the problem is if you are only telling me that it doesn't show correctly? What is not shown correctly? How does it look like?

Maybe simply the wrong painters are registered.
Re: TreeGrid question [message #1412891 is a reply to message #1412883] Wed, 27 August 2014 05:53
Arjan Kok is currently offline Arjan Kok
Messages: 31
Registered: July 2009
Member
I have attached two images : first the output from the printTree function:

index.php/fa/18952/0/

And second the output as rendered by the Nattable:

index.php/fa/18953/0/


Kind Regards,

Arjan.
  • Attachment: text.gif
    (Size: 15.48KB, Downloaded 17 times)
  • Attachment: table.gif
    (Size: 54.77KB, Downloaded 17 times)
Re: TreeGrid question [message #1412904 is a reply to message #1412891] Wed, 27 August 2014 06:30
Dirk Fauth is currently offline Dirk Fauth
Messages: 1262
Registered: July 2012
Senior Member
I don't know. Maybe a combination of wrong rendering of leafs and sorting. I can not answer that with the given information.
 Topic: Expand singe top row in tree grid
Expand singe top row in tree grid [message #1413049] Wed, 27 August 2014 12:37
Testr t is currently offline Testr t
Messages: 93
Registered: April 2013
Member
Hi,

With Nattable 1.1.0 i am able to use the ExpandAll and CollapseAll commands. Thanks. Also for the tree grid, is it possible to expand and collapse a single parent row? Lets say I have a tree grid with 10 top level rows (and each top row has nested sub rows with 3 or more child rows). I want to expand only row 3 programatically? Is there an option to do that? I am providing an option to expand and collapse all using context menus. There is also a requirement where for some rows i have to add option to expand that row alone.

Re: Expand singe top row in tree grid [message #1413078 is a reply to message #1413049] Wed, 27 August 2014 14:31
Dirk Fauth is currently offline Dirk Fauth
Messages: 1262
Registered: July 2012
Senior Member
No, that feature is currently not provided by NatTable. But technically you should be able to create a command handler that takes the row index and expands that node and all child nodes.
Forum: TMF (Xtext)
 Topic: References across files
Re: References across files [message #1412795 is a reply to message #1412521] Wed, 27 August 2014 00:50
Konstantin Stupnik is currently offline Konstantin Stupnik
Messages: 8
Registered: July 2009
Junior Member
Most definite culprit was line:
content-types/org.eclipse.core.runtime.text/file-extensions=def
in .plugins/org.eclipse.core.runtime/.settings/org.eclipse.core.runtime.pre
But for some reason removing this line alone didn't help. But removing this file and org.eclipse.core.resources.prefs with org.eclipse.core.variables.prefs as well helped (nothing really suspicious was in these files).

Christian thank you for your help anyway Smile

p.s. Is there a relatively easy way to exclude files from indexing?
Re: References across files [message #1412809 is a reply to message #1412795] Wed, 27 August 2014 01:49
Christian Dietrich is currently offline Christian Dietrich
Messages: 6225
Registered: July 2009
Senior Member
https://www.eclipse.org/forums/index.php/m/1405243/?srch=xtext+exclude#msg_1405243
 Topic: code generation using xtend
Re: code generation using xtend [message #1412807 is a reply to message #1412739] Wed, 27 August 2014 01:45
Christian Dietrich is currently offline Christian Dietrich
Messages: 6225
Registered: July 2009
Senior Member
hi,

i dont understand your question.
it is your grammar. you know your target environment (the c code to be generated)
this you have to write the xtend code that does the transformation.

do you have any specific question?
Re: code generation using xtend [message #1412813 is a reply to message #1412807] Wed, 27 August 2014 02:02
Christian Dietrich is currently offline Christian Dietrich
Messages: 6225
Registered: July 2009
Senior Member
btw shouldnt it be

def bodycompile(Resource it)
{
'''
//this is the body
'''


}


p.s. maybe some basic reading on xtend-lang.org may help,
specially https://www.eclipse.org/xtend/documentation.html#templates

[Updated on: Wed, 27 August 2014 02:02]

Report message to a moderator

 Topic: Marker id nnn not found
Marker id nnn not found [message #1412910] Wed, 27 August 2014 06:43
Thomas Hallgren is currently offline Thomas Hallgren
Messages: 3229
Registered: July 2009
Senior Member
I see the exception below from time to time when using our Xtext based editor (Geppetto). Our own code is not on the trace and so far I've not found the cause. Does anyone have a clue as to why this is happening?

I'm using Xtext 2.6.2.

!MESSAGE org.eclipse.xtext.ui.editor.validation.MarkerIssueProcessor  - Could not create marker.

!STACK 0
java.lang.reflect.InvocationTargetException
	at org.eclipse.ui.actions.WorkspaceModifyOperation.run(WorkspaceModifyOperation.java:121)
	at org.eclipse.xtext.ui.editor.validation.MarkerIssueProcessor.processIssues(MarkerIssueProcessor.java:58)
	at org.eclipse.xtext.ui.editor.validation.ValidationJob.run(ValidationJob.java:67)
	at org.eclipse.core.internal.jobs.Worker.run(Worker.java:54)
Caused by: org.eclipse.core.internal.resources.ResourceException: Marker id 219 not found.
	at org.eclipse.core.internal.resources.Marker.checkInfo(Marker.java:57)
	at org.eclipse.core.internal.resources.Marker.setAttribute(Marker.java:237)
	at org.eclipse.xtext.ui.editor.validation.MarkerCreator.setMarkerAttributes(MarkerCreator.java:45)
	at org.eclipse.xtext.ui.editor.validation.MarkerCreator.createMarker(MarkerCreator.java:31)
	at org.eclipse.xtext.ui.editor.validation.AddMarkersOperation.execute(AddMarkersOperation.java:106)
	at org.eclipse.ui.actions.WorkspaceModifyOperation$1.run(WorkspaceModifyOperation.java:106)
	at org.eclipse.core.internal.resources.Workspace.run(Workspace.java:2313)
	at org.eclipse.ui.actions.WorkspaceModifyOperation.run(WorkspaceModifyOperation.java:118)
	... 3 more
Root exception:
org.eclipse.core.internal.resources.ResourceException: Marker id 219 not found.
	at org.eclipse.core.internal.resources.Marker.checkInfo(Marker.java:57)
	at org.eclipse.core.internal.resources.Marker.setAttribute(Marker.java:237)
	at org.eclipse.xtext.ui.editor.validation.MarkerCreator.setMarkerAttributes(MarkerCreator.java:45)
	at org.eclipse.xtext.ui.editor.validation.MarkerCreator.createMarker(MarkerCreator.java:31)
	at org.eclipse.xtext.ui.editor.validation.AddMarkersOperation.execute(AddMarkersOperation.java:106)
	at org.eclipse.ui.actions.WorkspaceModifyOperation$1.run(WorkspaceModifyOperation.java:106)
	at org.eclipse.core.internal.resources.Workspace.run(Workspace.java:2313)
	at org.eclipse.ui.actions.WorkspaceModifyOperation.run(WorkspaceModifyOperation.java:118)
	at org.eclipse.xtext.ui.editor.validation.MarkerIssueProcessor.processIssues(MarkerIssueProcessor.java:58)
	at org.eclipse.xtext.ui.editor.validation.ValidationJob.run(ValidationJob.java:67)
	at org.eclipse.core.internal.jobs.Worker.run(Worker.java:54)
 Topic: Code Formatter and comments.
Code Formatter and comments. [message #1412849] Wed, 27 August 2014 04:17
Markus Duft is currently offline Markus Duft
Messages: 95
Registered: February 2013
Member
Hey. I have a slight problem with code formatting in my DSL. Assume this file contents:

/**
 * comment
 */
persistent XX {
	/**
	 * asd
	 */
	data-category XX;

			/**
			 * more comment
			 */
	attribute xx : common.BOOLEAN {
		property label "Something";
	}
}


running the code formatter with this config:

	override protected void configureFormatting(FormattingConfig c) {
		c.setAutoLinewrap(120);

		for (pair : ga.findKeywordPairs("{", "}")) {
			c.setNoLinewrap.before(pair.first);
			c.setLinewrap(1, 1, 1).after(pair.first);
			c.setLinewrap(1, 1, 2).before(pair.second);
			c.setLinewrap(1, 1, 2).after(pair.second);

			c.setIndentationIncrement.after(pair.first);
			c.setIndentationDecrement.before(pair.second);
		}

		for (kw : ga.findKeywords(";")) {
			c.setLinewrap(1, 1, 2).after(kw);
			c.setNoSpace.before(kw);
		}
		
		c.setNoLinewrap.before(ga.connectionAccess.toKeyword_4);

		// It's usually a good idea to activate the following three statements.
		// They will add and preserve newlines around comments
		c.setLinewrap(0, 1, 2).before(SL_COMMENTRule)
		c.setLinewrap(0, 1, 2).before(ML_COMMENTRule)
		c.setLinewrap(0, 1, 1).after(ML_COMMENTRule)
	}


yields this weird formatted comments. what am i doing wrong?? xtext version is 2.6.0.

/**
 * comment
 */
persistent XX {
/**
	 * asd
	 */
	data-category XX;

	/**
			 * more comment
			 */
	attribute xx : common.BOOLEAN {
		property label "Something";
	}
}


Any hints? Will I have to write my own formatter without help from the declarative one?

Thanks!
Re: Code Formatter and comments. [message #1412870 is a reply to message #1412849] Wed, 27 August 2014 05:03
Thomas Fritsch is currently offline Thomas Fritsch
Messages: 14
Registered: April 2013
Location: Germany
Junior Member
Looks very much like https://bugs.eclipse.org/bugs/show_bug.cgi?id=313133
Re: Code Formatter and comments. [message #1412905 is a reply to message #1412870] Wed, 27 August 2014 06:34
Markus Duft is currently offline Markus Duft
Messages: 95
Registered: February 2013
Member
Hey, thanks for the pointer, didn't find that one o.O

Will try the workaround, as I really need this working (I'm implementing format on save, so this must work then Wink)
Re: Code Formatter and comments. [message #1412956 is a reply to message #1412905] Wed, 27 August 2014 08:26
Markus Duft is currently offline Markus Duft
Messages: 95
Registered: February 2013
Member
Sadly this does not help at all... Neutral any hints on what I could try?
 Topic: right place to add save actions?
Re: right place to add save actions? [message #1412967 is a reply to message #940606] Wed, 27 August 2014 08:57
Markus Duft is currently offline Markus Duft
Messages: 95
Registered: February 2013
Member
The above made my day, thanks Smile
 Topic: Headless Xcore generation with edit plug-in
Headless Xcore generation with edit plug-in [message #1413021] Wed, 27 August 2014 11:24
Axel Guckelsberger is currently offline Axel Guckelsberger
Messages: 215
Registered: July 2009
Senior Member
Hi all,


I am trying to get Xcore generation working headlessly using a mwe2
workflow.

The xcore file has the following package annotation:

@GenModel(
    modelDirectory="org.example.mydsl/emf-gen",
    editPluginID="org.example.mydsl.edit",
    editDirectory="org.example.mydsl.edit/src-gen",
    forceOverwrite="true",
    updateClasspath="false",
    complianceLevel="5.0",
    operationReflection="false",
    runtimePlatform="RCP")


Inside the IDE everything works as expected with this.

The mwe2 workflow looks as follows:

var projectName = "org.example.mydsl"
var runtimeProject = "../${projectName}"

Workflow {

    bean = StandaloneSetup {
        scanClassPath = true
        platformUri = "${runtimeProject}/.."

        registerGenModelFile = 
"platform:/resource/org.eclipse.emf.ecore/model/Ecore.genmodel"
    }

    component = org.eclipse.emf.mwe.utils.DirectoryCleaner {
        directory = 'emf-gen'
    }

    component = de.itemis.xcore2java.XcoreReader {
        register = org.eclipse.emf.ecore.xcore.XcoreStandaloneSetup {}
        path = 'model'
        slot = 'model'
    }

    component = org.eclipse.xtext.generator.GeneratorComponent {
        register = org.eclipse.emf.ecore.xcore.XcoreStandaloneSetup {}
        slot = 'model'
        outlet = {
            path = 'emf-gen'
        }
    }
}


While the actual meta classes are generated properly in Jenkins the edit
plug-in's classes seem to be ignored (or generated to somewhere unexpected).

Any idea what needs to be done here? Is a second outlet required maybe?

TIA,

Axel
 Topic: Non-empty model in the beginning
Re: Non-empty model in the beginning [message #1412916 is a reply to message #1412624] Wed, 27 August 2014 06:59
Axel Guckelsberger is currently offline Axel Guckelsberger
Messages: 215
Registered: July 2009
Senior Member
Hi,

maybe you can work with code templates, like described here:
https://www.eclipse.org/Xtext/documentation.html#templates

After you defined a template for your desired initial model content, you
only need to find a way how to insert this template programmatically.

HTH,
Axel


Inti Gonzalez-Herrera wrote:

> Hi,
>
> I'm creating a language with a few related concepts as Type, Property and
> Entity. The idea is that entities contain properties and each property is
> from a type.
>
> The problem at hand is that there are some predefined properties for each
> entity, and of course these properties have predefined types. It is by far
> more easy to handle all the properties in the same way in the validator
> and the generator but, to do so, I need to include such properties by
> default in the model.
>
> Is there a way to customize the mechanism to create an "empty" model? I
> mean, I want to execute some Java/Xtend code when a model/file is created.
>
> If possible, I would like to avoid solutions as: create an ecore model
> "graphically", then import it because the language is under development it
> is not comfortable to modify a graphic model over and over.
>
> Thanks
Re: Non-empty model in the beginning [message #1413029 is a reply to message #1412916] Wed, 27 August 2014 11:49
Uli Merkel is currently offline Uli Merkel
Messages: 121
Registered: June 2013
Senior Member
Ok: the very best would be if the preamble text is added in the open process of the empty file.

But: what about putting all your initial specifications in the Proposal Provider.
So if you have an empty file and press CTRL+SPACE, you get this text.

Another option is the validator and a quickfix which enters your standard text.

[Updated on: Wed, 27 August 2014 11:50]

Report message to a moderator

Pages (4): [ «    1  2  3  4    »]


Current Time: Wed Aug 27 15:24:49 EDT 2014

Powered by FUDForum. Page generated in 0.04652 seconds