Eclipse Community Forums
Forum Search:

Search      Help    Register    Login    Home
Home » Modeling » GMT (Generative Modeling Technologies) » Doubt about a textual editor for QVT
Doubt about a textual editor for QVT [message #376315] Thu, 12 April 2007 10:56 Go to next message
Adolfo is currently offline AdolfoFriend
Messages: 2
Registered: July 2009
Junior Member
Hello,

Sorry if im not posting in the correct news, but i havent could find a
better place for it.

I m known that Ed Willink have been working in a textual editor for QVT
Relations, and im building an Eclipse's IDE for Operational Mapping, so i
think that Ed could help me in this job.

Good, as well as Relations, a bit of Operational Mapping's grammar is based
on OCL Expressions, and we have an eclipse proyect that resolve us a lot of
work in this way (MDT-OCL). So i have decided to generate my Operational
Mapping parser using LPG, extending OCL's grammar.

As you know, we have a problem when we are parsing wrong inputs, because the
parser generated does'nt build an AST, and it only locates all the errors in
such input. Locating all the errors in the input is a great and necessary
feature, but if i want a good quality IDE, definetely i need an AST.

I dont know if you are trying to solve these issues. Im working on them,
building my own recovery class and methods, but perhaps you could give me
some advice. Im not sure that implementing the recovery policies to try the
ast's building is a good idea.

what do you think about this ?

Thank you very much for the help.
Re: Doubt about a textual editor for QVT [message #378844 is a reply to message #376315] Fri, 03 August 2007 08:00 Go to previous message
Ed Willink is currently offline Ed WillinkFriend
Messages: 4154
Registered: July 2009
Senior Member
Hi Adolfo

Sorry for the delayed reply. I don't look at the newsgroups as often as I ought.

The latest GMT/UMLX release aligned with Eclipse 3.3 will be posted on Sunday.

It solves EMOF interoperability issues and provides multi-page model
sensitive editing for
- concrete syntax text
- abstract syntax Ecore Resource tree
- abstract syntax Ecore XMI
- abstract syntax EMOF XMI.

There are editors and AST builders for KM3, OCL, QVTc and QVTr.

I looked at QVTo but the recent grammar had a number of OCL contradictions.
I understand that the latest QVTo grammar is OCL compliant but not expressed
in a way that makes this self evident. If a BNF of the QVTo grammar can be
produced that is clearly an OCL extension, it should be fairly straightforward
to use the same OCL extending technology that I use for QVTc and QVTr to
support QVTo editing and CST to AST round-tripping.

If you want to use MDT OCL for QVTo I would strongly recommend looking
at the way I have made it extensible in the org.eclipse.gmt.umlx.ocl.cst
plug-in and extend it in org.eclipse.gmt.umlx.eqvtr.cst.

Regards

Ed Willink


Adolfo wrote:
> Hello,
>
> Sorry if im not posting in the correct news, but i havent could find a
> better place for it.
>
> I m known that Ed Willink have been working in a textual editor for QVT
> Relations, and im building an Eclipse's IDE for Operational Mapping, so i
> think that Ed could help me in this job.
>
> Good, as well as Relations, a bit of Operational Mapping's grammar is based
> on OCL Expressions, and we have an eclipse proyect that resolve us a lot of
> work in this way (MDT-OCL). So i have decided to generate my Operational
> Mapping parser using LPG, extending OCL's grammar.
>
> As you know, we have a problem when we are parsing wrong inputs, because the
> parser generated does'nt build an AST, and it only locates all the errors in
> such input. Locating all the errors in the input is a great and necessary
> feature, but if i want a good quality IDE, definetely i need an AST.
>
> I dont know if you are trying to solve these issues. Im working on them,
> building my own recovery class and methods, but perhaps you could give me
> some advice. Im not sure that implementing the recovery policies to try the
> ast's building is a good idea.
>
> what do you think about this ?
>
> Thank you very much for the help.
>
>
>
Re: Doubt about a textual editor for QVT [message #602834 is a reply to message #376315] Fri, 03 August 2007 08:00 Go to previous message
Ed Willink is currently offline Ed WillinkFriend
Messages: 4154
Registered: July 2009
Senior Member
Hi Adolfo

Sorry for the delayed reply. I don't look at the newsgroups as often as I ought.

The latest GMT/UMLX release aligned with Eclipse 3.3 will be posted on Sunday.

It solves EMOF interoperability issues and provides multi-page model
sensitive editing for
- concrete syntax text
- abstract syntax Ecore Resource tree
- abstract syntax Ecore XMI
- abstract syntax EMOF XMI.

There are editors and AST builders for KM3, OCL, QVTc and QVTr.

I looked at QVTo but the recent grammar had a number of OCL contradictions.
I understand that the latest QVTo grammar is OCL compliant but not expressed
in a way that makes this self evident. If a BNF of the QVTo grammar can be
produced that is clearly an OCL extension, it should be fairly straightforward
to use the same OCL extending technology that I use for QVTc and QVTr to
support QVTo editing and CST to AST round-tripping.

If you want to use MDT OCL for QVTo I would strongly recommend looking
at the way I have made it extensible in the org.eclipse.gmt.umlx.ocl.cst
plug-in and extend it in org.eclipse.gmt.umlx.eqvtr.cst.

Regards

Ed Willink


Adolfo wrote:
> Hello,
>
> Sorry if im not posting in the correct news, but i havent could find a
> better place for it.
>
> I m known that Ed Willink have been working in a textual editor for QVT
> Relations, and im building an Eclipse's IDE for Operational Mapping, so i
> think that Ed could help me in this job.
>
> Good, as well as Relations, a bit of Operational Mapping's grammar is based
> on OCL Expressions, and we have an eclipse proyect that resolve us a lot of
> work in this way (MDT-OCL). So i have decided to generate my Operational
> Mapping parser using LPG, extending OCL's grammar.
>
> As you know, we have a problem when we are parsing wrong inputs, because the
> parser generated does'nt build an AST, and it only locates all the errors in
> such input. Locating all the errors in the input is a great and necessary
> feature, but if i want a good quality IDE, definetely i need an AST.
>
> I dont know if you are trying to solve these issues. Im working on them,
> building my own recovery class and methods, but perhaps you could give me
> some advice. Im not sure that implementing the recovery policies to try the
> ast's building is a good idea.
>
> what do you think about this ?
>
> Thank you very much for the help.
>
>
>
Previous Topic:[TCS] Optional keywords in TCS vs. boolean attributes in metamodel...
Next Topic:[Announce] A Proposal and Prototype for an Eclipse Model Registry
Goto Forum:
  


Current Time: Thu Nov 27 23:43:46 GMT 2014

Powered by FUDForum. Page generated in 0.26283 seconds
.:: Contact :: Home ::.

Powered by: FUDforum 3.0.2.
Copyright ©2001-2010 FUDforum Bulletin Board Software