Eclipse Community Forums: TMF (Xtext) » Parsing and generating annotations

Help

Home

Home » Modeling » TMF (Xtext) » Parsing and generating annotations

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

Parsing and generating annotations [message #871102]

Thu, 10 May 2012 20:47

Jens-Uwe Moeller

Messages: 12
Registered: March 2012
Location: Hamburg

Junior Member

Hi,

I want to be able to write something similar like in Java:

/** documentation
/*
@annotation1
@annotation2
mylanguagetypeA name1 { xxx }

/** documentation
/*
@annotation2
@annotation1
mylanguagetypeB name2 { xxx }

Using "documentation = e.documentation" in the generation process allows me to access documentation without having defined this feature in the grammar at all - you i.e. xtext provide this feature to me.

When using annotations I get in conflict with the rule "one look ahead", when trying to define a grammar which is in its top rule similar like

Mylanguagetypes: MylanguagetypeA | MylanguagetypeB |MylanguagetypeC;

MylanguagetypeA: ... something for annotation ...
                 'mylanguagetypeA' ...

MylanguagetypeB: ... something for annotation ...
                 'mylanguagetypeB' ...

MylanguagetypeA resp. ..B resp. ..C are subject for code generation with the JVM inferer.

I can solve my problem easily by changing the language and integrating annotations behind the keywords 'mylanguagetypeA' etc. That way I can generate even complex annotations. But is there a better solution without changing the language?

Thanks in advance,
Jens

Report message to a moderator

Re: Parsing and generating annotations [message #871164 is a reply to message #871102]

Fri, 11 May 2012 08:21

Jan Koehnlein

Messages: 760
Registered: July 2009
Location: Hamburg

Senior Member

That's a typical usecase for assigned actions.
Have a look at the Xtend.xtext grammar to get an impression.

Am 10.05.12 22:47, schrieb Jens-Uwe Moeller:
> Hi,
>
> I want to be able to write something similar like in Java:
>
>
> /** documentation
> /*
> @annotation1
> @annotation2
> mylanguagetypeA name1 { xxx }
>
> /** documentation
> /*
> @annotation2
> @annotation1
> mylanguagetypeB name2 { xxx }
>
>
> Using "documentation = e.documentation" in the generation process allows
> me to access documentation without having defined this feature in the
> grammar at all - you i.e. xtext provide this feature to me.
>
> When using annotations I get in conflict with the rule "one look ahead",
> when trying to define a grammar which is in its top rule similar like
>
> Mylanguagetypes: MylanguagetypeA | MylanguagetypeB |MylanguagetypeC;
>
> MylanguagetypeA: ... something for annotation ...
> 'mylanguagetypeA' ...
>
> MylanguagetypeB: ... something for annotation ...
> 'mylanguagetypeB' ...
>
>
> MylanguagetypeA resp. ..B resp. ..C are subject for code generation with
> the JVM inferer.
>
> I can solve my problem easily by changing the language and integrating
> annotations behind the keywords 'mylanguagetypeA' etc. That way I can
> generate even complex annotations. But is there a better solution
> without changing the language?
>
> Thanks in advance,
> Jens
>
>
>

--
Need professional support for Eclipse Modeling?
Go visit: http://xtext.itemis.com

---
Get professional support from the Xtext committers at www.typefox.io

Report message to a moderator

Re: Parsing and generating annotations [message #871190 is a reply to message #871164]

Fri, 11 May 2012 11:25

Sebastian Zarnekow

Messages: 3108
Registered: July 2009

Senior Member

It's probably simpler to read the reference documentation about actions.

Regards,
Sebastian

Am 11.05.12 10:21, schrieb Jan Koehnlein:
> That's a typical usecase for assigned actions.
> Have a look at the Xtend.xtext grammar to get an impression.
>
> Am 10.05.12 22:47, schrieb Jens-Uwe Moeller:
>> Hi,
>>
>> I want to be able to write something similar like in Java:
>>
>>
>> /** documentation
>> /*
>> @annotation1
>> @annotation2
>> mylanguagetypeA name1 { xxx }
>>
>> /** documentation
>> /*
>> @annotation2
>> @annotation1
>> mylanguagetypeB name2 { xxx }
>>
>>
>> Using "documentation = e.documentation" in the generation process allows
>> me to access documentation without having defined this feature in the
>> grammar at all - you i.e. xtext provide this feature to me.
>>
>> When using annotations I get in conflict with the rule "one look ahead",
>> when trying to define a grammar which is in its top rule similar like
>>
>> Mylanguagetypes: MylanguagetypeA | MylanguagetypeB |MylanguagetypeC;
>>
>> MylanguagetypeA: ... something for annotation ...
>> 'mylanguagetypeA' ...
>>
>> MylanguagetypeB: ... something for annotation ...
>> 'mylanguagetypeB' ...
>>
>>
>> MylanguagetypeA resp. ..B resp. ..C are subject for code generation with
>> the JVM inferer.
>>
>> I can solve my problem easily by changing the language and integrating
>> annotations behind the keywords 'mylanguagetypeA' etc. That way I can
>> generate even complex annotations. But is there a better solution
>> without changing the language?
>>
>> Thanks in advance,
>> Jens
>>
>>
>>
>
>

--
Need professional support for Eclipse Modeling?
Go visit: http://xtext.itemis.com

Report message to a moderator

Previous Topic:	2.3.0 M6 JVM generation and Enums
Next Topic:	dangling reference error in XBlockExpression when accessing properties

Goto Forum:

-=] Back to Top [=-

[ Syndicate this forum (XML) ] [

]

Current Time: Tue Jul 07 12:59:01 GMT 2020

.:: Contact :: Home ::.

Breadcrumbs

Sign up to our Newsletter