Eclipse Community Forums: TMF (Xtext) » how to define a Dummy Block in Grammar?

Help

Home

Home » Modeling » TMF (Xtext) » how to define a Dummy Block in Grammar?

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

how to define a Dummy Block in Grammar? [message #556674]

Thu, 02 September 2010 10:03

No real name

Messages: 16
Registered: September 2010

Junior Member

Hello, I want to build a grammar iteratively for a subset of an existing language.

The key problem is to be able to define a dummy part - I don't care what's in right now.

for example:

Model:
loop=ForLoop;

ForLoop:
'for' type=Type varname=ID ';' cond=Condition; incr=Increment
dummy=Dummy // dummy can contain another for-loop
'end for'

Is it possible?

Best regars, Anatoli

Report message to a moderator

Re: how to define a Dummy Block in Grammar? [message #556679 is a reply to message #556674]

Thu, 02 September 2010 10:28

Bodo

Messages: 27
Registered: August 2010

Junior Member

Just introduce a new grammar rule like:

Dummy:
forLoop=ForLoop
|
;

Report message to a moderator

Re: how to define a Dummy Block in Grammar? [message #556684 is a reply to message #556674]

Thu, 02 September 2010 10:40

Alexander Nittka

Messages: 1193
Registered: July 2009

Senior Member

Hi,

to be a bit more verbose (yet untested). Note the changes in the ForLoop-Definition. It is to be preferred not to have white spaces within keywords.

ForLoop:
'for' type=Type varname=ID ';' cond=Condition; incr=Increment
dummy+=Dummy*
'end' 'for';

Dummy: ForLoop| content=Anything;

//pipe separated List of any terminal rule or keyword that may appear there
Anything: (ID|INT|STRING|';'|...);

Alex

Report message to a moderator

Re: how to define a Dummy Block in Grammar? [message #556707 is a reply to message #556679]

Thu, 02 September 2010 12:13

No real name

Messages: 16
Registered: September 2010

Junior Member

Thank you very much for your answeres: I'm one step further now, but there are still some problems:

@Bodo: is it a special construct or just a pattern how you would do that?

Dummy:
forLoop=ForLoop
|
; is not a valid syntax

@Alexander

I could skip lots of stuff going this way before I came across the following, let's stick to my example:

Model:
loop=ForLoop;

ForLoop:
for1=FOR type=Type varname=ID ';' cond=Condition; incr=Increment
dummy=Dummy // dummy can contain another for-loop
end=END for2=FOR

END:
end='end';

FOR:
for='for';

Dummy: ForLoop| content=Anything;

Anything: (ID|INT|STRING|';');

now I can have something like:

for int i; i < 10; i++

if a > 3

end if

end for

where I do not have the "if"-construct yet - it should be covered by Dummy, but it is not valid to have 'end' between 'for' and 'end' 'for'

so I extended Anything to:

Anything: {Anything}(ID|INT|STRING|';'|end=END);

but then I get a warning:

the decision cannot distinguish between alternative(s) 1,2 for at least one input sequence

full warning:

0 [main] INFO lipse.emf.mwe.utils.StandaloneSetup - Registering platform uri 'C:\Users\bar\workspace'
627 [main] INFO ipse.emf.mwe.utils.DirectoryCleaner - Cleaning C:\Users\bar\workspace\org.xtext.example.mydsl\..\org.xtext. example.mydsl\src-gen
639 [main] INFO ipse.emf.mwe.utils.DirectoryCleaner - Cleaning C:\Users\bar\workspace\org.xtext.example.mydsl\..\org.xtext. example.mydsl.ui\src-gen
791 [main] INFO ipse.xtext.generator.LanguageConfig - generating infrastructure for org.xtext.example.mydsl.MyDsl with fragments : ImplicitRuntimeFragment, ImplicitUiFragment, GrammarAccessFragment, EcoreGeneratorFragment, ParseTreeConstructorFragment, ResourceFactoryFragment, XtextAntlrGeneratorFragment, JavaValidatorFragment, ImportNamespacesScopingFragment, QualifiedNamesFragment, BuilderIntegrationFragment, FormatterFragment, LabelProviderFragment, TransformerFragment, OutlineNodeAdapterFactoryFragment, QuickOutlineFragment, QuickfixProviderFragment, JavaBasedContentAssistFragment, XtextAntlrUiGeneratorFragment
warning(202): ../org.xtext.example.mydsl/src-gen/org/xtext/example/mydsl/p arser/antlr/internal/InternalMyDsl.g:707:1: the decision cannot distinguish between alternative(s) 1,2 for at least one input sequence
7650 [main] INFO or.validation.JavaValidatorFragment - executing generate for org.eclipse.xtext.generator.validation.JavaValidatorFragment
warning(202): ../org.xtext.example.mydsl.ui/src-gen/org/xtext/example/myds l/ui/contentassist/antlr/internal/InternalMyDsl.g:1098:32: the decision cannot distinguish between alternative(s) 1,2 for at least one input sequence
8910 [main] INFO .emf.mwe2.runtime.workflow.Workflow - Done.

Report message to a moderator

Re: how to define a Dummy Block in Grammar? [message #556714 is a reply to message #556707]

Thu, 02 September 2010 12:29

No real name

Messages: 16
Registered: September 2010

Junior Member

fyi:

the for loop was a simplified example:

this is my actual (still simple) code, which throws the warning,

the warning is not there, when I remove end=END from rule Anything

grammar org.xtext.example.mydsl.MyDsl with org.eclipse.xtext.common.Terminals

generate myDsl "http://www.xtext.org/example/mydsl/MyDsl"

Model:
header=PackageHeader body=PackageBody;

PackageHeader:
literal=CreateOrReplacePackage packagename=ID is=IS content=PackageHeaderContent end=END

packegename=ID ';' ('/')?;

PackageHeaderContent:
dummy=Dummy;

PackageBody:
literal1=CreateOrReplacePackage literal2=Body packagename1=ID is=IS content=Dummy end=END

packagename2=ID ';' ('/')?;

CreateOrReplacePackage:
sentence='CREATE OR REPLACE PACKAGE';

Body:
sentence='BODY';

Dummy:
{Dummy}(dummy+=Anything)*;

IS:
is='IS';

END:
end='END';

Anything: {Anything}(ID|INT|STRING|';'|'('|')'|'*'|'/'|'-'|':'|'='|',' |is=IS|end=END);

Report message to a moderator

Re: how to define a Dummy Block in Grammar? [message #556718 is a reply to message #556714]

Thu, 02 September 2010 12:49

Meinte Boersma

Messages: 434
Registered: July 2009
Location: Leiden, Netherlands

Senior Member

The ANTLR warning (actually, it's more of an error) is there because the alternatives in your Anything rule overlap. I.e., there are two alternatives which consume the same token. It's quite easy to spot which one: 'IS' and 'END' both are consumed by the ID terminal rule as well.

In fact, I'm completely mystified why the IS and END rules exist at all, as they simply correspond to (the presence of) keywords. So, you can "inline" them and get rid of the ANTLR warning in one go.

Xtext blogs: executable models...again? | workshop material | custom scoping with Xtend

Report message to a moderator

Re: how to define a Dummy Block in Grammar? [message #556743 is a reply to message #556718]

Thu, 02 September 2010 13:57

No real name

Messages: 16
Registered: September 2010

Junior Member

Hello Meinte,

thanks for the explanation: I understand what you are saying.

I removed is=IS and end=END from Anything rule, but then I get a problem with the following at the subsequent IS belonging to FUNCTION (which I don't want to declare, because I don't want to have to declare everything at once, but iteratively) - I don't want to care what's inside the package body.

CREATE OR REPLACE PACKAGE sysstat IS
...
END sysstat;
/

CREATE OR REPLACE PACKAGE BODY sysstat IS

...

FUNCTION hour(hourx IN NUMBER DEFAULT 1) RETURN VARCHAR2 IS
BEGIN
RETURN '0';
END hour;

END sysstat;

Report message to a moderator

Previous Topic:	Java extensions for Xpand, editor support
Next Topic:	Validation-Error "An object may not circularly contain itself" validating an gmf-diagram

Goto Forum:

-=] Back to Top [=-

[ Syndicate this forum (XML) ] [

]

Current Time: Thu Apr 25 15:09:50 GMT 2024

.:: Contact :: Home ::.

Breadcrumbs

Sign up to our Newsletter