Eclipse Community Forums: TMF (Xtext) » ID and a more relaxed version living together

Help

Home

Home » Modeling » TMF (Xtext) » ID and a more relaxed version living together

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

ID and a more relaxed version living together [message #1755879]

Wed, 08 March 2017 20:52

Rodrigo Ruiz

Messages: 28
Registered: March 2017

Junior Member

Hi i need to support a structure like this:

varname=nodeOne/node-2/node-3

at first i implemented it changing the definition for ID, just including the '-' in it, but since i also support math expressions, if you do this:

sum= node1-node2 (without space in the middle)
the parser will think node1-node2 is a single ID.

If i try to include a more relaxed version of ID, like SEGMENT to live along with ID then the parser can't distinguish between ID and SEGMENT, forcing me to put an ^ before my id, which is bad for my users.
Do you have any tip for situations like this?

this is my definition for SEGMENT and ID (the standard one):

terminal SEGMENT: ('a'..'z'|'A'..'Z'|'_') (  'a'..'z'|'A'..'Z'|'_'|'-'|'0'..'9')*;
@Override
terminal ID  		: '^'?('a'..'z'|'A'..'Z'|'_') ('a'..'z'|'A'..'Z'|'_'|'0'..'9')*;

Thanks for your help!

Report message to a moderator

Re: ID and a more relaxed version living together [message #1755880 is a reply to message #1755879]

Wed, 08 March 2017 21:17

Christian Dietrich

Messages: 14665
Registered: July 2009

Senior Member

how do you know if x-y is a SEGEMENT or is ID-ID

if you can know it from place of occurence you could introduce a datatype rule

SEGMENT: ID ("-" (ID|INT))+;

Twitter : @chrdietrich
Blog : https://www.dietrich-it.de

Report message to a moderator