Eclipse Community Forums: TMF (Xtext)

Help

Home

Home » Modeling » TMF (Xtext) » Xtext good pracitces(I am looking for good Xtext usage practices)

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

Xtext good pracitces [message #1758236]

Sat, 25 March 2017 12:29

graph researcher

Messages: 8
Registered: March 2017

Junior Member

First, as a new user, I would like to know how different responsibilities szud be distributed between different elements of the DSL design/implementation in Xtext. For example, where the simplest constraints I can imagine should be implemented?
I have problmes while implementeing Xtext grammar for such atrivial example (the programm in the DSL I am trying to define):

<node isRooted="true">
<node isRooted="false">

Where the "true" and "false" string contraints representing booleans in my XML DSL language should be defined:
- in the Xtext grammar
- elsewhere (validation rules only or content assistant only, or both or in other place(s))

Is a good practice to define as week grammar (very limited contraints of this type) as possible and to move the implementation responsibility of them to other DSL design elements?

How the grammar should look like for such the construct?

Report message to a moderator

Re: Xtext good pracitces [message #1758325 is a reply to message #1758236]

Mon, 27 March 2017 16:50

Christian Dietrich

Messages: 14738
Registered: July 2009

Senior Member

Hi, your question is very vague so i cant give an answer.

generally you can encode stuff

- in the grammar
- in grammar related component (aka value converter)
- in MyDslValidator (semantic validation)

in your case you would add a boolean attribute to the grammar. that could look like

"isRooted" "=" (isRooted?="true"|"false")

The pratice i recommend is:

- if it can easily be done in the grammar, do it in the grammar,
- if not => use validation etc

Twitter : @chrdietrich
Blog : https://www.dietrich-it.de
Day Job: https://www.everest-systems.com

[Updated on: Mon, 27 March 2017 16:53]

Report message to a moderator

Re: Xtext good pracitces [message #1758562 is a reply to message #1758325]

Thu, 30 March 2017 07:09

Jan Koehnlein

Messages: 760
Registered: July 2009
Location: Hamburg

Senior Member

Another indicator whether to use validation or a strict grammar is the error message the user gets.
If you're OK with something like "Unexpected token 'foo'. Expected 'true' or 'false'", doing it in the grammar is fine.
If you rather want something like "The value of 'isRooted' can only be 'true' or 'false'" got for a lose grammar and strict validation.

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

Report message to a moderator

Re: Xtext good pracitces [message #1759767 is a reply to message #1758325]

Tue, 18 April 2017 10:36

graph researcher

Messages: 8
Registered: March 2017

Junior Member

Thank you for your answers. I isolated the problem just before your answered my question and realized that its reason was related to the way I were using strings in my Xtext grammar.
So, for the following DSL expression:
rooted = "true"
the appropriate grammar rule should look like that:
'rooted' '=' (isRooted ?= '"true"' | '"false"')
// single-quote double-quote true double-quote single-quote
In place of that I was trying to concatenate double quotes surrounded by single quotes with the true and false grammar terminal symbols (surronded by single quotes). I do not know the nature of the problems involved by my former approach. In the consequence I do not know how can I introduce terminal symbols into the string (surrounded by double quotes). But this is not the problem I have, just an observation.
Sorry, for being not precise enough before.

Report message to a moderator

Previous Topic:	Xtext and antlr4?
Next Topic:	is it possible to call @check method only once in validation?

Goto Forum:

-=] Back to Top [=-

[ Syndicate this forum (XML) ] [

]

Current Time: Sat Dec 21 14:08:35 GMT 2024

.:: Contact :: Home ::.

Breadcrumbs

Sign up to our Newsletter