Some Eclipse Foundation services are deprecated, or will be soon. Please ensure you've read this important communication.

Bug 354341

Summary:	Rely on XPand release rather than GMFT XPand fork
Product:	[Modeling] GMF-Tooling	Reporter:	Mickael Istria <mistria>
Component:	Core	Assignee:	Project Inbox <gmp.gmf-tooling-inbox>
Status:	RESOLVED WONTFIX	QA Contact:
Severity:	enhancement
Priority:	P3	CC:	borlander
Version:	unspecified
Target Milestone:	3.1
Hardware:	All
OS:	All
Whiteboard:

Description Mickael Istria

2011-08-10 04:48:52 EDT

GMF-Tooling relies on a forked XPand for code generation. This fork should be removed, and instead it the "official" XPand should be used.

Comment 1 Michael Golubev

2011-08-10 05:15:50 EDT

I am not sure exactly why the fork had been done, but I know it taken more than 2 developer months to do that. To my understanding them main reason behind was to use QVTO expressions.

I am not a big fun of that decision from the past, but changing it back now would requires 
a) rewriting (may be partially semi-automatically, but definitely not not 100% automatically) all the templates back to xtend/xpand
b) forcing users to rewrite every custom template
c) providing some verification tool that confirms that the custom templates works as expected after migration

At the time GMF started to use the fork I was responsible for UML2Tools templates, and was really unhappy being forced to spent weeks for items b) and c), and I got enough experience to strongly recommend to leave it as is. 

At least we need some reasons better than "lets just use the latest release".

Regards, 
Michael

Comment 2 Mickael Istria

2011-08-10 05:39:25 EDT

This idea comes to me after a discussion with Itemis guys on the Spray (equivalent to GMF-Tooling for Graphiti) mailing-list.
Those guys seem to hate GMF-Tooling just because there is a forked XPand.

But I was not aware of the reason why GMF forked XPand. I thought support for QVTo was built-in XPand, I didn't know that it is the core of the GMF fork of XPand.
Maybe XPand will one day include support for QVTo, then we won't need this fork anymore.

There is probably another way to use XPand and QVTo together, that we could use one day instead of the fork: http://wiki.eclipse.org/Model2Text_using_Xpand_and_QVT_for_Query

Comment 3 Michael Golubev

2012-09-11 21:45:06 EDT

Reconsidering for Kepler

Comment 4 Michael Golubev

2012-09-26 11:05:01 EDT

For plan item for Kepler release use https://bugs.eclipse.org/bugs/show_bug.cgi?id=386838

Comment 5 Michael Golubev

2014-03-19 05:16:05 EDT

Since Kepler SR2, GMF-Tooling provides set of xtend2 templates that allow to generate identical code.