[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index] [List Home]
Re: [ide-dev] Common Language Infrastructure

In my experience, the most useful way to integrate language tooling is to provide a common abstract model that is basic enough to be applicable for all major languages but also detailed enough to provide real value.

As an example, common concepts like module, function, method, class, field can capture a rough hierarchical structure for most languages.  If each language tooling would implement its own version of the model, this would provide a way to do navigation and resolve references across languages.  It would also provide a way to do semantically aware search (if combined with Sven's idea of a cross-language index).


On Wed, Aug 21, 2013 at 8:21 AM, Doug Schaefer <dschaefer@xxxxxxx> wrote:
That's a great one. For our work on Qt at BlackBerry, we're starting to ask that question since you're able to bind _javascript_ to C++ and visa versa. Being able to navigate between multiple languages would be a huge win and make multi-languages more than multiple languages, but actually integrate them.

I think the CDT index could be used for that. It already handles multiple languages (C, C++, Fortran) though not as integrated as this. Also not sure how well it would play with dynamic languages.

Doug.

From: Sven Efftinge <sven@xxxxxxxxxxx>
Reply-To: Discussions about the IDE <ide-dev@xxxxxxxxxxx>
Date: Wednesday, 21 August, 2013 11:16 AM
To: Discussions about the IDE <ide-dev@xxxxxxxxxxx>
Subject: Re: [ide-dev] Common Language Infrastructure

IMHO most important would be some kind of cross-language workspace index (i.e. a global symbol table), which contains and maintains 
information about what external symbols have been consumed and are exposed by a certain file.

Sven

On Aug 21, 2013, at 5:04 PM, Doug Schaefer <dschaefer@xxxxxxx> wrote:

Well, let's get right to it. Mickael Istria raised a bug on JDT to have to port over to DLTK since they have the same roots. (https://bugs.eclipse.org/bugs/show_bug.cgi?id=415563)

It's not a bad idea, but it does bring up an area that we've sorely been lacking, a common language infrastructure. There are bits and pieces here and there including Xtext (which is more than a bit or piece), as well as the LTK in the platform that we in the CDT use to co-ordinate our refactoring, and of course DLTK itself.

I'd like to open the floor though. What do you think we would need in a common language infrastructure? Would it really help or is there really nothing in common? If you work on a language project, what do you have that could be used to kickstart this (e.g. CDT's build system).

Doug.
_______________________________________________
ide-dev mailing list
ide-dev@xxxxxxxxxxx
https://dev.eclipse.org/mailman/listinfo/ide-dev


_______________________________________________
ide-dev mailing list
ide-dev@xxxxxxxxxxx
https://dev.eclipse.org/mailman/listinfo/ide-dev