[
Date Prev][
Date Next][
Thread Prev][
Thread Next][
Date Index][
Thread Index]
[
List Home]
Re: [wtp-dev] JSDT JS parser, internal representation, type inference and all that jazz
|
inlined
On 7 Apr 2016, at 3:00, Eugene Melekhov wrote:
Hi all,
I've been working on accommodation of the JSDT core to ES6, new
esprima parser etc. Right now I'm working on restoring
so called "bindings" used for code assistance etc.
Thank you for working on it, it is important for our Neon delivery.
I will finish the job that I've started and will implement initial
bindings resolver for single source but... I must admit
that in my opinion what we're doing at the moment is wrong.
I do not see it as a long term solution either, I think we have reached
the limits of nashorn based solution. We either
need a way to keep all this on JS world and integrate on the UI level,
which is what Angelo does on tern.java/ts.java.
Or we need to go with a full on Java solution.
Since you have brought it up I have also been looking at Closure
compiler and I agree it is a strong candidate for a
full-on-java solution. So far it looks OK, but I have not done any
testing with large JS projects, which was the real
pain with the old JSDT Compiler.
Another option that I am investigating is Node4J[1]. This is the sister
implementation of J2V8, it provides JNI
based bindings into node.js from Java. It is too early to tell if it is
a good option, TBH I am not even sure where
the project is at the moment.
[1] http://www.infoq.com/presentations/node4j-nodejs-java
We should also consider our due diligence as an eclipse hosted project.
During the Neon cycle, we have lost a lot of
time, between organizing our teams and waiting for eclipse IP process
and we effectively got started around Neon M5.
In order to avoid that my goal is to start the IP process for closure
compiler so that it is available as an option if
we choose to do so. Same for Node4J as soon as I can figure out what to
submit. :)
* We're trying to reuse old dom model inherited from JDK which is not
fully compliant with JS reality even after
addition of ES6 specific constructs.
* This model is not very convenient for internal and especially
external use.
Correct, we have already retired the internal AST model, I am hoping to
completely remove it after Neon. Unfortunately
abandoning DOM AST is not as easy as internal AST even with a “can
break APIs policy” because jsdt.ui uses DOM AST model.
Unless we have the enough contributors to change all that functionality,
we need to keep the DOM AST working no matter
how the javascript is parsed.
We can probably start pruning and shaping DOM AST to a more efficient
JavaScript model, which is something I have not
started doing yet.
* We have to convert this model to fro JSON/other AST formats.
* Integration with JS modules like esprima or Tern or anything alike
is resource consuming, complex, opaque, error prone
etc.
Actually several tools out there have already shown this approach can
work. I am concerned about the efficiency of the
mechanics of integration but the general principle can work.
So, what is the right way in my opinion? The answer is to completely
abandon or internal representation (ASTDom) model,
esprima parser and any other JS based tools like Tern and switch to
using Google closure compiler infrastructure instead.
This approach has the following advantages:
* Closure compiler is fast
* It's written in Java and there are no problem with integration.
* It's a mature project with vast and VERY active community
* It has it's own perfect internal representation (IR) and it's
possible to use just AST(rhino trees) if necessary
https://github.com/google/closure-compiler/tree/master/src/com/google/javascript/rhino
* IR has great traversal infrastructure
https://github.com/google/closure-compiler/blob/master/src/com/google/javascript/jscomp/NodeTraversal.java
* IR and a pluggable compiler allows to perform all tasks that we
need. IR can be used for for outline
view, symbol tables for code assist, search, navigation, auto
completion, refactoring and so on.
* It has call graphs, type flow graphs, type inference out of the box.
https://github.com/google/closure-compiler/blob/master/src/com/google/javascript/jscomp/TypeInference.java
* It has code validation, lint etc
https://github.com/google/closure-compiler/tree/master/src/com/google/javascript/jscomp/lint
* It has code printer
https://github.com/google/closure-compiler/blob/master/test/com/google/javascript/jscomp/CodePrinterTest.java
* It's possible to use it's IR to perform any other necessary tasks
like more elaborate type analysis that TAJS does.
In conclusion, I believe JSDT should switch to Google closure compiler
back-end as soon as it possible before it's too
late. It would allow to forget about almost all low-level problems and
concentrate on implementing functionality that is
visible and is really necessary for JSDT users:
I agree, I think this is the main principal, unfortunately the progress
had been slower than I was expecting but we should get
better now that the committers/contributors are settling down and
getting more efficient.
* Fast modern editor
* Elaborate code assistance
* Smart navigation
* Smart refactoring
* Integration of v8, node, npm, gulp, babel etc
* Debugging
* JSX support
* and so on, you name it
Thank you.
--
Eugene Melekhov
_______________________________________________
wtp-dev mailing list
wtp-dev@xxxxxxxxxxx
To change your delivery options, retrieve your password, or
unsubscribe from this list, visit
https://dev.eclipse.org/mailman/listinfo/wtp-dev