Hi folks, 

Martin, thanks for CCing me, some comments: 

Definitely DLTK has strong intention to move to debug-commons further, and 
we're very anxious about Werner's efforts on DBGp implementation. 

DLTK plans regarding debug is open and following: we *shall* have DBGp-based 
debugger by DLTK 1.0 RC0, which is scheduled on May 11th, so (and 
unfortunately) we have no time to cooperate with debug-commons right now. DLTK 
1.0 will be released with simple DBGp implementation on top of debug.rb. This 
solution is far from the best, but we need give end-users at least minimal 
possibility to debug their code. Also DLTK in general could not be considered 
as complete solution for the end-users, so vendors on top of DLTK could 
implement own debugging techniques.

After DLTK 1.0 we're going to switch to debug-commons. If debug-commons will 
have some DBGp-related stuff later in summer (murphee?), I believe we'll switch 
to that. If not, I believe we'll contribute it :)

We look forward to working with you guys, and
Kind Regards,
Andrey

-----Original Message-----
From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] 
Sent: Wednesday, May 02, 2007 1:05 AM
To: rubyeclipse-development@lists.sourceforge.net
Cc: Andrey Platov
Subject: Re: [RDT-Dev] debugger: stepReturn fix, next+, step+ support

Werner Schuster (murphee) wrote:
> Martin Krauskopf wrote:
>> BTW it seems that there will be something like jruby-debug soon. I hope 
>> I will have enough time to participate a lot on this piece. Might be 
>> another debug-commons child or jruby-extras, does not matter to much.
>>   
> jruby-debug would be good - BTW: yesterday I found out about a problem
> in JRuby and  set_trace_func
> based debugging: if JRuby's JIT is turned on, this won't work, because
> the code that the JIT emits doesn't call the trace function.
> I guess, there's not much to do about this before 1.0, but I guess
> filing this as a bug would is necessary for post-1.0.

If we would be able to come with jruby-debug "soon" (month, two?) I 
thing that set_trace_func should be completely deprecated as a technique 
for debugger implementation. It is quite unusable for any large 
applications like, yup, rails.
But should be supported for other interpreters as a fall-back.

>> Also it seems that DLTK people would join us in the future with 
>> debug-commons with DBGp implementation. Future seems quite good for 
>> ruby-debugging thanks to your initial contribution Markus! ;)

> Have they committed to sharing their DBGp implementation?

Andrey Platov wrote me that they would. But in the further future. They 
do not have enough time currently. I think they are quite open-minded 
folks. CCing Andrey then to talk for him.

> They have a Ruby version in their CVS but it's not based on the
> RDT/eclipse-debug code.
> 
> If there's an interest: I've been working on my own Ruby DBGp
> implementation (with the RDT code as backend) - it's not complete, and
> I'm still trying to get it to connect with DLTK, but it's coming along.

This is exactly the problem. We are all working on the same again and 
again. I think we could cooperate and have something we could use all. 
Then when there is bug, one of us just fix it and everybody has it for 
free. Then we can concentrate to work on another advanced features wrt. 
to Ruby debugging. Probably debug-commons could be the place to start? 
:) And in the future it could become something like Java has in the 
JPDA, maybe...

        m.


-------------------------------------------------------------------------
This SF.net email is sponsored by DB2 Express
Download DB2 Express C - the FREE version of DB2 express and take
control of your XML. No limits. Just data. Click to get it now.
http://sourceforge.net/powerbar/db2/
_______________________________________________
Rubyeclipse-development mailing list
Rubyeclipse-development@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/rubyeclipse-development

Reply via email to