Hi,

I have fixed my little bug... 

I prepare a web site and a GIT repository for it.

Regards,

Le vendredi 31 juillet 2009 à 22:04 +0200, Nicolas a écrit :
> Le jeudi 30 juillet 2009 à 15:49 -0400, Chris Frey a écrit :
> > Hi Nicolas,
> > 
> > Thanks very much for your work on this.  I don't have time to process this
> > today, but I hope to soon.
> > 
> > I'm secretly hoping that I'll get the binary packaging issues and distro
> > issues sorted out, and finish all documentation updates in the next week
> > or so.  Then I can release 0.15 officially.  From then on, I want to make
> > monthly releases, and keep the tree in a releasable state at all times.
> 
> Every week, I try to test the Debian package generation from GIT
> repository on Debian SID. No problem for the moment. If I meet an issue,
> I post you a mail.
> 
> In fact, for me before release the 0.15, you should be fix the bug about
> phone duplication (see my mail at 4 july 2009 and most of your
> proposition to fix it)
> 
> > So I may wait until after the release to fully incorporate your patches.
> > August may be JDWP month. :-)
> > 
> > Also, the patches are really big... I don't mind the odd huge patch,
> > but some mailing list subscribers might not like it if it happens too
> > often.  We can use this patch as "starting point 0" if you like, and you can
> > post patches against this to reduce the size.  Or, you could create a
> > git fork on repo.or.cz and post everything there.  Either one is fine with
> > me.
> 
> The patch is big, but doesn't touch the rest of barry.
> 
> I provide two things :
> - a patch to apply on the Barry source.
> - a tarball with bjdwp tool.
> 
> I'm going to wait your release 0.15 so as to include the barry patch.
> 
> I prefer avoid the fork and several branches.
> 
> During august, I'm in holidays... so "bjdwp" will wait september.
> 
> To avoid the big mail here, I'll post only a web link. And I save the
> patch content on my web site. (If I find the time, I create a GIT
> repository on my web server)
> 
> Regards,
> 
> > Thanks!
> > - Chris
> > 
> > 
> > On Thu, Jul 30, 2009 at 06:35:43PM +0200, Nicolas wrote:
> > > Hi,
> > > 
> > > I have worked about JDWP implementation with Barry.
> > > 
> > > At first, you have to apply the patch on GIT barry source in attachment.
> > > 
> > > Then, after re-build barry. You can build "bjdwp" application.
> > > 
> > > For the moment, you can only do :
> > > 
> > > 1??/ In first terminal :
> > > $ bjdwp localhost 8080
> > > 
> > > 2??/ In a second terminal :
> > > $ jdb -connect com.sun.jdi.SocketAttach:hostname=localhost,port=8080
> > > $ jdb> classes
> > > $ jdb> resume
> > > $ jdb> exit
> > > 
> > > You can test the command "jdb> suspend", but there is a little bug :(
> > > 
> > > In the first terminal (bjdwp), you can read JVM message (the
> > > "System.out.println")
> > > 
> > > I have implemented several commands (partially). I have to completed.
> > > 
> > > 
> > > The next step is :
> > > - reverse the BarryDemo.debug file (generated during the compilation).
> > > Indeed, RIM's JDE builds a COD file and a ".debug" file. This file is
> > > obviously usefull for JDWP. Otherwise, the "classes" command line
> > > doesn't show the applications to debug.
> > > - fix the "suspend" command. How ?
> > > 
> > > 
> > > In fact, when the JVM is running, you have to poll the device :
> > > 
> > > >>> URB 2416
> > >     00000000: 05 00 07 00 00 01 40
> > > <<< URB 2396
> > >     00000000: 00 00 0c 00 13 05 01 00 50 02 00 00
> > > <<< URB 2398
> > >     00000000: 05 00 06 00 00 0b
> > > <<< URB 2399
> > >     00000000: 05 00 0f 00 00 09 56 4d 3a 2d 44 41 20 30 0a
> > > 
> > > while a message is defined... If there isn't message :
> > > 
> > > >>> URB 2420
> > >     00000000: 05 00 07 00 00 01 40
> > > <<< URB 2400
> > >     00000000: 00 00 0c 00 13 05 01 00 51 02 00 00
> > > <<< URB 2402
> > >     00000000: 05 00 06 00 00 02
> > > <<< URB 2403
> > >     00000000: 05 00 06 00 00 00
> > > 
> > > You have to call the USB command GET_STATUS :
> > > 
> > > >>> URB 2424
> > >     00000000: 05 00 07 00 00 01 06
> > > <<< URB 2404
> > >     00000000: 00 00 0c 00 13 05 01 00 52 02 00 00
> > > <<< URB 2406
> > >     00000000: 05 00 06 00 00 00
> > > (wait a USB response...)
> > > <<< URB 2407
> > >     00000000: 05 00 06 00 00 01
> > > <<< URB 2408
> > >     00000000: 05 00 05 00 8d
> > > 
> > > This command is very important...
> > > 
> > > You can launch JDB commands only after the URB 2406...
> > > 
> > > So, we have to implement an infinite (or 60 seconde timeout) wait on USB
> > > bulk read... and if "bjdwp" receive an order from JDB, we have to
> > > interrupt the bulk read to excecute the order.
> > > 
> > > To resume, we have two cases :
> > > 
> > > First case :
> > > >>> URB 2424
> > >     00000000: 05 00 07 00 00 01 06
> > > <<< URB 2404
> > >     00000000: 00 00 0c 00 13 05 01 00 52 02 00 00
> > > <<< URB 2406
> > >     00000000: 05 00 06 00 00 00
> > > ...
> > > ...
> > > <<< URB 2407
> > >     00000000: 05 00 06 00 00 01
> > > <<< URB 2408
> > >     00000000: 05 00 05 00 8d
> > > 
> > > Second case :
> > > >>> URB 2424
> > >     00000000: 05 00 07 00 00 01 06
> > > <<< URB 2404
> > >     00000000: 00 00 0c 00 13 05 01 00 52 02 00 00
> > > <<< URB 2406
> > >     00000000: 05 00 06 00 00 00
> > > ...
> > > I have received an order from JDB
> > > I interrupt the bulk read, then, I process...
> > > When, the process is ended.
> > > >>> URB xxx
> > >     00000000: 05 00 07 00 00 01 40
> > > <<< URB xxx
> > >     00000000: 00 00 0c 00 13 05 01 00 51 02 00 00
> > > ...
> > > 
> > > 
> > > Chris, I have made the maximum to use the barry code style (packet,
> > > protocol, traces...) "bjdwp" is coded with the same style than barry.
> > > For the moment, I use "cmake", but when bjdwp will be good and useable,
> > > I will use autotools.
> > > 
> > > 
> > > If you have any questions about bjdwp, you can ask me...
> > > 
> > > 
> > > Regards,
> > > 
> > > -- 
> > > Nicolas VIVIEN
> > 
> > 
> > 
> > 
> > > ------------------------------------------------------------------------------
> > > Let Crystal Reports handle the reporting - Free Crystal Reports 2008 
> > > 30-Day 
> > > trial. Simplify your report design, integration and deployment - and 
> > > focus on 
> > > what you do best, core application coding. Discover what's new with 
> > > Crystal Reports now.  http://p.sf.net/sfu/bobj-july
> > > _______________________________________________
> > > Barry-devel mailing list
> > > Barry-devel@lists.sourceforge.net
> > > https://lists.sourceforge.net/lists/listinfo/barry-devel
> > 
> > 
> > ------------------------------------------------------------------------------
> > Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day 
> > trial. Simplify your report design, integration and deployment - and focus 
> > on 
> > what you do best, core application coding. Discover what's new with 
> > Crystal Reports now.  http://p.sf.net/sfu/bobj-july
> > _______________________________________________
> > Barry-devel mailing list
> > Barry-devel@lists.sourceforge.net
> > https://lists.sourceforge.net/lists/listinfo/barry-devel

-- 
Nicolas VIVIEN


------------------------------------------------------------------------------
Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day 
trial. Simplify your report design, integration and deployment - and focus on 
what you do best, core application coding. Discover what's new with 
Crystal Reports now.  http://p.sf.net/sfu/bobj-july
_______________________________________________
Barry-devel mailing list
Barry-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/barry-devel

Reply via email to