Re: [ajdt-dev] Hot Code Replace and LTW with Kepler and AJDT 2.2.3

[Andrew Eisenberg <andrew@xxxxxxxxxxxx>]

Hi Thomas,

AspectJ is not as good as pure Java at hot swapping.  If you are
running with load-time weaving and you edit a class that has an aspect
applied to it, then it is highly likely that you will not get hot
swapping to work.  If you edit a class that is not involved with any
aspects, then hot swapping should work as in Java.  I have not noticed
any difference when working in Kepler compared to other versions (and
I always launch my runtime workbenches with Equinox Weaving turned on
when working on AJDT).

Have you noticed a change compared to older versions of Eclipse?  A
difference in JDK would be more likely to affect things than Eclipse
version.

regards,
Andrew

ps- I'm not working for SpringSource any more so, I don't have much
time to work on AJDT these days.  Andy Clement will also be able to
look into problems.

On Wed, Oct 16, 2013 at 7:51 AM, Thomas Hofmann <email@xxxxxxxxxxxxxxxxx> wrote:
> Hi all,
>
> I am having trouble getting Hot Code Replace working in the debugger when my
> classes are woven with LTW.
>
> My setup is as follows:
>
> Eclipse Kepler 4.3.1
>
> and
>
> AJDT org.eclipse.ajdt_2.2.3.e43x-RELEASE-20130627-0800 which contains
> org.aspectj_1.7.3.20130613144500-a
>
> org.eclipse.equinox.weaving.sdk_1.0.200.I20130319-1000
>
> For compiler compliance settings I am using Java SE 1.7.
>
> I am developing an Kepler Eclipse RCP based application that makes use of
> Equinox Weaving.
>
> Every time a change something as simple as adding a new line in one of my
> classes I get an error about hot code replace not being possible. The
> details show the message "Scheme changed not implemented".
>
> I have searched google and found for example
> https://bugs.eclipse.org/bugs/show_bug.cgi?id=120375. All the posts that I
> found make me believe that this should be working with a current version of
> AspectJ / AJDT.
>
> I am asking whether or not this is a supported scenario and if it is how I
> could determine what causes the problem in my setup.
>
> Andrew E. could you please comment on this too?
>
> Thanks for your attention.
>
> Regards, Thomas
>
>
>
> _______________________________________________
> ajdt-dev mailing list
> ajdt-dev@xxxxxxxxxxx
> https://dev.eclipse.org/mailman/listinfo/ajdt-dev
>