RE: [aspectj-dev] AspectJ 1.5.2 rc1 available

["Wes" <wes@xxxxxxxxxxxxxx>]

I'm very sympathetic to most of Ron's suggestions/bugs, but I'd be inclined to
stay on the 3.2 train unless there is a serious regression that has no good
workaround, and to only make well-understood or well-tested fixes to the 
release candidate.

E.g., is the workaround for 148727 to use the Aspects class?
Can evicting non-aspects types be tested?
(actual answers in the bugs, of course)

Wes

> ------------Original Message------------
> From: "Ron Bodkin" <rbodkin@xxxxxxxxxxxxxx>
> To: "'AspectJ developer discussions'" <aspectj-dev@xxxxxxxxxxx>
> Date: Tue, Jun-27-2006 0:38 AM
> Subject: RE: [aspectj-dev] AspectJ 1.5.2 rc1 available
>
> Hi Andy,
> 
> It's exciting to see the 1.5.2 release close to completion. There are 
> three
> bug fixes and one big enhancement that I would really like to see 
> included
> in 1.5.2 final:
> 
> Bug fixes:
> 148727 - the change to marking aspect API methods aspectOf and has 
> Aspect
> with synthetic breaks any application which calls these methods on 
> library
> aspects (like ours).
> 
> 126118/145689 - with AJDT 20060616121018 / AspectJ 
> 1.5.2.20060530125805, I
> still see cases where e.g., missing types cause build failures with no
> problems reported. I think it's vital that AspectJ 1.5.2 and AJDT 1.4.0 
> ship
> with reliable problem reporting in Eclipse. I just switched to the 
> latest
> dev build on Eclipse 3.2 RC7 so hopefully this one is fixed? But if 
> not, I
> really think it's worth holding the release to get this right... It 
> creates
> a lot of problems for experienced developers and great concern/adoption
> issues among newer developers/projects: if you can't rely on finding
> problems in your Java code when you switch to AJDT, it makes it hard to 
> make
> the switch...
> 
> 145693 - after initial submission, I added a test to this case that 
> shows a
> problem with the weaver resolving types that are on the inpath so they 
> need
> to be added to the classpath too, which is something I don't think one
> should ever do.
> 
> Enhancements:
> 148773 - this is a small change to the load-time weaving world that 
> allows
> evicting non-aspect types (i.e., makes them expendable). This 
> significantly
> reduces overhead if you weave into a lot of types (e.g., it cuts total
> AspectJ memory overhead on startup in a test configuration from 160 to 
> 90%).
> This patch couldn't be submitted until recently, when LTWWorld was 
> finally
> integrated into the codebase...
> 
> -----Original Message-----
> From: aspectj-dev-bounces@xxxxxxxxxxx
> [mailto:aspectj-dev-bounces@xxxxxxxxxxx] On Behalf Of Andy Clement
> Sent: Monday, June 26, 2006 9:29 AM
> To: aspectj-users@xxxxxxxxxxx; AspectJ developer discussions
> Subject: [aspectj-dev] AspectJ 1.5.2 rc1 available
> 
> AspectJ1.5.2 release candidate 1 is now available from the AspectJ
> download page:
> 
> http://www.eclipse.org/aspectj/downloads.php
> 
> It includes over 60 bugs and enhancements we have addressed since
> AspectJ1.5.1a - see the readme for more details:
> 
> http://www.eclipse.org/aspectj/doc/released/README-152.html
> 
> (The readme includes a bugzilla query that lists all the changes in
> AspectJ1.5.2)
> 
> We would like to be in line with the release schedule for Eclipse 3.2
> and so we urge you to download and try this release candidate as soon
> as possible and raise any issues you find in Bugzilla.  We plan to go
> final at the end of this week if no serious problems are found.
> 
> The AJDT 1.4 release candidate for Eclipse 3.2 includes 1.5.2rc1.
> 
> enjoy!
> 
> The AspectJ Team.
> _______________________________________________
> aspectj-dev mailing list
> aspectj-dev@xxxxxxxxxxx
> https://dev.eclipse.org/mailman/listinfo/aspectj-dev
> 
> 
> _______________________________________________
> aspectj-dev mailing list
> aspectj-dev@xxxxxxxxxxx
> https://dev.eclipse.org/mailman/listinfo/aspectj-dev
>