On Sat, Dec 11, 2004 at 10:09:04PM +0000, Ganesh Sittampalam wrote:
> I think we're pretty much ready to release, and doing so shortly after the
> Soot release would be sensible. I hope this is OK with everyone.
>
> Please look through the optimisationsnew and differencesnew files
> (unlinked URLs under http://abc.comlab.ox.ac.uk/) and check them for
> accuracy. I don't really know what to say about interprocedural stuff, so
> that section is just a very rough attempt. I think we should say
> something, given that we have a tech report/PLDI submission claiming we
> did them.
I've replaced the interprocedural blurb with something more accurate.
And, BTW, we *DID* do them and they *ARE* available for download from
the Soot Subversion repository, and they *DO* work with the abc that
we're about to release. In general, it's very rare for a research
publication to be accompanied by a stable released piece of software
supported for industry use; many papers are not even accompanied by a
public repository like we already have for Paddle. And it won't even be
published until June!
I'm sorry, but for the longest time, until 10 days before the PLDI
deadline, all the reweaving and interprocedural stuff was "not a
priority". All of a sudden, that line has changed to "how come there's
not a supported stable release of them yet". There *would* have been a
stable Paddle release by now if I hadn't taken eight months away from
it to work on abc instead. Don't get me wrong; I enjoyed and continue
to enjoy the abc project. However, working on the abc project does not
contribute to me getting this stuff done, and if I don't get it done on
time, *I* will have to face the consequences. The people at NSERC and
my potential future employers are unlikely to consider things like "I
worked on abc instead" or "I instead spent my time hand-holding industry
users to make sense of a research framework" as acceptable excuses for
not having a degree.
I'm sorry if I sound a bit irritated. In real life, when one is only
very mildly irritated, it is immediately apparent to others, which
tends to help one not become more irritated. In e-mail, these kinds of
things don't communicate as easily. I would prefer to err on the side of
sounding more irritated than I am in e-mail if it contributes to keeping
me calm in reality.
Ondrej
> Here's an updated draft release announcement. Again, the section about
> AspectJ/Java 5 is a very rough draft. I've dropped everything else about
> future plans, because I think they're too up in the air. The contents of
> the CHANGELOG file (without the funny punctuation :-) will also be
> included.
>
> Cheers,
>
> Ganesh
>
> ---
> We are pleased to announce the release of abc 1.0.1, an incremental
> improvement on abc 1.0.0 incorporating some bug fixes and performance
> improvements (most notably from inlining advice bodies where appropriate).
> A full changelog can be found below.
>
> The release is a little later than we had hoped because we have been
> putting quite a lot of effort into discussing the design of AspectJ 5 with
> the ajc team. Although we will not be able to provide a release of abc
> supporting Java 5 or the new AspectJ features as quickly as ajc, it is a
> priority for us to do so as fast as possible.
>
> The release can be downloaded from http://abc.comlab.ox.ac.uk/download
> We would like to encourage all our users to register on our download page
> if they have not already done so; it is very helpful to have this
> information when applying for grants that will enable us to continue abc
> development.
>
> We would also like to encourage everyone that is extending abc in some way
> to make themselves known to us on the abc-dev mailing list. In particular
> it would be enormously helpful to know which internal interfaces in abc
> you are relying on; we have adopted a policy of redesigning the internals
> of abc where this is necessary to keep the code base maintainable and
> support future development, but in the places where extenders are relying
> on interfaces, we will try to keep this to a minimum and where necessary
> help to adapt your code. In the changelog below we have included a special
> section listing some small changes that we believe may affect you.
>
>
>
Received on Sun Dec 12 07:43:04 2004
This archive was generated by hypermail 2.1.8 : Sun Dec 12 2004 - 09:30:03 GMT