I've started this afresh, this is definitely a separate topic :-)
Apologies to the others for hijacking their topic.
Re: attrib4j.
I hadn't noticed that one but will have a look at it tonight. I did find another
before starting, can't remember what it was, but it didn't fit what I believe was
Hiya Bill.
Yep, I know. I was just letting him know that I am already working on some changes to
javassist to support the metadata attributes in the class info stuff.
I was originally wrapping the javassist classes but decided that it made more sense to
do it properly and add the support for t
Sorry, forgot to add that I'm not looking at the compiler part yet, just the class
info side of things.
View the original post :
http://www.jboss.org/index.html?module=bb&op=viewtopic&p=3823415#3823415
Reply to the post :
http://www.jboss.org/index.html?module=bb&op=posting&mode=reply&p=382341
I'm in the middle of the annotation stuff for the meta data. There's not that much
more to do but I've been finding it hard to fit it in over the last week.
View the original post :
http://www.jboss.org/index.html?module=bb&op=viewtopic&p=3823414#3823414
Reply to the post :
http://www.jboss.
Integer.TYPE instead of int.class should do.
View the original post :
http://www.jboss.org/index.html?module=bb&op=viewtopic&p=3822441#3822441
Reply to the post :
http://www.jboss.org/index.html?module=bb&op=posting&mode=reply&p=3822441
---
apers in other contexts but never
the tests. I hadn't realised the problem with longs (and
presumably doubles) existed.
Thanks again,
Kev
Kevin Conner
This is a personal e-mail. Any views or opinions presented are
solely those of the author and do not necessarily represent those
of
> Unlike longs, integer assignments are atomic. If we keep the
> statistics in a volatile int, we make sure different threads see the
> most current value, without the need to lock.
Just a minor point, double and long assigments are atomic if volatile.
Kev
Kevin Conner
was hoping for a bit
> more intelligence for supporting multiple jdk versions concurrently.
Eclipse 1 used to do this on a per project basis, they removed it
when moving to Eclipse 2. I don't know what the reasons are for
this but it is one of the features I miss :-(.
Kev
Kevin Co
that the latest version is obtained from main memory and updates are
prevented.
The only safe use of the volatile keyword is for primitives or for
references
to thread safe objects.
HTH,
Kev
Kevin Conner
This is a personal e-mail. Any views or opinions presented are
solely those of the
> Do you know of JEOPS (sourceforge)? It was the only rules engine I
> could find that had a reasonable license.
Did you look at mandarax?
Kev
Kevin Conner
This is a personal e-mail. Any views or opinions presented are
solely those of the author and do not necessarily represent
left it at that :-(. I suspect it may be similar
to the WebSphere problem that I had to work around.
I'll have a quick look at it again and see if I can get any
further, although we have just upgraded Solaris and it may
no longer happen :-).
Thanks again,
Kev
Kevin Conner
This is a perso
for netboot deployments. There is
also a proposed fix which we would like to see incorporated
into your code if it is still applicable.
Thanks,
Kev
Kevin Conner
This is a personal e-mail. Any views or opinions presented are
solely those of the author and do not necessarily represent those
ice as fast.
The move to 1.4.0 changed the map to an internal implementation, but still
hash based. The code hasn't changed between 1.4.0 and 1.4.1 so, I assume,
the gain is in the VM.
Kev
Kevin Conner
Orchard Information Systems Limited
Newcastle Technopole, Kings Manor
Newcastl
> Please, share the vision.
>
> I would like to contribute here but don't know where to start
> as it seems
> everything might be changing.
I second that.
Kev
Kevin Conner
Orchard Information Systems Limited
Newcastle Technopole, Kings Manor
Newcastle Upon T
databases
(progress) from barfing at identifiers which are also recognised keywords.
Are you interested in this? If so I'll post a patch on Monday morning.
Kev
Kevin Conner
Orchard Information Systems Limited
Newcastle Technopole, Kings Manor
Newcastle Upon Tyne, NE1 6PA. United Kingdo
> Can you post this patch at source forge? Assign it to dsundstrom.
Yep, no problem. It will have to be in the morning though (UK time).
I've made changes to the source after David's comments, just removed
the autocommit/commit/rollback stuff.
Kev
Kevin Conner
Orchar
erstand the
current CMP code correctly?
Thanks for the feedback,
Kev
Kevin Conner
Orchard Information Systems Limited
Newcastle Technopole, Kings Manor
Newcastle Upon Tyne, NE1 6PA. United Kingdom
Registered in England, Number 1900078
Tel: +44 (0)
om head, and then will look at moving it into CVS. It's
> just so that I can try out the changes before updating our version
> of jboss.
Now that 3.2 is out I'll have a look at basing this code on that
release.
Kev
P.S. I've included both mailing lists so that
and you should change it. :-)
Kev
Kevin Conner
Orchard Information Systems Limited
Newcastle Technopole, Kings Manor
Newcastle Upon Tyne, NE1 6PA. United Kingdom
Registered in England, Number 1900078
Tel: +44 (0) 191-2032536 Fax: +44 (0) 191 2302515
-
ant scripts to get it all to play nice.
We must have been lucky I guess :-). We're using version 1.2.1a
and all we've had to do is write our own label incrementer. I'll
watch out for this in future though.
Thanks,
Kev
Kevin Conner
Orchard Information Systems Limited
Newcas
efit of using
this in jboss if there is already a mechanism to build on a regular
interval (which there is). If you were starting from scratch then I
would certainly recommend cruisecontrol but why change if your current
mechanism is sufficient?
Kev
Kevin Conner
Orchard Information System
chines (NT/2000 + JDK1.3.1/JDK1.4.0).
We will eventually use the solaris box to host some of our test environments
but
not until we have some JDKs on at least 1.3.1 :-).
Kev
Kevin Conner
Orchard Information Systems Limited
Newcastle Technopole, Kings Manor
Newcastle Upon Tyne, NE1 6PA. U
Sorry, I must have missed that. Our email is still filtering
through from our earlier outage.
Glad to hear it's under control. Still, it would have been fun :-)
Kev
Kevin Conner
Orchard Information Systems Limited
Newcastle Technopole, Kings Manor
Newcastle Upon Tyne, NE1 6PA.
rrow (once I get progress working with jboss :-))
Kev
Kevin Conner
Orchard Information Systems Limited
Newcastle Technopole, Kings Manor
Newcastle Upon Tyne, NE1 6PA. United Kingdom
Registered in England, Number 1900078
Tel: +44 (0) 191-2032536 Fax: +44 (0) 191 2302515
-
> Kevin Conner wrote:
> > As an aside, having just looked at the 3.0.1 code (this may
> > have changed since then), commit, delistResource, enlistResource,
> > registerSynchronization, rollback and setRollbackOnly are protected
> > by the lock method (and hence sync
nt x ; // instance variable
public int testInc() {
x++;
return x;
}
public int testAdd() {
x = x + 1;
return x;
}
}
Kev
Kevin Conner
Orchard Information Systems Limited
Newcastle Technopole, Kings Manor
Newcastle Upon Tyne, NE1 6PA.
the case? If you do then
I would genuinely like to hear about it.
> If you want to carry on flogging this horse, carry onI'm
> happy as is.
:-)
Kev
Kevin Conner
Orchard Information Systems Limited
Newcastle Technopole, Kings Manor
Newcastle Upon Tyne, NE1 6PA. United Kingdom
Reg
o a local
>
> Exactly my point... :-)
...
I asked the question because I am not aware of any. The compiler
should not be allowed to optimise the access to the local copy
because it changes the semantics of the code. You implied in your
previous email that this may be happening, I am asking
es of the txCapsule
reference. Can the compiler choose to ignore this and make two accesses
to the volatile reference, in essence making it the same as the original?
Kev
Kevin Conner
Orchard Information Systems Limited
Newcastle Technopole, Kings Manor
New
ception ex) {
> > throw new IllegalStateException("No transaction.");
> > - txCapsule.commit();
> > + }
>
> Interesting... Is this actually faster then synchronizing? Have you
> measured it?
How about:
public void commit()
{
fina
30 matches
Mail list logo
