Eclipse Community Forums
Forum Search:

Search      Help    Register    Login    Home
Home » Language IDEs » AJDT » AJDT 1.2.1 any new releases planned to fix memory usage
AJDT 1.2.1 any new releases planned to fix memory usage [message #63239] Thu, 30 March 2006 14:54 Go to next message
Eclipse User
Originally posted by: schuchert.yahoo.com

I noticed that there are fixes to the development build of AJDT 1.3 that
reduces the memory used by 50%. I could really use this but there is not a
development build for 1.2.1, which leads me to believe that 1.2.1 is
officially dead.

I'm stuck using RAD, which is tied to Eclipse 3.0. Are there any plans to
release another version of ADJT for Eclipse 3.0, or should I plan to stick
to using ant tasks to build.

When I use AJDT, eclipse will use all of 1.2 GB before, which causes
problems on my measly 2 GB system. I have given RAD more memory, but then
I start swapping when I am also running a server, which makes the
performance so bad it is not usable.

Thanks!

Brett
Re: AJDT 1.2.1 any new releases planned to fix memory usage [message #63262 is a reply to message #63239] Thu, 30 March 2006 18:13 Go to previous messageGo to next message
Matt Chapman is currently offline Matt Chapman
Messages: 429
Registered: July 2009
Senior Member
Hi Brett,

It's not dead, just resting! We have the capability to produce a 1.2.2
build - no-one has asked for one until now. Putting the upcoming release
of AspectJ 1.5.1 into AJDT 1.2 will take some work, but I think it should
be possible. It's not possible to backport many of the AJDT changes
because they are dependent on the eclipse version, but most of the memory
improvements are in the compiler. We will try this after AJDT 1.3.1 is
done (which should be next week).

Regards,

Matt.

Brett L. Schuchert wrote:

> I noticed that there are fixes to the development build of AJDT 1.3 that
> reduces the memory used by 50%. I could really use this but there is not a
> development build for 1.2.1, which leads me to believe that 1.2.1 is
> officially dead.

> I'm stuck using RAD, which is tied to Eclipse 3.0. Are there any plans to
> release another version of ADJT for Eclipse 3.0, or should I plan to stick
> to using ant tasks to build.

> When I use AJDT, eclipse will use all of 1.2 GB before, which causes
> problems on my measly 2 GB system. I have given RAD more memory, but then
> I start swapping when I am also running a server, which makes the
> performance so bad it is not usable.

> Thanks!

> Brett
Re: AJDT 1.2.1 any new releases planned to fix memory usage [message #63353 is a reply to message #63262] Fri, 31 March 2006 12:45 Go to previous messageGo to next message
Eclipse User
Originally posted by: schuchert.yahoo.com

Matt,

First, thanks for your extremely quick reply. I appreciate it.

We've been using AspectJ since 11/04. In our initial uses we only needed
to weave one project (manually) to get our stuff to work (required for
both test and production). The build process weaves other aspects that
were really only required for produciton.

We've recently added a few more aspects for test support. One requires
weaving in 4 projects (minimum, 7 typically) for it to work. We organize
our applications by projects and layers within a project. Each project
has, typically, 8 projects (EAR, WAR, EJB, Business, Integraion, Domain,
Utility, Test). We also have a core application (reference architecture)
upon which other projects depend. A typical workspace as aroune 20
projects and I'd guess 500K - 1M SLOC.

This one aspect mocks out all of our DAO's by recording input and
responses and then plays them back on subsequent test executions. It
speads up our tests by roughtly 60 - 80%, which makes our developers more
likely to use them.

We have around 40 developers and AJDT would make their life much easier.
I'd go so far as to say that without it, this new mocking of DAO's would
effectivley be useless.

However, we only have 2 gigs of memory and without the memory improvements
we are right on the edge of usable. So any improvement there would be a
major boon.

We are less concerned with AspectJ 1.5.1, but given the opportunity to
move to it, we would. Since we are using RAD and WebSphere 6.0.2, we are
suck using JDK 1.4.2. I'm under the impression that most of the
differences from 1.2 to 1.5 are related to support for annotations and
generices, which we are ready to use as soon as we can. I'd love to see
IBM update WebSphere and RAD so we could use JDK 1.5. I suppose I should
be luck, we started with JDK 1.2 (10/02), skipped 1.3 and move to 1.4.2
around 8 months ago...

Again, thank you. I have no experience building eclipse plugings, but I do
have a lot of experienc with WSAD and a good deal with RAD (which is very
bugy). If you can use my help in any way, I'm happy to offer it.
Re: AJDT 1.2.1 any new releases planned to fix memory usage [message #63376 is a reply to message #63262] Fri, 31 March 2006 12:46 Go to previous messageGo to next message
Eclipse User
Originally posted by: schuchert.yahoo.com

Matt,

First, thanks for your extremely quick reply. I appreciate it.

We've been using AspectJ since 11/04. In our initial uses we only needed
to weave one project (manually) to get our stuff to work (required for
both test and production). The build process weaves other aspects that
were really only required for production.

We've recently added a few more aspects for test support. One requires
weaving in 4 projects (minimum, 7 typically) for it to work. We organize
our applications by projects and layers within a project. Each project
has, typically, 8 projects (EAR, WAR, EJB, Business, Integration, Domain,
Utility, Test). We also have a core application (reference architecture)
upon which other projects depend. A typical workspace as around 20
projects and I'd guess 500K - 1M SLOC.

This one aspect mocks out all of our DAO's by recording input and
responses and then plays them back on subsequent test executions. It
speeds up our tests by roughly 60 - 80%, which makes our developers more
likely to use them.

We have around 40 developers and AJDT would make their life much easier.
I'd go so far as to say that without it, this new mocking of DAO's would
effectively be useless.

However, we only have 2 gigs of memory and without the memory improvements
we are right on the edge of usable. So any improvement there would be a
major boon.

We are less concerned with AspectJ 1.5.1, but given the opportunity to
move to it, we would. Since we are using RAD and WebSphere 6.0.2, we are
suck using JDK 1.4.2. I'm under the impression that most of the
differences from 1.2 to 1.5 are related to support for annotations and
generics, which we are ready to use as soon as we can. I'd love to see IBM
update WebSphere and RAD so we could use JDK 1.5. I suppose I should be
luck, we started with JDK 1.2 (10/02), skipped 1.3 and move to 1.4.2
around 8 months ago...

Again, thank you. I have no experience building eclipse plugins, but I do
have a lot of experience with WSAD and a good deal with RAD (which is very
buggy). If you can use my help in any way, I'm happy to offer it.
Re: AJDT 1.2.1 any new releases planned to fix memory usage [message #63582 is a reply to message #63376] Mon, 03 April 2006 11:38 Go to previous messageGo to next message
Matt Chapman is currently offline Matt Chapman
Messages: 429
Registered: July 2009
Senior Member
Brett,

Thanks for the information, it's very interesting to hear about your
AspectJ usage.

I thought from your first post that you are using AJDT 1.2.1 (the release
versions, from December 20, 2005). Is that correct? If so, that contains
AspectJ 5, so you are already using it. It runs just fine on JDK 1.4.2 -
Java 5 is only required for those specific features.

The AspectJ 1.5.1 release will contain the memory improvements, so we can
then try to produce an AJDT 1.2.2 release containing that.

Regards,

Matt.

Brett L. Schuchert wrote:
> Matt,

> First, thanks for your extremely quick reply. I appreciate it.

> We've been using AspectJ since 11/04. In our initial uses we only needed
> to weave one project (manually) to get our stuff to work (required for
> both test and production). The build process weaves other aspects that
> were really only required for production.

> We've recently added a few more aspects for test support. One requires
> weaving in 4 projects (minimum, 7 typically) for it to work. We organize
> our applications by projects and layers within a project. Each project
> has, typically, 8 projects (EAR, WAR, EJB, Business, Integration, Domain,
> Utility, Test). We also have a core application (reference architecture)
> upon which other projects depend. A typical workspace as around 20
> projects and I'd guess 500K - 1M SLOC.

> This one aspect mocks out all of our DAO's by recording input and
> responses and then plays them back on subsequent test executions. It
> speeds up our tests by roughly 60 - 80%, which makes our developers more
> likely to use them.

> We have around 40 developers and AJDT would make their life much easier.
> I'd go so far as to say that without it, this new mocking of DAO's would
> effectively be useless.

> However, we only have 2 gigs of memory and without the memory improvements
> we are right on the edge of usable. So any improvement there would be a
> major boon.

> We are less concerned with AspectJ 1.5.1, but given the opportunity to
> move to it, we would. Since we are using RAD and WebSphere 6.0.2, we are
> suck using JDK 1.4.2. I'm under the impression that most of the
> differences from 1.2 to 1.5 are related to support for annotations and
> generics, which we are ready to use as soon as we can. I'd love to see IBM
> update WebSphere and RAD so we could use JDK 1.5. I suppose I should be
> luck, we started with JDK 1.2 (10/02), skipped 1.3 and move to 1.4.2
> around 8 months ago...

> Again, thank you. I have no experience building eclipse plugins, but I do
> have a lot of experience with WSAD and a good deal with RAD (which is very
> buggy). If you can use my help in any way, I'm happy to offer it.
Re: AJDT 1.2.1 any new releases planned to fix memory usage [message #63671 is a reply to message #63582] Mon, 03 April 2006 14:26 Go to previous messageGo to next message
Eclipse User
Originally posted by: schuchert.yahoo.com

Matt Chapman wrote:

> Brett,

> I thought from your first post that you are using AJDT 1.2.1 (the release
> versions, from December 20, 2005). Is that correct? If so, that contains

That is correct, we are using AJDT 1.2.1.

> The AspectJ 1.5.1 release will contain the memory improvements, so we can
> then try to produce an AJDT 1.2.2 release containing that.

OK, then I was confused. I didn't realize that the memory performance was
in ApsectJ. For some reason I was thinking it was related to AJDT, not
AspectJ.

That being the case, we ARE interested in AspectJ 1.5.1.

Thanks,

Brett
Re: AJDT 1.2.1 any new releases planned to fix memory usage [message #65551 is a reply to message #63671] Fri, 26 May 2006 16:46 Go to previous message
Eclipse User
Originally posted by: matthew.smith.valtech.com

Quick follow-up:

With the release of AJDT 1.2.2 for eclipse 3, we saw reduced memory usage
in RAD, but it was not enough to get us going with AJDT. We've since
moved into a Load-time weaving approach with RAD 6 (eclipse 3) and ibm jdk
1.4.2 for our development environments. (Based on some of the work
described here:
http://blogs.codehaus.org/people/avasseur/archives/001140_as pectj_5_load_time_weaving_with_java_13_using_aspectwerkz.htm l)

- Matt
Re: AJDT 1.2.1 any new releases planned to fix memory usage [message #592751 is a reply to message #63239] Thu, 30 March 2006 18:13 Go to previous message
Matt Chapman is currently offline Matt Chapman
Messages: 429
Registered: July 2009
Senior Member
Hi Brett,

It's not dead, just resting! We have the capability to produce a 1.2.2
build - no-one has asked for one until now. Putting the upcoming release
of AspectJ 1.5.1 into AJDT 1.2 will take some work, but I think it should
be possible. It's not possible to backport many of the AJDT changes
because they are dependent on the eclipse version, but most of the memory
improvements are in the compiler. We will try this after AJDT 1.3.1 is
done (which should be next week).

Regards,

Matt.

Brett L. Schuchert wrote:

> I noticed that there are fixes to the development build of AJDT 1.3 that
> reduces the memory used by 50%. I could really use this but there is not a
> development build for 1.2.1, which leads me to believe that 1.2.1 is
> officially dead.

> I'm stuck using RAD, which is tied to Eclipse 3.0. Are there any plans to
> release another version of ADJT for Eclipse 3.0, or should I plan to stick
> to using ant tasks to build.

> When I use AJDT, eclipse will use all of 1.2 GB before, which causes
> problems on my measly 2 GB system. I have given RAD more memory, but then
> I start swapping when I am also running a server, which makes the
> performance so bad it is not usable.

> Thanks!

> Brett
Re: AJDT 1.2.1 any new releases planned to fix memory usage [message #592803 is a reply to message #63262] Fri, 31 March 2006 12:45 Go to previous message
Eclipse User
Originally posted by: schuchert.yahoo.com

Matt,

First, thanks for your extremely quick reply. I appreciate it.

We've been using AspectJ since 11/04. In our initial uses we only needed
to weave one project (manually) to get our stuff to work (required for
both test and production). The build process weaves other aspects that
were really only required for produciton.

We've recently added a few more aspects for test support. One requires
weaving in 4 projects (minimum, 7 typically) for it to work. We organize
our applications by projects and layers within a project. Each project
has, typically, 8 projects (EAR, WAR, EJB, Business, Integraion, Domain,
Utility, Test). We also have a core application (reference architecture)
upon which other projects depend. A typical workspace as aroune 20
projects and I'd guess 500K - 1M SLOC.

This one aspect mocks out all of our DAO's by recording input and
responses and then plays them back on subsequent test executions. It
speads up our tests by roughtly 60 - 80%, which makes our developers more
likely to use them.

We have around 40 developers and AJDT would make their life much easier.
I'd go so far as to say that without it, this new mocking of DAO's would
effectivley be useless.

However, we only have 2 gigs of memory and without the memory improvements
we are right on the edge of usable. So any improvement there would be a
major boon.

We are less concerned with AspectJ 1.5.1, but given the opportunity to
move to it, we would. Since we are using RAD and WebSphere 6.0.2, we are
suck using JDK 1.4.2. I'm under the impression that most of the
differences from 1.2 to 1.5 are related to support for annotations and
generices, which we are ready to use as soon as we can. I'd love to see
IBM update WebSphere and RAD so we could use JDK 1.5. I suppose I should
be luck, we started with JDK 1.2 (10/02), skipped 1.3 and move to 1.4.2
around 8 months ago...

Again, thank you. I have no experience building eclipse plugings, but I do
have a lot of experienc with WSAD and a good deal with RAD (which is very
bugy). If you can use my help in any way, I'm happy to offer it.
Re: AJDT 1.2.1 any new releases planned to fix memory usage [message #592817 is a reply to message #63262] Fri, 31 March 2006 12:46 Go to previous message
Eclipse User
Originally posted by: schuchert.yahoo.com

Matt,

First, thanks for your extremely quick reply. I appreciate it.

We've been using AspectJ since 11/04. In our initial uses we only needed
to weave one project (manually) to get our stuff to work (required for
both test and production). The build process weaves other aspects that
were really only required for production.

We've recently added a few more aspects for test support. One requires
weaving in 4 projects (minimum, 7 typically) for it to work. We organize
our applications by projects and layers within a project. Each project
has, typically, 8 projects (EAR, WAR, EJB, Business, Integration, Domain,
Utility, Test). We also have a core application (reference architecture)
upon which other projects depend. A typical workspace as around 20
projects and I'd guess 500K - 1M SLOC.

This one aspect mocks out all of our DAO's by recording input and
responses and then plays them back on subsequent test executions. It
speeds up our tests by roughly 60 - 80%, which makes our developers more
likely to use them.

We have around 40 developers and AJDT would make their life much easier.
I'd go so far as to say that without it, this new mocking of DAO's would
effectively be useless.

However, we only have 2 gigs of memory and without the memory improvements
we are right on the edge of usable. So any improvement there would be a
major boon.

We are less concerned with AspectJ 1.5.1, but given the opportunity to
move to it, we would. Since we are using RAD and WebSphere 6.0.2, we are
suck using JDK 1.4.2. I'm under the impression that most of the
differences from 1.2 to 1.5 are related to support for annotations and
generics, which we are ready to use as soon as we can. I'd love to see IBM
update WebSphere and RAD so we could use JDK 1.5. I suppose I should be
luck, we started with JDK 1.2 (10/02), skipped 1.3 and move to 1.4.2
around 8 months ago...

Again, thank you. I have no experience building eclipse plugins, but I do
have a lot of experience with WSAD and a good deal with RAD (which is very
buggy). If you can use my help in any way, I'm happy to offer it.
Re: AJDT 1.2.1 any new releases planned to fix memory usage [message #592943 is a reply to message #63376] Mon, 03 April 2006 11:38 Go to previous message
Matt Chapman is currently offline Matt Chapman
Messages: 429
Registered: July 2009
Senior Member
Brett,

Thanks for the information, it's very interesting to hear about your
AspectJ usage.

I thought from your first post that you are using AJDT 1.2.1 (the release
versions, from December 20, 2005). Is that correct? If so, that contains
AspectJ 5, so you are already using it. It runs just fine on JDK 1.4.2 -
Java 5 is only required for those specific features.

The AspectJ 1.5.1 release will contain the memory improvements, so we can
then try to produce an AJDT 1.2.2 release containing that.

Regards,

Matt.

Brett L. Schuchert wrote:
> Matt,

> First, thanks for your extremely quick reply. I appreciate it.

> We've been using AspectJ since 11/04. In our initial uses we only needed
> to weave one project (manually) to get our stuff to work (required for
> both test and production). The build process weaves other aspects that
> were really only required for production.

> We've recently added a few more aspects for test support. One requires
> weaving in 4 projects (minimum, 7 typically) for it to work. We organize
> our applications by projects and layers within a project. Each project
> has, typically, 8 projects (EAR, WAR, EJB, Business, Integration, Domain,
> Utility, Test). We also have a core application (reference architecture)
> upon which other projects depend. A typical workspace as around 20
> projects and I'd guess 500K - 1M SLOC.

> This one aspect mocks out all of our DAO's by recording input and
> responses and then plays them back on subsequent test executions. It
> speeds up our tests by roughly 60 - 80%, which makes our developers more
> likely to use them.

> We have around 40 developers and AJDT would make their life much easier.
> I'd go so far as to say that without it, this new mocking of DAO's would
> effectively be useless.

> However, we only have 2 gigs of memory and without the memory improvements
> we are right on the edge of usable. So any improvement there would be a
> major boon.

> We are less concerned with AspectJ 1.5.1, but given the opportunity to
> move to it, we would. Since we are using RAD and WebSphere 6.0.2, we are
> suck using JDK 1.4.2. I'm under the impression that most of the
> differences from 1.2 to 1.5 are related to support for annotations and
> generics, which we are ready to use as soon as we can. I'd love to see IBM
> update WebSphere and RAD so we could use JDK 1.5. I suppose I should be
> luck, we started with JDK 1.2 (10/02), skipped 1.3 and move to 1.4.2
> around 8 months ago...

> Again, thank you. I have no experience building eclipse plugins, but I do
> have a lot of experience with WSAD and a good deal with RAD (which is very
> buggy). If you can use my help in any way, I'm happy to offer it.
Re: AJDT 1.2.1 any new releases planned to fix memory usage [message #592992 is a reply to message #63582] Mon, 03 April 2006 14:26 Go to previous message
Eclipse User
Originally posted by: schuchert.yahoo.com

Matt Chapman wrote:

> Brett,

> I thought from your first post that you are using AJDT 1.2.1 (the release
> versions, from December 20, 2005). Is that correct? If so, that contains

That is correct, we are using AJDT 1.2.1.

> The AspectJ 1.5.1 release will contain the memory improvements, so we can
> then try to produce an AJDT 1.2.2 release containing that.

OK, then I was confused. I didn't realize that the memory performance was
in ApsectJ. For some reason I was thinking it was related to AJDT, not
AspectJ.

That being the case, we ARE interested in AspectJ 1.5.1.

Thanks,

Brett
Re: AJDT 1.2.1 any new releases planned to fix memory usage [message #594024 is a reply to message #63671] Fri, 26 May 2006 16:46 Go to previous message
Matthew Smith is currently offline Matthew Smith
Messages: 2
Registered: July 2009
Junior Member
Quick follow-up:

With the release of AJDT 1.2.2 for eclipse 3, we saw reduced memory usage
in RAD, but it was not enough to get us going with AJDT. We've since
moved into a Load-time weaving approach with RAD 6 (eclipse 3) and ibm jdk
1.4.2 for our development environments. (Based on some of the work
described here:
http://blogs.codehaus.org/people/avasseur/archives/001140_as pectj_5_load_time_weaving_with_java_13_using_aspectwerkz.htm l)

- Matt
Previous Topic:A question about adding AspectJ nature to a plugin project programmatically
Next Topic:A question about adding AspectJ nature to a plugin project programmatically
Goto Forum:
  


Current Time: Wed Jul 30 15:26:56 EDT 2014

Powered by FUDForum. Page generated in 0.02703 seconds