You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@maven.apache.org by Jerome Lacoste <je...@coffeebreaks.org> on 2004/06/08 03:19:20 UTC
Re: ant vs m1 vs m2
On Fri, 2004-05-28 at 09:43, Jason van Zyl wrote:
> Howdy,
>
> Just for anyone interested in following the dev of m2:
>
> <jason> ant:
> <jason> real 0m3.678s
> <jason> user 0m2.270s
> <jason> sys 0m0.140s
>
> <jason> m1:
> <jason> real 0m5.600s
> <jason> user 0m4.180s
> <jason> sys 0m0.080s
>
> <jason> m2:
> <jason> real 0m2.468s
> <jason> user 0m0.910s
> <jason> sys 0m0.000s
>
> This is for the most simple of builds with one source file. Just really
> an illustration of initialization time.
Interesting.
Do you know where does that difference come from?
Is there a way to enable profiling? I remember Jira having a very nice filter allowing to enable profiling. I wonder if it's possible and interesting to introduced something like that within maven.
But I guess that maven is more predictable than an application using various SQL queries on top of various DBMS.
Jerome
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@maven.apache.org
For additional commands, e-mail: dev-help@maven.apache.org
Re: ant vs m1 vs m2
Posted by Jerome Lacoste <je...@coffeebreaks.org>.
On Tue, 2004-06-08 at 02:13, Jason van Zyl wrote:
> On Mon, 2004-06-07 at 21:19, Jerome Lacoste wrote:
> >
> > > This is for the most simple of builds with one source file. Just really
> > > an illustration of initialization time.
> >
> > Interesting.
> >
> > Do you know where does that difference come from?
>
> As I haven't even started optimizing or profiling I would speculate it
> results from trying to shed all dependencies possible, and using tools
> like the xpp3 parser to serialize/deserialize the model and thus
> avoiding the grossly ineffecient commons-digester/beanutils/collections.
>
> > Is there a way to enable profiling?
>
> No, I would just use a profiler or possibly an aspect.
>
> > I remember Jira having a very nice filter allowing to enable profiling. I wonder if it's possible and interesting to introduced something like that within maven.
>
> I don't really see any benefit in doing that i.e. building that into
> maven when you can do profiling with a tool.
See my response below
> > But I guess that maven is more predictable than an application using various SQL queries on top of various DBMS.
>
> It's not a matter of being predictable, but rather doing analysis where
> appropriate. I wouldn't bother introducing a complicated mechanism in
> maven itself for profiling when an external tool would be much more
> appropriate.
The benefit in Jira is that:
- first it is very simple to use and even to integrate to the app
- it is run by the user, not by the developer
The second point is crucial, because you may not have access to the user's environment and/or data.
So he might face a problem you may never be able to reproduce and sometimes you cannot ask a user to do the profiling job.
After it's just a question of how much support the dev team wants to provide.
For personal curiosity I will see if I can allocate some cycles to use the thing on maven.
Jerome
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@maven.apache.org
For additional commands, e-mail: dev-help@maven.apache.org
Re: ant vs m1 vs m2
Posted by Jason van Zyl <jv...@maven.org>.
On Mon, 2004-06-07 at 21:19, Jerome Lacoste wrote:
>
> > This is for the most simple of builds with one source file. Just really
> > an illustration of initialization time.
>
> Interesting.
>
> Do you know where does that difference come from?
As I haven't even started optimizing or profiling I would speculate it
results from trying to shed all dependencies possible, and using tools
like the xpp3 parser to serialize/deserialize the model and thus
avoiding the grossly ineffecient commons-digester/beanutils/collections.
> Is there a way to enable profiling?
No, I would just use a profiler or possibly an aspect.
> I remember Jira having a very nice filter allowing to enable profiling. I wonder if it's possible and interesting to introduced something like that within maven.
I don't really see any benefit in doing that i.e. building that into
maven when you can do profiling with a tool.
> But I guess that maven is more predictable than an application using various SQL queries on top of various DBMS.
It's not a matter of being predictable, but rather doing analysis where
appropriate. I wouldn't bother introducing a complicated mechanism in
maven itself for profiling when an external tool would be much more
appropriate.
>
> Jerome
>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: dev-unsubscribe@maven.apache.org
> For additional commands, e-mail: dev-help@maven.apache.org
--
jvz.
Jason van Zyl
jason@maven.org
http://maven.apache.org
happiness is like a butterfly: the more you chase it, the more it will
elude you, but if you turn your attention to other things, it will come
and sit softly on your shoulder ...
-- Thoreau
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@maven.apache.org
For additional commands, e-mail: dev-help@maven.apache.org
Re: ant vs m1 vs m2
Posted by Jason van Zyl <jv...@maven.org>.
On Tue, 2004-06-08 at 00:27, Brett Porter wrote:
> Hey Jeff,
>
> That looks nice! I'll definitely take that for a spin on my new project. For me,
> I especially like the servlet filter.
>
> m2 guys - worth using this in m2, or is there already something available in
> plexus that does likewise?
A simple aspect does the trick as you don't need any profiling
capabilities during runtime in an app this assuming you done the work up
front during development. But AspectJ 1.2 would allow this at runtime if
you so desired. We also now have licenses for JProfiler and YourKit to
actually profile the m2 code.
--
jvz.
Jason van Zyl
jason@maven.org
http://maven.apache.org
happiness is like a butterfly: the more you chase it, the more it will
elude you, but if you turn your attention to other things, it will come
and sit softly on your shoulder ...
-- Thoreau
Re: ant vs m1 vs m2
Posted by Brett Porter <br...@apache.org>.
Hey Jeff,
That looks nice! I'll definitely take that for a spin on my new project. For me,
I especially like the servlet filter.
m2 guys - worth using this in m2, or is there already something available in
plexus that does likewise?
- Brett
Quoting Jeff Turner <je...@apache.org>:
> On Mon, Jun 07, 2004 at 09:19:20PM -0400, Jerome Lacoste wrote:
> > On Fri, 2004-05-28 at 09:43, Jason van Zyl wrote:
> [...]
> > Is there a way to enable profiling? I remember Jira having a very nice
> > filter allowing to enable profiling. I wonder if it's possible and
> > interesting to introduced something like that within maven.
>
> http://opensource.atlassian.com/profiling/
>
>
> --Jeff
>
> > But I guess that maven is more predictable than an application using
> > various SQL queries on top of various DBMS.
> >
> > Jerome
> >
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: dev-unsubscribe@maven.apache.org
> For additional commands, e-mail: dev-help@maven.apache.org
>
Re: ant vs m1 vs m2
Posted by Jeff Turner <je...@apache.org>.
On Mon, Jun 07, 2004 at 09:19:20PM -0400, Jerome Lacoste wrote:
> On Fri, 2004-05-28 at 09:43, Jason van Zyl wrote:
[...]
> Is there a way to enable profiling? I remember Jira having a very nice
> filter allowing to enable profiling. I wonder if it's possible and
> interesting to introduced something like that within maven.
http://opensource.atlassian.com/profiling/
--Jeff
> But I guess that maven is more predictable than an application using
> various SQL queries on top of various DBMS.
>
> Jerome
>
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@maven.apache.org
For additional commands, e-mail: dev-help@maven.apache.org