On Thursday 25 March 2010 14:43:08, Hans Dockter wrote:
> I'm not sure what the main reason for that is. There is always the Groovy
> startup time (the Groovy team is working on making this faster). For a
> clean build you also pay for collecting the incremental build information.

I hope, you don't take me wrong!

My ant build is outdated and the incremental part is no more reliable. 
I did not come to gradle for being bored!

I don't like maven and by searching for alternatives I discovered ivy. But ivy 
did not convince me either. Then reading some blogs about the good and bad of 
ivy I read about gradle - and that made me nosy ;)
Then reading the first pages from gradle I already was convinced, that I found 
the tool, I was searching for.

The migration of my build-scripts was harder than expected - I think, there's 
a lot of description missing for people that don't know maven, ivy, groovy 
and gradle ;)
But - after all - two weeks from first touch to get all running is not so bad 
for me.

So - one reason (amoung many others) was the desire to get a reliable 
incremental build without any unwanted surprises.

The clean build was just a way to create similar conditions for ant and 
gradle.

> And I'm sure there are some other areas where we can optimize the execution
> behavior of Gradle.

Any hint is welcome!

> Yet our main focus is to get the _average_ build time for enterprise builds
> down.

No doubt: you did an excellent job with gradle and I'm already at the point, 
that I don't wonna change :)
May be, gradle needs some more newbees with the same missing knowledge like me 
to improve the already very good userguide ...

I like injection too, but it is known, that it is nearly impossible to get rid 
of connections without documentation.

> > I would like to move that commandline parameter to a configuration
> > setting (some day), so that I don't have to use a wrapper script but can
> > select the remote lookup by use of different tasks.
>
> Gradle should provide some DSL elements to express that. Would you mind
> filing a Jira for this with fix for 1.0?

Ok. 

I also like the idea to bind a dependency to a repository (as you stated in 
other post) - so the lookup-time could be decreased dramatically.

kind regards

Geronimo

---------------------------------------------------------------------
To unsubscribe from this list, please visit:

    http://xircles.codehaus.org/manage_email


Reply via email to