Re: Memory usage regression in core

2015-01-10 Thread Jason van Zyl
What projects were you building? You may very well have triggered a problem. On Jan 10, 2015, at 5:26 PM, Kristian Rosenvold wrote: > Sorry about this, there was something wrong about my environment. > Memory usage is slightly higher, but nothing out of the ordinary. > > Kristian > > > 2015-

Re: Memory usage regression in core

2015-01-10 Thread Jason van Zyl
I think it needs to be fixed so I'll take a look. Igor noticed when I made the change: http://mail-archives.apache.org/mod_mbox/maven-dev/201410.mbox/%3C5448E62D.7040607%40ifedorenko.com%3E But I also feel the change is one that's necessary to get toward an immutable set of projects before the

Re: Memory usage regression in core

2015-01-10 Thread Kristian Rosenvold
Sorry about this, there was something wrong about my environment. Memory usage is slightly higher, but nothing out of the ordinary. Kristian 2015-01-10 22:29 GMT+01:00 Jason van Zyl : > If it's causing you issues I'll look sooner, but it is something I planned to > look at. We're building 900 m

Re: Memory usage regression in core

2015-01-10 Thread Jason van Zyl
If it's causing you issues I'll look sooner, but it is something I planned to look at. We're building 900 modules and we're using more memory but no as drastic as you're seeing. On Jan 10, 2015, at 4:13 PM, Kristian Rosenvold wrote: > I just tracked down the commit that broke memory usage in

Re: Memory usage regression in core

2015-01-10 Thread Jason van Zyl
Some memory increase is expected, as it doesn't allow anything to be loaded lazily. We didn't see as drastic as increase as you saw but I think it reveals other problems in the core. On Jan 10, 2015, at 4:13 PM, Kristian Rosenvold wrote: > I just tracked down the commit that broke memory usag

Memory usage regression in core

2015-01-10 Thread Kristian Rosenvold
I just tracked down the commit that broke memory usage in core, https://github.com/apache/maven/tree/6cf9320942c34bc68205425ab696b1712ace9ba4 it does not revert cleanly towards trunk, I filed MNG-5751 about this issue. Kristian