Re: Using immutables.org for Java code

2014-12-17 Thread Maxim Khutornenko
Sure, that's a nice feature. Though its application would be limited to unit tests as we almost never change mutable objects in prod code. On Wed, Dec 17, 2014 at 1:03 PM, Kevin Sweeney wrote: > One of the most exciting features of immutables.org collections (over our > custom-generated ones) are

Re: Using immutables.org for Java code

2014-12-17 Thread Kevin Sweeney
One of the most exciting features of immutables.org collections (over our custom-generated ones) are the structural-sharing "wither" methods [1]. So instead of 2 deep copies in our version: ITaskConfig productionTaskConfig = ITaskConfig.build(taskConfig.newBuilder().setProduction(true)); we do a

Re: Using immutables.org for Java code

2014-12-16 Thread Kevin Sweeney
I think adoption would be blocked on an upgrade to Guava 18.0, which we should probably do anyway On Tue, Dec 16, 2014 at 10:31 AM, Zameer Manji wrote: > Nothing in particular stands out to me but I think it would be a nicer > overall solution then our bespoke compiler. > Perhaps if we ever feel

Re: Using immutables.org for Java code

2014-12-16 Thread Zameer Manji
Nothing in particular stands out to me but I think it would be a nicer overall solution then our bespoke compiler. Perhaps if we ever feel the need to change the compiler we should first evaluate if we can move to a library like immutables.org. On Tue, Dec 16, 2014 at 9:44 AM, Maxim Khutornenko w

Re: Using immutables.org for Java code

2014-12-16 Thread Maxim Khutornenko
It looks quite interesting as a general case for immutability. As for replacing our immutable thrift implementation, is there anything in particular you noticed in immutables.org that we currently lack?