Re: For review: Clojure Namespace Isolation

2014-12-08 Thread Ralph Ritoch
Thank you. I have signed the contributor agreement and applied to join clojure-dev (request pending). On Monday, December 8, 2014 3:38:15 AM UTC+8, Alan Forrester wrote: On 7 December 2014 at 19:27, Fluid Dynamics a209...@trbvm.com javascript: wrote: On Sunday, December 7, 2014 11:12:37

Re: For review: Clojure Namespace Isolation

2014-12-08 Thread Gary Verhaegen
Naive question: does that solve the dependency problem where two explicit dependencies depend on different versions of a transitive one and something breaks because of it? On Monday, 8 December 2014, Ralph Ritoch rrit...@gmail.com wrote: Thank you. I have signed the contributor agreement and

Re: For review: Clojure Namespace Isolation

2014-12-08 Thread Ralph Ritoch
Gary, What your stating is possible with this, but it wouldn't be automatic. You could load the one dependency in one isolation, and load the second dependency in a second isolation. This is where a bit of hard work would be needed to keep the pointers returned when you create the two

Re: For review: Clojure Namespace Isolation

2014-12-08 Thread Ralph Ritoch
Gary, One additional note. When your about to use a dependency you will still need to enter into the isolation of that dependency for any feature that accesses namespaces not in your current environment (such as any call to the dependency which uses conflicting namespaces). Not every feature

Re: For review: Clojure Namespace Isolation

2014-12-08 Thread Ralph Ritoch
One important piece of information about this feature is that it doesn't manipulate classloaders. This feature was designed to be used with OSGi or applications that handle their own classloading issues. In a typical application when entering a namespace isolation you may also need to bind a

Re: For review: Clojure Namespace Isolation

2014-12-08 Thread Gary Verhaegen
Thanks for these clarifications! On Monday, 8 December 2014, Ralph Ritoch rrit...@gmail.com wrote: One important piece of information about this feature is that it doesn't manipulate classloaders. This feature was designed to be used with OSGi or applications that handle their own

Re: For review: Clojure Namespace Isolation

2014-12-07 Thread Bozhidar Batsov
I think you should post this to clojure-...@googlegroups.com —  Cheers, Bozhidar On December 7, 2014 at 07:32:38, Ralph Ritoch (rrit...@gmail.com) wrote: Hello,    I have created a way to create namespace isolation within clojure. The code is available on github @

Re: For review: Clojure Namespace Isolation

2014-12-07 Thread Fluid Dynamics
On Sunday, December 7, 2014 11:12:37 AM UTC-5, Bozhidar Batsov wrote: I think you should post this to cloju...@googlegroups.com To where? Your post displayed like that, and I couldn't get a usable tooltip expanding the address by mousing over it either. -- You received this message because

Re: For review: Clojure Namespace Isolation

2014-12-07 Thread Alan Forrester
On 7 December 2014 at 19:27, Fluid Dynamics a2093...@trbvm.com wrote: On Sunday, December 7, 2014 11:12:37 AM UTC-5, Bozhidar Batsov wrote: I think you should post this to cloju...@googlegroups.com To where? Your post displayed like that, and I couldn't get a usable tooltip expanding the

For review: Clojure Namespace Isolation

2014-12-06 Thread Ralph Ritoch
Hello, I have created a way to create namespace isolation within clojure. The code is available on github @ https://github.com/rritoch/clojure . What this new code does is it replaces the clojure.lang.Namespace/namespaces static property with a clojure.lang.NamespaceContainer object that