Scott Hernandez wrote:

> I want to bring this up as a topic of its own. We have discussed what 
> the behavior of loading assemblies and custom tasks before, but I 
> wanted to make sure there was consensus, and a clear understanding of 
> how things work.
>
>  
>
> Let me start by stating how it works now (as far as I can tell).
>
> 1.) The Project ctor loads the executing Assembly and any Assembly 
> named *Tasks.dll. The Assemblies are scans and classes that are 
> derived from Task (not abstract) are added (from the BaseDirectory of 
> the executing Domain).
>
> 2.) The TaskDefTask can also add an Assembly during a run.
>
>  
>
seems like a good summation to me

> Note: With my latest changes to TaskFactory, The Project.BaseDirectory 
> is also scanned for *Tasks.dll(s).
>
>  
>
why scan the project base directory for tasks ? Is it a good idea to put 
task dlls at the base of a source tree ?

> Now, I think we have discussed that the TaskDefTask should go away. I 
> also think there was some discussion about the directory names, and 
> locations we should be scanning. It seems to me that some of this 
> stuff should go in a NAnt configuration file.
>
>  
>
TaskDefTask should definately die. The path list is a fine candidate for 
somthing to go into the nant config file.  By default we should only 
scan the nant bin directory and maybe a /tasks below it. Anything else 
would come from the config file. I imagine for most uses the default 
directories will be sufficient.

Ian




_______________________________________________________________

Have big pipes? SourceForge.net is looking for download mirrors. We supply
the hardware. You get the recognition. Email Us: [EMAIL PROTECTED]
_______________________________________________
Nant-developers mailing list
[EMAIL PROTECTED]
https://lists.sourceforge.net/lists/listinfo/nant-developers

Reply via email to