﻿<?xml version="1.0" encoding="utf-8"?><rss version="2.0"><channel><title>Ayende @ Rahien</title><link>http://ayende.com</link><description>Ayende @ Rahien</description><copyright>Copyright (C) Ayende Rahien  2004 - 2021 (c) 2026</copyright><ttl>60</ttl><item><title>Jason commented on On Fluent NHibernate</title><description>@Sean
  
Check out AllTypes.Pick().FromAssembly() this eliminates the need to manually register each object into Windsor.
</description><link>http://ayende.com/3514/on-fluent-nhibernate#comment12</link><guid>http://ayende.com/3514/on-fluent-nhibernate#comment12</guid><pubDate>Wed, 13 Aug 2008 12:51:42 GMT</pubDate></item><item><title>Sean Kearon commented on On Fluent NHibernate</title><description>@Ayende, no I didn't and I have no idea how I managed to miss that one.  I've just spent five minutes with Windsor and found container.AddComponent&lt;I, T&gt;().  I'm going to get my eyes checked...Thanks :)
  
</description><link>http://ayende.com/3514/on-fluent-nhibernate#comment11</link><guid>http://ayende.com/3514/on-fluent-nhibernate#comment11</guid><pubDate>Wed, 13 Aug 2008 06:59:57 GMT</pubDate></item><item><title>Tobin Harris commented on On Fluent NHibernate</title><description>I agree with @Ayende in that what we *really* want is convention based mapping, which essentially leads to NO mapping (or automatically inferred mapping, to put it another way). 
  
  
I blogged about achieving this by extending the Mapping.Attributes library, but Fluent NHibernate is in a much better position to get this right IMHO.
  
</description><link>http://ayende.com/3514/on-fluent-nhibernate#comment10</link><guid>http://ayende.com/3514/on-fluent-nhibernate#comment10</guid><pubDate>Wed, 13 Aug 2008 05:53:12 GMT</pubDate></item><item><title>Jay commented on On Fluent NHibernate</title><description>foreach(Type maybeEntity in Assembly.Load("My.Model").GetTypes())
  
{
  
    if(maybeEntity.Namespace == "My.Model.Entities")
  
        continue;
  
    yield return maybeEntity;
  
}
  
  
--or--
  
  
return Assembly.Load("My.Model").GetTypes().Where(t =&gt; t.Namespace != "My.Model.Entities");
  
</description><link>http://ayende.com/3514/on-fluent-nhibernate#comment9</link><guid>http://ayende.com/3514/on-fluent-nhibernate#comment9</guid><pubDate>Tue, 12 Aug 2008 22:51:17 GMT</pubDate></item><item><title>Ayende Rahien commented on On Fluent NHibernate</title><description>sean,
  
You do realize that you can fully configure Windsor from your code, right?
</description><link>http://ayende.com/3514/on-fluent-nhibernate#comment8</link><guid>http://ayende.com/3514/on-fluent-nhibernate#comment8</guid><pubDate>Tue, 12 Aug 2008 20:08:08 GMT</pubDate></item><item><title>Sean Kearon commented on On Fluent NHibernate</title><description>I agree with James, I like what I'm seeing, it's extremely exciting and makes NH much more accessible. I'm currently skirting around the edges of deciding to use NH or something else.  I want to replace my current ORM approach in a new commercial project I have bubbling.  I've spent a good amount of time around NH for the last couple of years and the main thing that has always put me off NH is the XML, period.  Loosing the compiler support is not really tenable for me.  The fluent interface is very sweet indeed and I think will be widely attractive.  
  
  
Your approach above is very good too, but less readable than the fluent interface.  P.S. I also choose Ninject and Unity over Winsor solely because I can configure them fluently in code...I really can't stand using XML config.  I also haven't had time to get into the Boo/Bisnor approach yet - although I have a pdf copy of your book, so things may change!
  
  
All in all, I'd be very happy to ditch the XML and buy back the compiler and R# support any day.  
</description><link>http://ayende.com/3514/on-fluent-nhibernate#comment7</link><guid>http://ayende.com/3514/on-fluent-nhibernate#comment7</guid><pubDate>Tue, 12 Aug 2008 20:01:47 GMT</pubDate></item><item><title>James Kovacs commented on On Fluent NHibernate</title><description>Personally I think that Fluent NHibernate is very exciting and promising. The two big things it brings to NHibernate are better refactoring support and better testability. Refactoring support means that ReSharper can perform property renames. (Currently you get runtime errors from NHibernate because the mapping document and class don't agree. Even without ReSharper, a property rename becomes a compile error rather than a run-time error.) Testability is improved with the PersistenceSpecification&lt;T&gt; class as it gives you a standard way to test your mappings. Yes, convention support would be nice, but right now Fluent NHibernate has some nice building blocks to get there. Overall I really like what I'm seeing!
</description><link>http://ayende.com/3514/on-fluent-nhibernate#comment6</link><guid>http://ayende.com/3514/on-fluent-nhibernate#comment6</guid><pubDate>Tue, 12 Aug 2008 17:34:52 GMT</pubDate></item><item><title>Sean commented on On Fluent NHibernate</title><description>I'm very interested in this project and I do see the point. I'm not a fan of AR for nontrivial uses though so that does color my judgement.
  
  
I don't really need it to be ORM neutral, I'm happy with it being just for NH.
</description><link>http://ayende.com/3514/on-fluent-nhibernate#comment5</link><guid>http://ayende.com/3514/on-fluent-nhibernate#comment5</guid><pubDate>Tue, 12 Aug 2008 16:59:44 GMT</pubDate></item><item><title>Mike commented on On Fluent NHibernate</title><description>@James
  
  
Don’t get discouraged you are off to a good start. I would like to see ORM Neutral added to the roadmap so you are not locked to NH.
  
</description><link>http://ayende.com/3514/on-fluent-nhibernate#comment4</link><guid>http://ayende.com/3514/on-fluent-nhibernate#comment4</guid><pubDate>Tue, 12 Aug 2008 16:03:16 GMT</pubDate></item><item><title>Buddy Stein commented on On Fluent NHibernate</title><description>If you don't see the point, then either do I.    Next!
</description><link>http://ayende.com/3514/on-fluent-nhibernate#comment3</link><guid>http://ayende.com/3514/on-fluent-nhibernate#comment3</guid><pubDate>Tue, 12 Aug 2008 14:48:32 GMT</pubDate></item><item><title>James Gregory commented on On Fluent NHibernate</title><description>This project has only been alive (in my hands) for about 3 weeks. It's still very early days yet, and it concerns me that people are passing judgement long before we've been able to fully implement some of our star features.
  
  
Convention support is already available, but it's not complete. We're working on that. We are aiming for eventual full convention based configuration.
</description><link>http://ayende.com/3514/on-fluent-nhibernate#comment2</link><guid>http://ayende.com/3514/on-fluent-nhibernate#comment2</guid><pubDate>Tue, 12 Aug 2008 14:41:06 GMT</pubDate></item><item><title>Jason commented on On Fluent NHibernate</title><description>I think this line
  
if(maybeEntity.Namespace == "My.Model.Entities")
  
                continue;
  
should read be negated. 
  
if(maybeEntity.Namespace != "My.Model.Entities")
  
                continue;
  
or else your returning every object but the domain.
  
  
"Fluent NHibernate, at least in its current stage, requires a mapping class per entity. At this point, I might as well use the XML again."
  
allows for R# refactorings and compile time checks.  You get this with AR, but the argument there is AR "pollutes" the domain.  With FI for NH the code could be located in the infrastructure layer, just like the Mapping files are an embedded resource.
</description><link>http://ayende.com/3514/on-fluent-nhibernate#comment1</link><guid>http://ayende.com/3514/on-fluent-nhibernate#comment1</guid><pubDate>Tue, 12 Aug 2008 14:27:47 GMT</pubDate></item></channel></rss>