Comments on: Bypassing mass assignment for update_attributes http://blog.lrdesign.com/2009/03/bypassing-mass-assignment-for-update_attributes/ Logical Reality Design: Web Design and Software Development Wed, 14 Jul 2010 20:46:12 +0000 hourly 1 http://wordpress.org/?v=3.0 By: Evan http://blog.lrdesign.com/2009/03/bypassing-mass-assignment-for-update_attributes/comment-page-1/#comment-3103 Evan Tue, 17 Mar 2009 21:22:27 +0000 http://blog.lrdesign.com/?p=51#comment-3103 Oh, indeed - yours is the correct way, and I've modified the post. I had been lazy and not looked up update_attributes in the API, simply assuming that the difference between update_attributes and update_attributes! was the same as the difference between toggle and toggle! ... i.e. that without the bang, it updates the attribs in memory but does not save the record. Instead, of course, it's the same as the difference between save and save!; the difference is not whether it saves, but whether it returns an error if the record is invalid and the save fails. This is a case where a little more consistency from the API would be lovely... :-) Oh, indeed – yours is the correct way, and I’ve modified the post.

I had been lazy and not looked up update_attributes in the API, simply assuming that the difference between update_attributes and update_attributes! was the same as the difference between toggle and toggle! … i.e. that without the bang, it updates the attribs in memory but does not save the record.

Instead, of course, it’s the same as the difference between save and save!; the difference is not whether it saves, but whether it returns an error if the record is invalid and the save fails.

This is a case where a little more consistency from the API would be lovely… :-)

]]> By: Michael Hartl http://blog.lrdesign.com/2009/03/bypassing-mass-assignment-for-update_attributes/comment-page-1/#comment-3100 Michael Hartl Tue, 17 Mar 2009 18:13:00 +0000 http://blog.lrdesign.com/?p=51#comment-3100 What's funny is that I recently added the same methods independently. You (though not the other readers of this blog) can verify this by doing a git pull and a git log on the private project we've been working on together; you'll see that I added unsafe_update_attributes(!) on February 12, and the code is virtually identical to yours. (The only difference, apart from the method order, is a missing 'save' in your unsafe_update_attributes, which I think you need to add, since unsafe_update_attributes should change the attributes in the database as well as the ones in memory.) What’s funny is that I recently added the same methods independently. You (though not the other readers of this blog) can verify this by doing a git pull and a git log on the private project we’ve been working on together; you’ll see that I added unsafe_update_attributes(!) on February 12, and the code is virtually identical to yours. (The only difference, apart from the method order, is a missing ‘save’ in your unsafe_update_attributes, which I think you need to add, since unsafe_update_attributes should change the attributes in the database as well as the ones in memory.)

]]>