Comments for kete's kete  

I actually think this is breaking standard capistrano behavior and thus makes it harder for new Kete users (but previous Cap users) to get started with deployment.

One small refinement for the displays of the total. If they could be links to the equivalent of the “more” link.

You may have fixed this in a later commit, but line 6 seems to be incomplete. It should also specify related_item_type = ‘Topic’. The idea is on the right track though, for topics, you need to count both types of relationship.

Yes, the tests were moved to the generic homepage_test.rb file.

Even though these tests are meant to test that caching is working, they actually test other aspects of the archive by type functionality. Have you moved them to another file? I don’t think you should ditch them in total.

Looks much better, still one quirk for the _flagging_contributor layout. The “created” and “last edited” bits are split onto two lines and shouldn’t be.

Yep, you’re comment is right. I would not do a query per portrait object look up app/controllers/account_controller.rb#376. Instead I would do this outside the loop:

Looks pretty good. Now let’s see some corresponding tests…

See http://kete.lighthouseapp.com/projects/14288-kete/tickets/126 for an update on the situation with regard to search result duplication.

Changes implemented in the following commit:

The commit has since been removed.

How about some local variables for those arrays to cut down duplication:

The commit has since been removed.

The commit has since been removed.

The commit has since been removed.

I’m not really liking the sound of your methodology. Instead of doing the right thing, it reads like you are doing the easy thing. You also didn’t communicate those problems to me earlier.

Well login, add item, logout, and item/basket delete buttons aren’t hard to miss. And changing them to visit not only shaved off 30s, but also got rid of a lot of errors I was getting on REE. Calling two click_buttons or a visit then click_link in a row was causing havoc. Not sure why. Might talk to webrat maintainer.

I would be hesitant to replace most of these “click_link” lines with direct visit calls. If I’m not mistaken, by calling “click_link” we are, thanks to webrat, in effect testing that the link exists. Something we should test!

Of course, substituting “tag.tag_name” with tag.name…

Here’s what I came up with: