Em 23-04-2012 20:04, David Chelimsky escreveu:
On Mon, Apr 23, 2012 at 5:39 PM, Rodrigo Rosenfeld Rosas
<rr.ro...@gmail.com> wrote:
I have a set of examples that should run with a specific set of records in
the database.
Since setting those records is an expensive operation I'd like to perform it
just once per context.
For example:
context 'sample tree' do
before(:all) { create_tree_records }
example ...
end
The problem with this is that while before(:each) and the examples will run
in a transaction that will be rolled back at the end of each example, that
won't happen to the records created/modified by the before(:all) block.
It seems my database vendor (PostgreSQL) supports nested transactions
(savepoints):
http://www.postgresql.org/about/
I should also notice that I'm using Sequel and it seems that transactions
are reentrant in Sequel:
http://cheat.errtheblog.com/s/sequel/
Database#transaction is re-entrant:
DB.transaction do # BEGIN issued only here
DB.transaction
dataset<< {:first_name => 'Inigo', :last_name => 'Montoya'}
end
end # COMMIT issued only here
So, this logic wouldn't work for me. I need a save point in a before(:all),
so that I'd restore it on a before(:each).
Is there any recommendation to make this work in my specs?
Nothing built into RSpec now, but you could disable the Rails
transaction management and use DatabaseCleaner on your own.
Actually I'm not using the Rails transaction management anymore since I
moved all my models to Sequel::Model. I've created a section in my last
article on how to setup Sequel, Rails and RSpec. Please check the RSpec
section:
http://rosenfeld.herokuapp.com/en/articles/2012-04-18-getting-started-with-sequel-in-rails
The issue with using DatabaseCleaner is that it doesn't support nested
transactions too. And I don't want to use the truncate strategy as there
are some tables that shouldn't be cleared.
This is currently being discussed in
https://github.com/rspec/rspec-rails/issues/496 as well. Feel free to
join that convo.
I don't know if this is possible to accomplish with all supported
databases. Maybe there is a simpler approach for PostgreSQL by using
savepoints. I'll make some experiments similar to the approach I've used
in my article to create those savepoints.
I'd just like to know if there was something already done following this
idea, but if there isn't, I guess I'll have to do it myself :(
Thank you very much for your (always) fast response :)
Cheers,
Rodrigo.
_______________________________________________
rspec-users mailing list
rspec-users@rubyforge.org
http://rubyforge.org/mailman/listinfo/rspec-users
