rspec users - May 2008 - RailsExampleGroup

So I have a describe which is a subclass of RailsExampleGroup. I inserted
some data into my database, and reconnect! my connection. Before reconnect,
finding the data I just created is fine. The data cannot be found, after
reconnection, the data is gone. It''s not in the database. It looks to
me the
data is not actually inserted into the DB at all, just somehow buffered
somewhere. Would anybody know why? thanks.

Now the workaround is to have a separate class doing all DB operations, but
I am still curious why RSpec does this, for speed reason?

Yi
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
<http://rubyforge.org/pipermail/rspec-users/attachments/20080523/4615b8b5/attachment.html>

On 23 May 2008, at 19:53, Yi Wen wrote:
> It''s not in the database. It looks to me the data is not actually
> inserted into the DB at all, just somehow buffered somewhere.
It''s stored in a transaction, so the only place you can see the data  
is from the same connection.  Make another connection (by  
reconnecting) and you are protected from seeing that data.

Why do you want to see this intermediate state of the database from  
another connection anyway?  It''s an unusual thing to want to do.

Ashley


-- 
http://www.patchspace.co.uk/
http://aviewfromafar.net/

Well, I use this for driving selenium RC. So the approach I am using
is to setup some data, and then invoke selenium to interact with a
test server. Apparently the test server uses another connection. Would
you suggest a better way to handle this? It will be the same for doing
functional testing on web services. What''s the "standard way"
of doing
it?

Yi

On 23 May 2008, at 19:53, Yi Wen wrote:
> It''s not in the database. It looks to me the data is not actually
> inserted into the DB at all, just somehow buffered somewhere.
It''s stored in a transaction, so the only place you can see the data
is from the same connection.  Make another connection (by
reconnecting) and you are protected from seeing that data.

Why do you want to see this intermediate state of the database from
another connection anyway?  It''s an unusual thing to want to do.

Ashley

On 23 May 2008, at 21:35, Yi Wen wrote:
> Well, I use this for driving selenium RC. So the approach I am using
> is to setup some data, and then invoke selenium to interact with a
> test server. Apparently the test server uses another connection. Would
> you suggest a better way to handle this? It will be the same for doing
> functional testing on web services. What''s the "standard
way" of doing
> it?

Ah ok, I get you now.  I did the same thing and it stumped me too, I  
was completely baffled for a whole day.  What you''ve got with selenium-
rc is not this:

+--ruby-process--+
|                |
|     rails      |
|       ^        |
|       |        |----->(database)
|     rspec      |
|                |
+----------------+


but this:

+--ruby-process--+
|                |
|     rails      |--------
|                |        \
+----------------+        |
         ^                 |
         |                 |
+--native-proc---+        |
|                |        |
|     browser    |        |
|                |        |
+----------------+        V
         ^             (database)
         |                 ^
+--java-process--+        |
|                |        |
|  selenium-rc   |        |
|                |        |
+----------------+        |
         ^                 |
         |                 |
+--ruby-process--+        |
|                |        |
|     rspec      |--------/
|                |
+----------------+

Nothing like a bit of ascii art to liven up a mailing list :)

My solution was to create a custom story listener, which is a class  
that looks like this (copied from a post I made in April):

class MyStoryListener
  def run_started(number_of_scenarios); end
  def story_started(story, narrative); end
  def story_ended(story, narrative); end
  def scenario_started(story, scenario); end
  def scenario_succeeded(story, scenario); end
  def scenario_pending(story, scenario, error); end
  def scenario_failed(story, scenario, error); end
  def run_ended; end
  # method_missing allows you to remove hooks you don''t need
  def method_missing(*args); end
end

Spec::Story::Runner.register_listener(MyStoryListener.new)

What I did was to have the run_started hook delete all non-static data  
out of the database.

There''s other things you can do to make the process more useful for  
development.  I''ve wanted for ages to package some code for this  
together.  I''ve got a Merb app I want to start soon, and this will be  
an essential pre-requisite.  So hopefully I''ll get that done soon and  
I''ll post back here.  (However my pet projects usually get  
sidetracked...)


Hope this helps for now anyway.


Ashley


-- 
http://www.patchspace.co.uk/
http://aviewfromafar.net/