Daniel Guettler's profile

Daniel Guettler commented at 1:53 PM
@Jeremy : The reason why I kept them separate is because they need slightly different treatment to actually get the helpers. Therefor I think it is best to keep...

Daniel Guettler commented at 11:02 PM
My original tests for Rails 2.3.5 came out at 6.1 however today they ran slightly faster and come out at 5.0
```
Rails 2.3.5
======================================...
```
Daniel Guettler commented at 10:41 PM
Aaron I can confirm even better performance after latest patches
```
Rails 3 before 'Minor performance improvment in notifications/fanout and active_record/log_su...
```

Daniel Guettler commented at 9:42 PM

Follow up on my last post adding some benchmarks


Base case Rails 2.3.5
=================================================
>> puts Benchmark.measure{5000.times{...

Daniel Guettler commented at 8:59 PM
I don't think it is needed to define all the methods. What I tried to do before with little performance impact was to call the methods directly on the relation ...
Daniel Guettler commented at 7:34 PM

@aaron : I attached updated patch.

@drjoke : no this doesn't have any impact on find yet. Performance hit in find comes more or less completely from Arel

Daniel Guettler commented at 1:02 PM
Ok, attached patch against the latest version of rails master following the above guidelines.
Latest version now comes pretty close to the 2.3.5 base case:
```
Rai...
```
Daniel Guettler commented at 5:14 AM
Well you can get a little but more performance out of it if you modify the log_subscriber 'sql' method a tiny bit.
```
Rails 3 with modified log_subscriber
=======...
```
Daniel Guettler commented at 4:53 AM
@Patrik : Unfortunately you're right. I could squeeze out a tiny performance gain by changing the code a little, but most of the remaining difference comes from...