[cucumber] Performance of 0.1.99?


#1

Has anyone noticed a slowdown in performance of their cucumber suite
from 0.1 to 0.2alpha?

I haven’t ruled out something in our suite itself, but the slowdown
appeared to occur concurrently with our upgrade to 0.2alpha.

Cheers,
Luke

Luke M.
removed_email_address@domain.invalid
http://www.lukemelia.com/


#2

On Sat, Mar 14, 2009 at 7:00 AM, Luke M. removed_email_address@domain.invalid wrote:

Has anyone noticed a slowdown in performance of their cucumber suite from
0.1 to 0.2alpha?

I haven’t ruled out something in our suite itself, but the slowdown
appeared to occur concurrently with our upgrade to 0.2alpha.

I haven’t noticed anything. Do you want to create a benchmark?

Aslak


#3

On 14 Mar 2009, at 10:06, aslak hellesoy wrote:

On Sat, Mar 14, 2009 at 7:00 AM, Luke M. removed_email_address@domain.invalid
wrote:
Has anyone noticed a slowdown in performance of their cucumber suite
from 0.1 to 0.2alpha?

Ours is getting slower all the time (as we add more features) so it’s
very hard to say. I’ve noticed it seems to take a little longer to
crank up because the features are all parsed first - I’m not sure that
happened before.

Matt W.
http://blog.mattwynne.net
http://www.songkick.com


#4

On Mar 14, 2009, at 6:06 AM, aslak hellesoy wrote:

On Sat, Mar 14, 2009 at 7:00 AM, Luke M. removed_email_address@domain.invalid
wrote:
Has anyone noticed a slowdown in performance of their cucumber suite
from 0.1 to 0.2alpha?

I haven’t noticed anything. Do you want to create a benchmark?

I just ran a subset of our suite with 0.1.x and then with master.
Times were pretty close. Guess I need to keep hunting for the
culprit. :slight_smile:

Cheers,
Luke


#5

Matt W. wrote:

happened before.
Cuke parsed all the features (well the ones you tell it to) in its
previous incarnations. The major difference is the new Abstract syntax
tree structure where we represent a more complex object structure than
before (but not massively I think). And hence we have more complex tree
traversal but again I don’t see it being huge.

It would be interesting to see some performance comparisons.


Joseph W.
http://blog.josephwilk.net