# Fwd: WorkItem Issue #6095 - "IronRuby 1.1.3 Can't count" **URL:** \ **Category:** IronRuby\ **Created:** [May 4, 2011, 11:38pm UTC](https://www.ruby-forum.com/t/fwd-workitem-issue-6095-ironruby-1-1-3-cant-count/206569 "2011-05-04T23:38:18Z")\ **Posts on this page:** 7\ **Page:** 1
**Author:** ![dubstep](https://www.ruby-forum.com/letter_avatar_proxy/v4/letter/d/cdc98d/32.png) [@dubstep](https://www.ruby-forum.com/u/dubstep)\ **Post date:** [May 4, 2011, 11:38pm UTC](https://www.ruby-forum.com/t/fwd-workitem-issue-6095-ironruby-1-1-3-cant-count/206569/1 "2011-05-04T23:38:18Z")
Hey all, I thought I’d look at issue #6095 on codeplex ([http://ironruby.codeplex.com/workitem/6095](http://ironruby.codeplex.com/workitem/6095)) as my first contribution. Having a dig through the code lead me to ArrayOps which then lead me IListOps in the IronRuby.Libraries project. I noticed IListOps.Length had the [RubyMethod(“count”)] attribute method which only returned the arrays count. Have to some digging (and unneeded code writing then reverting) I looked at the ruby doc and saw Array mixed in Enumerable, looked around and found that IListOps was indeed marked to include Enumerable and sure enough found the 3 count methods. One for no parameters, one with, and one with a block that did all they were supposed to do. So the RubyMethod declaration for count in IListOps was overriding the method declarations in Enumerable once the ruby class was put together at runtime. I removed the RubyMethod declaration on IListOps compiled and ran the IronRuby console, tried out all three versions of count and they all worked. Now, I can see why the count declaration on IListOps was put there, because accessing the internal count property on RubyArray is much quicker than counting and iterating over the collection, however, the ruby methods length and size for quick access to the number of elements in the RubyArray. I see two ways of fixing the bug, 1. removing the [RubyMethod(“count”)] declaration from IListOps.Length which would make count inefficent on IList’s. 2. reimplementing the object and block variations of count in IListOps In my humble opinion, 1 would be best, but since this is my first change I’d thought I’d get some input. Cheers, – Nik Radford Mobile: 07884 254 866 Email: removed\_email\_address@domain.invalid – Nik Radford Mobile: 07884 254 866 Email: removed\_email\_address@domain.invalid ---
**Author:** ![nickrahn](https://www.ruby-forum.com/letter_avatar_proxy/v4/letter/n/dfb087/32.png) [@nickrahn](https://www.ruby-forum.com/u/nickrahn)\ **Post date:** [May 5, 2011, 1:14am UTC](https://www.ruby-forum.com/t/fwd-workitem-issue-6095-ironruby-1-1-3-cant-count/206569/2 "2011-05-05T01:14:39Z")
On Wed, May 4, 2011 at 2:02 PM, Nicholas R. [removed\_email\_address@domain.invalid](mailto:removed_email_address@domain.invalid)wrote: > Hey all, > > I thought I’d look at issue #6095 on codeplex > ([http://ironruby.codeplex.com/workitem/6095](http://ironruby.codeplex.com/workitem/6095)) as my first contribution. > I see two ways of fixing the bug, > 1. removing the [RubyMethod(“count”)] declaration from > IListOps.Length which would make count inefficent on IList’s. > 2. reimplementing the object and block variations of count in IListOps > > In my humble opinion, 1 would be best, but since this is my first > change I’d thought I’d get some input. I’m partial to option 2. As Ruby is already a slow language, we should try to optimize it wherever possible. In addition, 2 keeps to the trend in other .NET languages. For example, I believe LINQ takes advantage of existing methods on Lists when it can. Ryan ---
**Author:** ![nickrahn](https://www.ruby-forum.com/letter_avatar_proxy/v4/letter/n/dfb087/32.png) [@nickrahn](https://www.ruby-forum.com/u/nickrahn)\ **Post date:** [May 5, 2011, 10:39am UTC](https://www.ruby-forum.com/t/fwd-workitem-issue-6095-ironruby-1-1-3-cant-count/206569/3 "2011-05-05T10:39:34Z")
> > I’m partial tooption 2. As Ruby is already a slow language, we should try > > to optimize it wherever possible. In addition, 2 keeps to the trend in other > > .NET languages. For example, I believe LINQ takes advantage of existing > > methods on Lists when it can. > > Likewise. Because arrays are used absolutely everywhere in ruby, making a > common method like count slower will have a whole lot of hidden performance > penalties… Anything we can do to make these core classes faster is a win > for everyone Alright, I’ll get onto doing option 2 when I get home this evening. Thanks guys. ---
**Author:** ![nickrahn](https://www.ruby-forum.com/letter_avatar_proxy/v4/letter/n/dfb087/32.png) [@nickrahn](https://www.ruby-forum.com/u/nickrahn)\ **Post date:** [May 5, 2011, 5:04am UTC](https://www.ruby-forum.com/t/fwd-workitem-issue-6095-ironruby-1-1-3-cant-count/206569/4 "2011-05-05T05:04:05Z")
> > > > I’m partial to option 2. As Ruby is already a slow language, we should try > to optimize it wherever possible. In addition, 2 keeps to the trend in other > .NET languages. For example, I believe LINQ takes advantage of existing > methods on Lists when it can. Likewise. Because arrays are used absolutely everywhere in ruby, making a common method like count slower will have a whole lot of hidden performance penalties… Anything we can do to make these core classes faster is a win for everyone ---
**Author:** ![nickrahn](https://www.ruby-forum.com/letter_avatar_proxy/v4/letter/n/dfb087/32.png) [@nickrahn](https://www.ruby-forum.com/u/nickrahn)\ **Post date:** [May 7, 2011, 10:39am UTC](https://www.ruby-forum.com/t/fwd-workitem-issue-6095-ironruby-1-1-3-cant-count/206569/5 "2011-05-07T10:39:05Z")
> Quick question guys, when pushing this to my repo on github, should I > push to a topic branch or push to master? Well I pushed it to a topic branch, just in case you guys required changes 🙂 As the contributing documentation says, would someone mind giving me a code review? Code can be found at [https://github.com/sekhat/main/tree/bugfix/issue-6095](https://github.com/sekhat/main/tree/bugfix/issue-6095) Cheers. ---
**Author:** ![nickrahn](https://www.ruby-forum.com/letter_avatar_proxy/v4/letter/n/dfb087/32.png) [@nickrahn](https://www.ruby-forum.com/u/nickrahn)\ **Post date:** [May 8, 2011, 3:50am UTC](https://www.ruby-forum.com/t/fwd-workitem-issue-6095-ironruby-1-1-3-cant-count/206569/6 "2011-05-08T03:50:00Z")
Yes, this is the right process - fix it in a branch and send a pull request. The change looks great! Merged into main and closed the bug. Thanks, Tomas ---
**Author:** ![nickrahn](https://www.ruby-forum.com/letter_avatar_proxy/v4/letter/n/dfb087/32.png) [@nickrahn](https://www.ruby-forum.com/u/nickrahn)\ **Post date:** [May 7, 2011, 10:09am UTC](https://www.ruby-forum.com/t/fwd-workitem-issue-6095-ironruby-1-1-3-cant-count/206569/7 "2011-05-07T10:09:28Z")
On 5 May 2011 09:32, Nicholas R. [removed\_email\_address@domain.invalid](mailto:removed_email_address@domain.invalid) wrote: > Alright, I’ll get onto doing option 2 when I get home this evening. > > Thanks guys. Quick question guys, when pushing this to my repo on github, should I push to a topic branch or push to master? – Nik Radford Mobile: 07884 254 866 Email: [removed\_email\_address@domain.invalid](mailto:removed_email_address@domain.invalid)