Hacker News new | past | comments | ask | show | jobs | submit login

Does this contain the fixes with require to make loading rails and other libraries startup faster? I didn't see anything relating to it in the change logs.



it seems so https://github.com/ruby/ruby/blob/v1_9_3_preview1/ChangeLog#...

Notice this does not solve rails becoming slower every release (I recently recovered a 1.2 project and it runs the tests something like five times faster then a fresh one with 3.0)


I know, right?


Good question. I'm guessing we should at least see some improvement pathname library C rewrite. But I can't tell if the require fix made it into 1.9.3.

For context:

http://www.rubyinside.com/ruby-1-9-3-faster-loading-times-re...


I can confirm that yes, the load.c patch as outlined in the Ruby Inside post IS in Ruby 1.9.3 preview 1 :-)


Yes. That was fixed in r31875 which was committed before the 1.9.3 branch was made.

  http://svn.ruby-lang.org/cgi-bin/viewvc.cgi?view=revision&revision=31875


While the load.c patch made it in this time, is that a good indicator of things making it into 1.9.3 from head? Just thinking about Module#mix here..


Why do you want Module#mix in favor of Module#prepend. The former introduces a whole new system for mixins alongside the old one, while the latter simply makes the existing system work for extending existing classes (a first-class, no caveats alias_method_chain)


I wasn't really flying the flag for Module#mix, just intrigued as to the pattern of things making it into head and then either making it out into tagged releases (or not).

That said, I confess I was a fan of mix, but it might be in my favor for things to get more complicated sometimes.. ;-)


Only bug fixes (not new features like Module#mix) will be merged from trunk to the 1.9.3 branch.


FWIW, the improvements are not that significant (based on the last time I tried the patch a month or so ago).


I had a 36% improvement in a mid-sized Rails 3 app start time using the core load.c patch. Will be trying again on 1.9.3preview1 soon but in theory it should remain similar. DHH also tweeted noting a significant improvement at the time with one of his apps.

It may not be significant to you or in your use case but it seems in some cases there are definitely reasonable gains to be had, which can't be a bad thing.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: