[ruby-core:38526] [Backport92 - Backport #5099][Open] Backport r31875 load path performance problem
Aaron Patterson
Issue #5099 has been reported by Aaron Patterson.
----------------------------------------
Backport #5099: Backport r31875 load path performance problem
http://redmine.ruby-lang.org/issues/5099
Author: Aaron Patterson
Status: Open
Priority: Normal
Assignee: Yuki Sonoda
Category:
Target version:
Would it be possible to backport r31875 to 1.9.2? I think 1.9.2 users are still suffering.
Thanks!
Josh Kalderimis
Issue #5099 has been updated by Josh Kalderimis.
Hi Guys,
Please backport this fix to 1.9.2.
Travis CI (http://travis-ci.org) shows some great examples of the issues caused by the slow requires in 1.9.2.
For example the Rails test suite (http://travis-ci.org/#!/rails/rails/builds/64519), especially the Railties tests, explodes from 13 minutes when we test it against 1.8.7 to 28mins on 1.9.2.
The Rails test suite does a lot of isolated testing, so although this is an exaggerated case, it goes to show how this requires issue slows down testing and development in general.
Here are some other tests which are slower to run on 1.9.2:
http://travis-ci.org/#!/intridea/omniauth/builds/57098
http://travis-ci.org/#!/rspec/rspec-rails/builds/64368
and finally
http://travis-ci.org/#!/mongoid/mongoid/builds/64365 (this is a great example of 1.8.7 vs 1.9.2 vs 1.9.3)
Kind Regards,
Josh and the Travis CI team
Michael Klishin
Issue #5099 has been updated by Michael Klishin.
I second this request. It is surprising how little attention this performance regression had gotten. Many applications went from 5 seconds load time to 35. Running a single test in some codebases now takes seconds instead of 0.1th of a second it used to take.
Please backport this to 1.9.2-head and make a release soon. Thank you.
Nicolás Sanguinetti
Issue #5099 has been updated by Nicolás Sanguinetti.
+1!!! :)
Arun Agrawal
Issue #5099 has been updated by Arun Agrawal.
+1 for this fix
Tom Brown
Issue #5099 has been updated by Tom Brown.
+1
Chris Oliver
Issue #5099 has been updated by Chris Oliver.
+1
Tee Parham
Issue #5099 has been updated by Tee Parham.
+1
Luis Lavena
Issue #5099 has been updated by Luis Lavena.
Hello,
I believe Aaron's request, Josh Kalderimis reasoning and explanation are enough to describe this issue.
Please avoid flooding it with meaningless +1 that do not add value. It is clear the importance of this and the 1.9.2 maintainer will determine if the changes can be considered for backport or not.
Thank you.
Michael Klishin
Issue #5099 has been updated by Michael Klishin.
Luis: I am sorry but this and YAML parser bug backports have been ignored for months. It is time to step up and bring this to 1.9.2 maintainers attention. Issues like this make me strongly believe that Ruby Core Team doesn't give a damn about 1.9.2. If 35 freaking seconds is not enough of a reason to *backport*, I don't know what is.
Oleg Bovykin
Issue #5099 has been updated by Oleg Bovykin.
+1
Luis Lavena
Issue #5099 has been updated by Luis Lavena.
Michael Klishin wrote:
> Luis: I am sorry but this and YAML parser bug backports have been ignored for months. It is time to step up and bring this to 1.9.2 maintainers attention. Issues like this make me strongly believe that Ruby Core Team doesn't give a damn about 1.9.2. If 35 freaking seconds is not enough of a reason to *backport*, I don't know what is.
The reasoning behind backports has been explained in the past. Bugs are candidates for backports as long it doesn't alter the default behavior of Ruby/stdlib/etc.
If you want to make more easy the life of the 1.9.2 maintainer, please, checkout the branch and perform the manual backport of the corresponding commits, then attach a diff here for the maintainer properly apply.
Sometimes what you consider "35 freaking seconds" takes time when you need to perform lot of other tasks.
Please, don't be disrespectful on other's time availability. If you're not pleased, then prove you're a better release maintainer and step up.
Chris Cherry
Issue #5099 has been updated by Chris Cherry.
+1, Definitely!
Jan Šedivý
Issue #5099 has been updated by Jan Šedivý.
+1
Chris White
Issue #5099 has been updated by Chris White.
File ruby-1.9.2-p290-load-path-backport.diff added
Attached is a backport patch to be applied against ruby-1.9.2-p290. All of ruby compiles and make test shows no complaints. Though the function signatures are the same and I don't see the patch breaking anything. Please apply an give it a shot to a 1.9.2 in an ISOLATED NON-PRODUCTION ENVIRONMENT, unless that is you're feeling adventurous :)
Bohuslav Blín
Issue #5099 has been updated by Bohuslav Blín.
+1.
Thank's Aaron for opening this painful issue.
Vishnu Atrai
Issue #5099 has been updated by Vishnu Atrai.
It would be better to have it in ruby 1.9.2 . This patch looks good.
Ivan Samsonov
Issue #5099 has been updated by Ivan Samsonov.
+1. It's very important for me and my projects
Emil Ahlbäck
Issue #5099 has been updated by Emil Ahlbäck.
Definitely +1!