Is it time to start new projects in Rails 3? I'm nervous about using beta versions but at the same time I really like what they are doing and don't want to deal with legacy 2.3.5 issues with these apps.
Is it better to wait these things out, or buckle-up, deal with early adopter issues and get a head start on the future. Thanks for any light you can shed.
I'd say no, rails v3.beta3 still has some serious issues that cause it to crash (unexpectedly). As noted from the Riding Rails blog:
Note that Ruby 1.8.7 p248 and p249 has marshaling bugs that crash both Rails 2.3.x and Rails 3.0.0. Ruby 1.9.1 outright segfaults on Rails 3.0.0, so if you want to use Rails 3 with 1.9.x, jump on 1.9.2 trunk for smooth sailing.
Also of note:
Known regressions: Rails crashes unless configuration.action_controller.session is set, config.thread_safe does not work, Unable to run a RJS partial from an HTML template, Backtrace silencers oftem remove application lines from test failures backtraces, Active Record double escapes error_messages_for
I am doing all of my new development work on Rails 3. I have found a couple of issues in general stability, but they are really in the fringe/edge cases - I have a system running on Heroku that is quite stable. The core platform itself is quite stable and the various APIs should not significantly change heading to release.
You need to ask yourself what you've got to lose, but one thing is clear: Rails 2.x is history.
Related
We have an application which is currently on Rails version 2.3.12 and Ruby version 1.8.7. We want to update our application to Rails 4.0 and Ruby 2.1.0. We have around 200 models and 150 controllers.
I would like to know how big an effort is it to undergo the upgrade process. Also can you provide steps which can be followed for the upgrade. Should we first upgrade Ruby and then Rails or vice-versa?
What you want to achieve is gonna be an epic effort. I can't provide you the step by step instructions because it's impossible to cover all the cases in a single answer.
I recommend to not upgrade Ruby and Rails in parallels, but in small steps.
The complexity of the upgrade itself is huge, but it's not impossible as long as you have a reasonable test coverage of you application. I really hope you have tests. If you don't, then you can't even think to start an upgrade like this and you first need to ensure a minimum coverage for the product.
My suggestion is to follow this roadmap
Upgrade Ruby from 1.8.7 to 1.9.3. Rails 2.3 is supposed to be compatible with Ruby 1.9 and you'll get the benefit of a more recent Ruby version. In fact, several gems have dropped support for Ruby 1.8.
Leave the app settle and run for a while, just to make sure you can find and fix incompatibilities. Every time you find an issue, reproduce it with a test, then fix it.
Upgrade Rails from 2.3 to 3.0.
Same as point 2.
Upgrade Rails from 3.0 to the latest 3.x version.
Same as point 2.
Upgrade Ruby from 1.9.3 to 2.0. The upgrade should be trivial.
Same as point 2.
Upgrade Rails from 3.x to the latest version
A few notes
The reason I suggested you to not upgrade from Rails 2.3 directly to 4, it's because it's almost impossible.
The reason I suggested you to not upgrade form Rails 2.3 directly to the latest 3.x, it's because from 3.0 to 3.x there is the asset pipeline new feature that is gonna cause you several headaches. It's better to handle this step separately and not in a big upgrade.
Ruby 1.9 handles encoding differently than Ruby 1.8. Rails 2.3 is not well prepared for this.
The issues often lead to hard errors, and may occur at unexpected places (like when you want to show a error message)
A good (or scary) overview is on this blog entry:
http://www.rvdh.de/2010/01/06/why-you-cant-run-rails-23-apps-on-ruby-19/
I also lost a lot of time trying to bring Rails 2.3.x to work with Ruby 1.9.x
(Maybe it is possible if you are always use plain ascii)
So you need first update to Rails 3.0 and then to Ruby 1.9.3 (or both at the same time)
The asset pipeline is nice when it works, but upgrading is difficult enough.
So I would just disable it until you are on Rails 4.0.
There is a lot changed with the views, like form_for helper.
It is good when you have tests that render each view at least once. You then will get deprecation warning for things that have changed.
If this is not realistic, then you should try to test all things of your app manually, and
grep the log files for the deprecation warnings.
Many deprecations in Rails 3.0 will lead to errors or wrong behaviour in Rails 3.1.
For updating the ruby and rails version to new is not take much more time but in your code you have to change all the quires and model scopes and whatever which does't work in new rails version of deprecated in new ruby and rails versions.And one more thing may be many gems which have support to ruby 1.8.7 will not support the new ruby version and same thing for the rail.
My suggestions is just, you first check the dependencies of your application with new versions and the get a decisions.I have knowledge about it because i have also upgrade a full application from ruby 1.8.7, rails 2.3.5 to ruby 1.9.3 and rails 3.2.13.
I am running an application on Rails 2.3.8. I am planning to upgrade it to Rails 4.0 (which is in RC). What will be the easiest way for me to do this? Do I need to first upgrade to Rails 3.x?
Note: in my current implementation, I am using starling and ferret; as part of upgrade I am also considering to move to sidekiq and sunspot
This is a multi-step process, and depending on the size of your application, it can take a long time. At each step, you'll want to test your application for bugs and problems and broken gems (because they are most certainly going to crop up). I have included links for the most complicated steps. Here is the path of least pain:
Update to Rails 2.3.18.
Update to Ruby 1.9.3.
Update to Rails 3.0.
Part 1
Part 2
Part 3
Update to Rails 3.1.
Update to Rails 3.2.latest.
Update to Ruby 2.0.latest.
Update to Rails 4.0.
Update from Rails 4.0 to Rails 4.1.
Update from Rails 4.1 to Rails 4.2.
Update from Rails 4.2 to Rails 5.0.
Update from Rails 5.0 to Rails 5.1.
Bonus: If you have a large application, this is going to take a long time. If you have a large team, long-running branches become a huge headache because of recurring merge conflicts. One strategy for mitigating this is to dual boot your application with both versions of Rails so that you can have the new version running on your master branch, rather than on a long-running branch of its own.
Aaron Gray's answer is very useful, but sometimes it's very difficult after upgrading to find the differences between rails version in the code. I mean, many times after upgrading something doesn't work. Maybe could be some changes in the file or some file could be added or removed in newer version. There is a nice tool, how you can find out all differences between version and compare them. It's on Rails differences and could be very helpful to discover some potential mistakes and bugs.
Rails 3 is out and every one is excited (etc etc).
However, I'm not ready to update to it yet for a couple reasons:
Not all gems I use are supported.
Just finally got my head around 2.3.x.
My client expects software that is reliable.
So, right now I'm running 2.3.5 and will be upgrading to 2.3.8 soon. But what about any security patches or parallel feature updates that 3 might get? Will there be a 2.3.9? Will there be a 2.4? Does anyone know?
Looking on https://rails.lighthouseapp.com/projects/8994-ruby-on-rails there is a 2.3.10 milestone, so they will at least have that.
You can see from the commit log that there are still commits to the 2.3 branch eventhough it's not as active as the 3.x branch. 37signals are still using 2.3 on many of their apps that's why you will see many patches/commits coming from Jeremy, one of 37signals employee. I don't think they will make any major changes that will break your apps though.
Judging by the lastest commit to 2.3.9pre ("preparing for 2.3.9") and the open tickets (none), it looks like Rails 2.3.9 will be released any day now.
Actually upgrade to 3.0 doesn't break that much. There's few gotchas, but they are mostly very well documented. Most common gems are already compatibile with rails 3 (at least the newest beta versions). I am sure, that by the time you would finish upgrading all of those would have stable relases.
I would go rather for 3.0 instead of going for an upgrade in 2.3.x branch.
It is very crucial to have lots of well placed tests for your application. This way you can easily see if anything breaks.
I found much many problems after switching to Ruby 1.9.x, then after going from Rails 2.3.x to Rails 3. Actually now I am using REE 1.8.7 with Rails3.
I'm developing an application with Ruby on Rails that I want to maintain for at least a few years, so I'm concerned about the next version coming up soon.
Going from Rails 1 to Rails 2 was such a big pain that I didn't bother and froze my gems and let the application die, alone, in the dark.
On this project I don't want to do that. First because this new version looks awesome, but also because this application may turn into a real product.
How can I prepare my application so that it will be upgradable with as little changes as possible.
How time consuming do you think switching version will be?
And what about my server? Deployment?
I'm already looking at deprecation notices... what else can I do?
The best thing you could do would be to follow development of Rails 3 via blogs and the Github repository and keep up a copy of your app along with it.
The official Ruby on Rails blog is updated with "What's new in Edge" posts every once in awhile. There are other blogs that often write about new things in edge as well. Larger features are often highlighted in these blogs, so you know about all the cool new features you can play with.
I'm not sure how close Rails 3 is to release (last I heard the core team was talking about a release at RailsConf 2009 in May), but you can always freeze the edge version of Rails into your application and just see what breaks. If you are using git, or another DVCS, you might make a branch specifically for Rails 3 and periodically update Rails to the latest edge code. Just be aware that edge Rails is a moving target so things in your app may break or fix themselves as you are pulling in newer Rails code.
Update:
Jeremy McAnally has a ton of info on upgrading from Rails 2 to Rails 3 on his blog.
http://omgbloglol.com/
I don't think there is going to be a major problem. Going off what was said in that initial report the Rails team realized that they can't do a major rewrite like they did from 1 to 2.
They even say:
I’m sure there’ll be some parts of Rails 3 that are incompatible, but we’ll try to keep them to a minimum and make it really easy to convert a Rails 2.x application to Rails 3.
I would be more concerned going from Merb to Rails 3.
The single most important thing you can do to make it easy to migrate to a new version of rails is to have a comprehensive test suite. Without a good test suite, I would never have the confidence that the new version of rails hasn't broken something in my app. On the current Rails app I'm working on, we started on Rails 2.1.1 back in October of 2008. Since then, we've migrated to Rails 2.1.2, 2.2.2, 2.3.2, 2.3.3 and now 2.3.4. I did the migrations to 2.3.2, 2.3.3 and 2.3.4...and for the 2.3.2 and 2.3.3 upgrades, we had some failing tests that alerted us to problems we would not have discovered without having such a good test suite. The failing tests actually alerted us to a regressive bug in rails that there was a patch for on the Rails lighthouse but that was not included in the release (since it was discovered, right after the release).
Once you've got that test suite in place, just stay current with each rails release (waiting a couple weeks to upgrade is fine, just don't skip any of the releases).
Yehuda Katz (a member of the Rails core team) has stated that there will most likely be a transitional release, containing deprecation warnings and such.
So as long as you have a good test suite to expose the inevitable upgrade problems, and stay current with the Rails release, the migration to Rails 3 should not be too difficult.
As simple as:
One
Two
Three
Great screencasts from Ryan Bates.
For preparing your application, the best way it what Jared said. Follow the Rails3 development.
For the time consuming, I think it depends of how you've followed the rails3 development before it's release.
And for the deployment, it shouldn't take too much problems. Rails 3 will be using Rack. So you can start it with mongrel, passenger or any server/gateway it shouldn't give you any problem.
There are some major changes in Rails 3, I posted about my experience upgrading my app to Rails 3 here: http://rails3.community-tracker.com/permalinks/5/notes-from-the-field-upgrading-to-rails-3
A good start in preparing would be to migrate over to using bundler. And doing a very deep review of strings that will go through the new XSS protection scheme.
There are going to be some automated compatibility checkers. Also, keep an eye on http://www.railsplugins.org/ so that you know if the libraries you depend on are going to be upgraded. The Rails Core team seems to be giving a lot of advance notice to the community this time around, so any lib that is actively maintained should be good to go.
Just do one thing
take a backup of your old version project first and then
on terminal(command prompt) write
rails new path/of/the/project
for example if my 2.3.* project is at home/rails_projects/myproject then
rails new home/rails_projects/myproject
or
cd home/rails_projects
rails new myproject
It will ask if there is any modifications done in any /config or other files. Do appropriate.
About half a year ago, when I started to learn Ruby and Rails, I first tried Ruby 1.9 but I soon gave up, because at that time nothing worked out of the box and almost every helping blog or tutorial was designed for Ruby 1.8.
What about now? (Dec 2009) Is it possible to get an existing Rails application running by a Ruby and Rails beginner without running into problems which can only be fixed by an absolute Ruby and Rails professional?
Unfortunately I dind't have good experiences with Ruby 1.9 and Rails.
You can read more here: Has anyone successfully deployed a Rails project with Ruby 1.9.1?
My opinion is that migrating an existing Rails app from Ruby 1.8.x to Ruby 1.9.1 is not as easy as you would think, event with an excellent test suite.
I'm also quite sure that most of the problems arise from trying to convert an existing application because you are working with an established code base.
Starting with a new Rails app with Ruby 1.9 should probably easier because you can trace a problem as soon as you write a single line of code so you can easily isolate which component is not compatible.
When migrating an existing app I had hard time trying to figure out which stack level was actually not compatible with Ruby 1.9. And there are more than one incompatible libraries at the same time I can't tell you how it's difficult to understand which one should be fixed first and which one originated the final error.
In 99.9% of the cases yes, there were rare cases where u might find problems but they should be solved with the new release.
As for the external gems and plugins, most of them now are fully compatible with ruby 1.9, however sepcial cases might exist but I'm not aware of any right now.