4 Rail Environments Upgraded. 0 Problems. Great work everyone, I really appreciate it. ~ Ben -- Ben Reubenstein http://www.benr75.com
Ben Reubenstein wrote:> 4 Rail Environments Upgraded. 0 Problems. Great work everyone, I > really appreciate it. > > ~ Ben > -- > Ben Reubenstein > http://www.benr75.comI upgraded to Rails 1.1 and found it to have only a few issues. Don''t be surprised if sites that worked on 1.0 stop functioning on 1.1. For one example, see: http://www.ruby-forum.com/topic/59801 I also read the site5 upgraded to rails 1.1 and then rolled back to 1.0 because of problems reported by users. I hope a small bugfix release addressing the top ~10 most annoying bugs (like regressions) come out soon. It would be great to see rails 1.1.x as the default on hosted servers so new users can benefit from all the great new features before they learn to freeze edge. Frankly, given the huge number of changes and many new features, it is simply amazing there hasn''t been more problems with rails 1.1. Overall a very impressive release indeed. Huge thanks to all the core members and people who submitted patches. -- Posted via http://www.ruby-forum.com/.
> I upgraded to Rails 1.1 and found it to have only a few issues. Don''t be > surprised if sites that worked on 1.0 stop functioning on 1.1. For one > example, see: > > http://www.ruby-forum.com/topic/59801That page actually has a recommendation disguised as a complaint, the bit where they say "because Rails doesn''t have a VERSION constant" -- this could be a useful thing to add. -- Giles Bowkett www.gilesgoatboy.org
You can specify a version constant if you are running edge rails. Now would be a good time to start using edge rails so that you can lock into the 1.1 release. On Thu, 2006-03-30 at 10:06 -0700, Giles Bowkett wrote:> > I upgraded to Rails 1.1 and found it to have only a few issues. Don''t be > > surprised if sites that worked on 1.0 stop functioning on 1.1. For one > > example, see: > > > > http://www.ruby-forum.com/topic/59801 > > That page actually has a recommendation disguised as a complaint, the > bit where they say "because Rails doesn''t have a VERSION constant" -- > this could be a useful thing to add. > > -- > Giles Bowkett > www.gilesgoatboy.org > _______________________________________________ > Rails mailing list > Rails@lists.rubyonrails.org > http://lists.rubyonrails.org/mailman/listinfo/railsCharlie Bowman http://www.recentrambles.com -------------- next part -------------- An HTML attachment was scrubbed... URL: http://wrath.rubyonrails.org/pipermail/rails/attachments/20060330/b59d7dec/attachment.html
> I hope a small bugfix release addressing the top ~10 most annoying bugs > (like regressions) come out soon. It would be great to see rails 1.1.x > as the default on hosted servers so new users can benefit from all the > great new features before they learn to freeze edge.Do you have such a top 10? The only major issues I''ve heard of is Prototype performance in IE (fixed), Oracle adapter issue (fixed), and a failure to communicate that the environment.rb style from 0.13.x would need updating. Regarding the problem of freezing, we''ll be making that mandatory on "rails myapp" now. It''ll automatically extract all the gems into vendor/rails. All applications should be distributed with the version of Rails they''re certified to work with in vendor/rails. I hope we''ll see a new typo release shortly that does this. -- David Heinemeier Hansson http://www.loudthinking.com -- Broadcasting Brain http://www.basecamphq.com -- Online project management http://www.backpackit.com -- Personal information manager http://www.rubyonrails.com -- Web-application framework
David Heinemeier Hansson wrote:>> I hope a small bugfix release addressing the top ~10 most annoying bugs >> (like regressions) come out soon. It would be great to see rails 1.1.x >> as the default on hosted servers so new users can benefit from all the >> great new features before they learn to freeze edge. > > Do you have such a top 10? The only major issues I''ve heard of is > Prototype performance in IE (fixed), Oracle adapter issue (fixed), and > a failure to communicate that the environment.rb style from 0.13.x > would need updating. >The list should be decided by you and core members to maintain objectivity. For example, the Oracle adapter issue is very serious to those that use Oracle--but not for me. Tickets about regressions (things that worked in previous revisions) should probably be included unless the changes are too time-consuming or risky for a bugfix release. For example #3704 and #4234 are probably worth investigating. After regressions, tickets like #4378 are probably worth considering because it can bite a lot of people without warning. Frankly, at the rate you''re fixing bugs, maybe you can simply bless a revision as 1.1.1 in a matter of days. Why bother? Because of perception, testing, and easier collaboration. It won''t matter to hobbiests but to companies running ecommerce sites, bugfix-only stable releases will give them a sense of comfort--which they need in order to migrate away from Java and PHP. Mark my words...you will begin to see a lot of articles about Rails being ''too risky'' or ''too buggy'' in articles sponsored by vendors/developers of competing tools. They will point to events like site5 rolling back their servers to rails 1.0 after encountering issues with 1.1. Have them find out what went wrong and see if the problems can be fixed and site5 migrated to 1.1 before the FUD articles appear. The beauty of rails is that we can actually fix those pesky ''major new release issues'' before the sponsored FUD articles show up in IT magazines read by decision-makers. You know...the same folks that get withdrawal symptoms if they don''t see the word "enterprise" in every product description--sadly, the same ones that dictate what developers are allowed to use.> Regarding the problem of freezing, we''ll be making that mandatory on > "rails myapp" now. It''ll automatically extract all the gems into > vendor/rails. >Fantastic!> All applications should be distributed with the version of Rails > they''re certified to work with in vendor/rails. I hope we''ll see a new > typo release shortly that does this.Keep up the great work! And don''t let the upcoming avalanche of sponsored FUD about rails get you down. -- Posted via http://www.ruby-forum.com/.
Hi, I have upgraded to rails 1.1. Thus, I''m wondering, if I should have done so in light of all the errors reports that I have been posted here? I''m new to the Rails development and I simply would like to get through all the tutorials and move to developing for projects. Thus, if anyone can provide with a clear picture as to the state of Rails 1.1, it would be GREATLY appreciated. Thanks in advance, -Conrad On 3/30/06, John Smith <nospam1@nospam1.lan> wrote:> > David Heinemeier Hansson wrote: > >> I hope a small bugfix release addressing the top ~10 most annoying bugs > >> (like regressions) come out soon. It would be great to see rails 1.1.x > >> as the default on hosted servers so new users can benefit from all the > >> great new features before they learn to freeze edge. > > > > Do you have such a top 10? The only major issues I''ve heard of is > > Prototype performance in IE (fixed), Oracle adapter issue (fixed), and > > a failure to communicate that the environment.rb style from 0.13.x > > would need updating. > > > > The list should be decided by you and core members to maintain > objectivity. For example, the Oracle adapter issue is very serious to > those that use Oracle--but not for me. > > Tickets about regressions (things that worked in previous revisions) > should probably be included unless the changes are too time-consuming or > risky for a bugfix release. For example #3704 and #4234 are probably > worth investigating. > > After regressions, tickets like #4378 are probably worth considering > because it can bite a lot of people without warning. > > Frankly, at the rate you''re fixing bugs, maybe you can simply bless a > revision as 1.1.1 in a matter of days. Why bother? Because of > perception, testing, and easier collaboration. It won''t matter to > hobbiests but to companies running ecommerce sites, bugfix-only stable > releases will give them a sense of comfort--which they need in order to > migrate away from Java and PHP. > > Mark my words...you will begin to see a lot of articles about Rails > being ''too risky'' or ''too buggy'' in articles sponsored by > vendors/developers of competing tools. They will point to events like > site5 rolling back their servers to rails 1.0 after encountering issues > with 1.1. Have them find out what went wrong and see if the problems > can be fixed and site5 migrated to 1.1 before the FUD articles appear. > > The beauty of rails is that we can actually fix those pesky ''major new > release issues'' before the sponsored FUD articles show up in IT > magazines read by decision-makers. You know...the same folks that get > withdrawal symptoms if they don''t see the word "enterprise" in every > product description--sadly, the same ones that dictate what developers > are allowed to use. > > > Regarding the problem of freezing, we''ll be making that mandatory on > > "rails myapp" now. It''ll automatically extract all the gems into > > vendor/rails. > > > > Fantastic! > > > All applications should be distributed with the version of Rails > > they''re certified to work with in vendor/rails. I hope we''ll see a new > > typo release shortly that does this. > > Keep up the great work! And don''t let the upcoming avalanche of > sponsored FUD about rails get you down. > > -- > Posted via http://www.ruby-forum.com/. > _______________________________________________ > Rails mailing list > Rails@lists.rubyonrails.org > http://lists.rubyonrails.org/mailman/listinfo/rails >-------------- next part -------------- An HTML attachment was scrubbed... URL: http://wrath.rubyonrails.org/pipermail/rails/attachments/20060331/6183b5ee/attachment-0001.html
John Smith wrote:> Frankly, at the rate you''re fixing bugs, maybe you can simply bless a > revision as 1.1.1 in a matter of days. Why bother? Because of > perception, testing, and easier collaboration. It won''t matter to > hobbiests but to companies running ecommerce sites, bugfix-only stable > releases will give them a sense of comfort--which they need in order to > migrate away from Java and PHP. > > Mark my words...you will begin to see a lot of articles about Rails > being ''too risky'' or ''too buggy'' in articles sponsored by > vendors/developers of competing tools. They will point to events like > site5 rolling back their servers to rails 1.0 after encountering issues > with 1.1. Have them find out what went wrong and see if the problems > can be fixed and site5 migrated to 1.1 before the FUD articles appear.Companies running high availability sites didn''t get bitten by the Rails 1.1 bugs. They already knew not to deploy a new piece of software on the first day it was released. If you have a site that you are running for yourself or your clients that needs to be up 24x7, you don''t deploy software without testing it. Not just your software, but any software. Even if there aren''t any bugs in the third-party software that you are deploying (ha!) the deployment might uncover some previously hidden bugs in your own code. Just read the release notes. There were over 500 changes in Rails 1.1. If you are doing test driven development, you aren''t going to deploy 500 changes without testing. As to the various ISPs that apparently just pulled Rails 1.1 down and deployed it to all their servers, without testing and without communicating the change to their customers in advance, I don''t know what to think. I can''t imagine what they were thinking. Were they trying to please their customers who were clamoring for Rails 1.1? -- Ray
John Smith wrote: [...]> Mark my words...you will begin to see a lot of articles about Rails > being ''too risky'' or ''too buggy'' in articles sponsored by > vendors/developers of competing tools. They will point to events like > site5 rolling back their servers to rails 1.0 after encountering issues > with 1.1. Have them find out what went wrong and see if the problems > can be fixed and site5 migrated to 1.1 before the FUD articles appear. > > The beauty of rails is that we can actually fix those pesky ''major new > release issues'' before the sponsored FUD articles show up in IT > magazines read by decision-makers. You know...the same folks that get > withdrawal symptoms if they don''t see the word "enterprise" in every > product description--sadly, the same ones that dictate what developers > are allowed to use.One thing that "enterprises" will expect is the ability to get bug fixes without adding new features, and to continue to get bug fixes for a release long after the next release is out. Is this part of the plan for Rails? regards Justin
Wow... This thread has turned into the exact opposite of what I intended. There have been a lot of posts with issues on the new 1.1, and I wanted to put a positive one out there so that the core members and other contributors realize that they make the lives a real world developers easier. Rails 1.1fixed a bug I couldn''t figure out, motivated me to get all of my systems running Ruby 1.84, and introduced many enhancements that in only a couple of days have shortened my code base. I keep seeing this word ''enterprise'' on this list and around the net. I hate that word. I equate it with having an app that requires a couple hundred IT folks to keep running. All businesses, large or small, depend on the up time and reliability of their apps. Ruby on Rails is a powerful web application development framework that requires fewer developers to maintain... PERIOD. If Sun released a new version of Java today, how long would it take to work out the bugs that it creates in current java apps? I can guarantee that most Java developers out their would not even attempt to move to the new version for months if at all. I have a friend whose full time job is purely to manage Java versions for the company so that their apps aren''t broken. When Rails is deployed properly and tested thoroughly, issues with new releases are minimal, especially when developers are allowed to follow the development of the framework on a daily basis. So to answer Conrad, Rails 1.1 is where it''s at*. Many of my Rails co-developers have been riding the edge version for months. 37 signals if I am not mistaken runs on the bleeding edge of the code base too. Now that there is a certified 1.1, start there and don''t go back. You''ll miss great features like polymorphic associations, RJS templates, an overall faster app, and much, much more. So again, Rails 1.1 ~ Kick Ass. Thanks David and crew ~ ~ Ben * Two turn tables and a microphone... On 3/31/06, Justin Forder <justin@justinforder.me.uk> wrote:> > John Smith wrote: > [...] > > Mark my words...you will begin to see a lot of articles about Rails > > being ''too risky'' or ''too buggy'' in articles sponsored by > > vendors/developers of competing tools. They will point to events like > > site5 rolling back their servers to rails 1.0 after encountering issues > > with 1.1. Have them find out what went wrong and see if the problems > > can be fixed and site5 migrated to 1.1 before the FUD articles appear. > > > > The beauty of rails is that we can actually fix those pesky ''major new > > release issues'' before the sponsored FUD articles show up in IT > > magazines read by decision-makers. You know...the same folks that get > > withdrawal symptoms if they don''t see the word "enterprise" in every > > product description--sadly, the same ones that dictate what developers > > are allowed to use. > > One thing that "enterprises" will expect is the ability to get bug fixes > without adding new features, and to continue to get bug fixes for a > release long after the next release is out. > > Is this part of the plan for Rails? > > regards > > Justin > > _______________________________________________ > Rails mailing list > Rails@lists.rubyonrails.org > http://lists.rubyonrails.org/mailman/listinfo/rails >-- Ben Reubenstein http://www.benr75.com -------------- next part -------------- An HTML attachment was scrubbed... URL: http://wrath.rubyonrails.org/pipermail/rails/attachments/20060331/e50ac086/attachment-0001.html
Hey Ben, thanks for the information and I really appreciate it. Thus, I''ll get going with Rails 1.1. Thanks again, -Conrad On 3/31/06, Ben Reubenstein <benr@x-cr.com> wrote:> > Wow... This thread has turned into the exact opposite of what I intended. > There have been a lot of posts with issues on the new 1.1, and I wanted to > put a positive one out there so that the core members and other contributors > realize that they make the lives a real world developers easier. Rails > 1.1 fixed a bug I couldn''t figure out, motivated me to get all of my > systems running Ruby 1.84, and introduced many enhancements that in only > a couple of days have shortened my code base. > > I keep seeing this word ''enterprise'' on this list and around the net. I > hate that word. I equate it with having an app that requires a couple > hundred IT folks to keep running. All businesses, large or small, depend on > the up time and reliability of their apps. Ruby on Rails is a powerful > web application development framework that requires fewer developers to > maintain... PERIOD. If Sun released a new version of Java today, how long > would it take to work out the bugs that it creates in current java apps? I > can guarantee that most Java developers out their would not even attempt to > move to the new version for months if at all. I have a friend whose full > time job is purely to manage Java versions for the company so that their > apps aren''t broken. > > When Rails is deployed properly and tested thoroughly, issues with new > releases are minimal, especially when developers are allowed to follow the > development of the framework on a daily basis. > > So to answer Conrad, Rails 1.1 is where it''s at*. Many of my Railsco-developers have been riding the edge version for months. 37 signals if I > am not mistaken runs on the bleeding edge of the code base too. Now that > there is a certified 1.1, start there and don''t go back. You''ll miss > great features like polymorphic associations, RJS templates, an overall > faster app, and much, much more. > > So again, Rails 1.1 ~ Kick Ass. Thanks David and crew ~ > > ~ Ben > > * Two turn tables and a microphone... > > > On 3/31/06, Justin Forder < justin@justinforder.me.uk> wrote: > > > > John Smith wrote: > > [...] > > > Mark my words...you will begin to see a lot of articles about Rails > > > being ''too risky'' or ''too buggy'' in articles sponsored by > > > vendors/developers of competing tools. They will point to events like > > > > > site5 rolling back their servers to rails 1.0 after encountering > > issues > > > with 1.1. Have them find out what went wrong and see if the problems > > > can be fixed and site5 migrated to 1.1 before the FUD articles appear. > > > > > > > > The beauty of rails is that we can actually fix those pesky ''major new > > > release issues'' before the sponsored FUD articles show up in IT > > > magazines read by decision-makers. You know...the same folks that get > > > > > withdrawal symptoms if they don''t see the word "enterprise" in every > > > product description--sadly, the same ones that dictate what developers > > > are allowed to use. > > > > One thing that "enterprises" will expect is the ability to get bug fixes > > > > without adding new features, and to continue to get bug fixes for a > > release long after the next release is out. > > > > Is this part of the plan for Rails? > > > > regards > > > > Justin > > > > _______________________________________________ > > Rails mailing list > > Rails@lists.rubyonrails.org > > http://lists.rubyonrails.org/mailman/listinfo/rails > > > > > > -- > > Ben Reubenstein > http://www.benr75.com > > _______________________________________________ > Rails mailing list > Rails@lists.rubyonrails.org > http://lists.rubyonrails.org/mailman/listinfo/rails > > >-------------- next part -------------- An HTML attachment was scrubbed... URL: http://wrath.rubyonrails.org/pipermail/rails/attachments/20060331/ffed20a7/attachment.html
> Tickets about regressions (things that worked in previous revisions) > should probably be included unless the changes are too time-consuming or > risky for a bugfix release. For example #3704 and #4234 are probably > worth investigating.+1> Frankly, at the rate you''re fixing bugs, maybe you can simply bless a > revision as 1.1.1 in a matter of days.Call me crazy, but it would be incredibly cool to just take the release schedule to the level which is normal for Web 2.0 apps like Flickr, which builds and releases every couple hours. Do incredibly-minor-version releases every single day. It would take some work to get a sufficiently kick-ass testing/integration framework in place, but that would be work worth doing. It would also permanently clobber the "enterprise" wankers. :-)> Mark my words...you will begin to see a lot of articles about Rails > being ''too risky'' or ''too buggy'' in articles sponsored by > vendors/developers of competing tools.True, but that was happening already anyway. -- Giles Bowkett www.gilesgoatboy.org
> As to the various ISPs that apparently just pulled Rails 1.1 down and > deployed it to all their servers, without testing and without > communicating the change to their customers in advance, I don''t know > what to think. I can''t imagine what they were thinking. Were they trying > to please their customers who were clamoring for Rails 1.1?I think they just got carried away in all the excitement. -- Giles Bowkett www.gilesgoatboy.org
Ben Reubenstein wrote:> Wow... This thread has turned into the exact opposite of what I > intended. > There have been a lot of posts with issues on the new 1.1, and I wanted > to > put a positive one out there so that the core members and other > contributors > realize that they make the lives a real world developers easier. > Rails 1.1fixed a bug I couldn''t figure out, motivated me to get all of > my systems > running Ruby 1.84, and introduced many enhancements that in only a > couple > of days have shortened my code base. > > I keep seeing this word ''enterprise'' on this list and around the net. I > hate that word. I equate it with having an app that requires a couple > hundred IT folks to keep running. All businesses, large or small, > depend on > the up time and reliability of their apps. Ruby on Rails is a powerful > web > application development framework that requires fewer developers to > maintain... PERIOD. If Sun released a new version of Java today, how > long > would it take to work out the bugs that it creates in current java apps? > I > can guarantee that most Java developers out their would not even attempt > to > move to the new version for months if at all. I have a friend whose > full > time job is purely to manage Java versions for the company so that their > apps aren''t broken. > > When Rails is deployed properly and tested thoroughly, issues with new > releases are minimal, especially when developers are allowed to follow > the > development of the framework on a daily basis. > > So to answer Conrad, Rails 1.1 is where it''s at*. Many of my Rails > co-developers have been riding the edge version for months. 37 signals > if I > am not mistaken runs on the bleeding edge of the code base too. Now > that > there is a certified 1.1, start there and don''t go back. You''ll miss > great > features like polymorphic associations, RJS templates, an overall faster > app, and much, much more. > > So again, Rails 1.1 ~ Kick Ass. Thanks David and crew ~ > > ~ Ben > > * Two turn tables and a microphone...I''m glad you were here to respond to Justin''s questions. It shows everyone that without paying any overpriced ''enterprise'' fees, we respond quickly with relevant answers that are very practical. I only have one tiny correction to make in your otherwise fantastic response. I think comparing Rails to Java is comparing a framework to a language. Maybe a comparison to Struts or Zope at the next opportunity... Anyway, here''s Justin''s original question in case we want to put this question into the Rails FAQ along with with your answer. I think it''ll prevent his question from being asked again because the answer is so precise and clear. Justin>> One thing that "enterprises" will expect is the ability to get bug Justin>> fixes without adding new features, and to continue to get bug fixes Justin>> for a release long after the next release is out. Justin>> Justin>> Is this part of the plan for Rails? I''d put this into the wiki''s FAQ page but I''m pressed for time and need to run. We need more people like you to help answer questions. Too bad forums for other software is full of fanboys with religious ferver. I''m sure you''ve made rails much more appealing to many others by your response. See? Conrad already jumped on board because of you! Rock on! -- Posted via http://www.ruby-forum.com/.
Ben Reubenstein wrote:> Wow... This thread has turned into the exact opposite of what I > intended. There have been a lot of posts with issues on the new 1.1, > and I wanted to put a positive one out there so that the core members > and other contributors realize that they make the lives a real world > developers easier. Rails 1.1 fixed a bug I couldn''t figure out, > motivated me to get all of my systems running Ruby 1.84, and introduced > many enhancements that in only a couple of days have shortened my code > base.Hi, Ben. I was glad to see your post, and you are not responsible for the twists and turns that the thread took after you started it.> I keep seeing this word ''enterprise'' on this list and around the net. I > hate that word. I equate it with having an app that requires a couple > hundred IT folks to keep running. All businesses, large or small, > depend on the up time and reliability of their apps. Ruby on Rails is a > powerful web application development framework that requires fewer > developers to maintain... PERIOD. If Sun released a new version of Java > today, how long would it take to work out the bugs that it creates in > current java apps? I can guarantee that most Java developers out their > would not even attempt to move to the new version for months if at > all. I have a friend whose full time job is purely to manage Java > versions for the company so that their apps aren''t broken.Yes - I haven''t yet had the chance to work with Java 5.0, which came out in mid-2004. Both the clients I''m working for (a leading UK insurance company, and a foreign exchange settlement bank) are IBM-based, so even green-field projects use WebSphere App Server v6, which runs on Java 1.4.2. (Even so, we have used Rails successfully for prototyping in one project.) At the application level, Java upgrades don''t break much and do bring advantages. We just can''t get those advantages because of the weight of application server between us and the language.> When Rails is deployed properly and tested thoroughly, issues with new > releases are minimal, especially when developers are allowed to follow > the development of the framework on a daily basis. > > So to answer Conrad, Rails 1.1 is where it''s at*. Many of my Rails > co-developers have been riding the edge version for months. 37 signals > if I am not mistaken runs on the bleeding edge of the code base too. > Now that there is a certified 1.1, start there and don''t go back. > You''ll miss great features like polymorphic associations, RJS templates, > an overall faster app, and much, much more.37 signals knows exactly what is going on at the edge. I''m sure they don''t automatically update their servers after every commit! As the audience for Rails grows, it becomes too much to expect every developer to make these judgements (and the more Rails is positioned as being easy to use, the truer this will be). The fact that 37 signals runs on the edge is a mixed blessing for other users. On the plus side, it means that released code has been proved in production. On the down side, it means that 37 signals are not feeling the pain of working with the current release. Until 1.0, every Rails release I saw was followed quickly be a 0.n.m.1 release which fixed some teething troubles. 1.0 had at least one serious bug (relating to the saving of multiple associations) - but there was never a 1.0.1 release. Are users expected to find and apply patches? Should everyone switch to Edge Rails to get fixes? (If so, the concept of a release loses all relevance.) To echo my earlier post, I *do* think it is important for Rails to have maintenance releases, and where an upgrade is a major change it will be important for the earlier release to continue to be maintained for a while.> > So again, Rails 1.1 ~ Kick Ass. Thanks David and crew ~Absolutely! Justin
Ray Baxter wrote:> As to the various ISPs that apparently just pulled Rails 1.1 down and > deployed it to all their servers, without testing and without > communicating the change to their customers in advance, I don''t know > what to think. I can''t imagine what they were thinking. Were they trying > to please their customers who were clamoring for Rails 1.1?I think that if you''re running an application on shared hosting, you need to ensure that you have require_gem statements locking you to the version you need. Now, I''d love to see Rails create those for you-- inserting the chunk of environment.rb shown in the Wiki (http://wiki.rubyonrails.com/rails/pages/HowtoLockToSpecificRailsVersions) would make everyone''s lives a bit easier, I think. -- Posted via http://www.ruby-forum.com/.