<?xml version="1.0" encoding="UTF-8"?><rss version="2.0"
	xmlns:content="http://purl.org/rss/1.0/modules/content/"
	xmlns:dc="http://purl.org/dc/elements/1.1/"
	xmlns:atom="http://www.w3.org/2005/Atom"
	xmlns:sy="http://purl.org/rss/1.0/modules/syndication/"
		>
<channel>
	<title>Comments on: Rails 2.3.5 still broken on 1.9.1</title>
	<atom:link href="http://fukamachi.org/wp/2009/11/29/rails-2-3-5-still-broken-on-1-9-1/feed/" rel="self" type="application/rss+xml" />
	<link>http://fukamachi.org/wp/2009/11/29/rails-2-3-5-still-broken-on-1-9-1/</link>
	<description>「偶然世界」で出逢い</description>
	<lastBuildDate>Fri, 12 Mar 2010 07:49:44 +0000</lastBuildDate>
	<generator>http://wordpress.org/?v=2.9.1</generator>
	<sy:updatePeriod>hourly</sy:updatePeriod>
	<sy:updateFrequency>1</sy:updateFrequency>
		<item>
		<title>By: Walter Schreppers</title>
		<link>http://fukamachi.org/wp/2009/11/29/rails-2-3-5-still-broken-on-1-9-1/comment-page-1/#comment-45396</link>
		<dc:creator>Walter Schreppers</dc:creator>
		<pubDate>Tue, 26 Jan 2010 04:25:08 +0000</pubDate>
		<guid isPermaLink="false">http://fukamachi.org/wp/?p=1978#comment-45396</guid>
		<description>I&#039;m using rails 2.3.3 happily with ruby 1.9.1.
With rails 2.3.4 there are the weird session errors ( ^...bla bla in application.html.erb)
it had all to do with a timing attack they wanted to prevent (basically if your not running a site for the FBI, I believe it&#039;s not important. Anyway I&#039;d use ssl on that kind of thing which puts the security burdon on the apache daemon anyways. I mean ok it is good to secure the sessions but 2.3.4 just plain broke due to this timing attack patch!

When 2.3.5 came out with the patch that worked for 1.9.1 I tried using that but unfortunately every app became ultra slow. So some good advice is to stick to 2.3.3 with 1.9.1 you get speedy apps and speedy database access!

To install it:
gem install rails -v 2.3.3

Anyway that&#039;s what I&#039;m sticking with until 2.3.6 with 1.9.x works equally fast/stable.

As for stability I can vouch that 1.9.1 with 2.3.3 is indeed stable (was running for months with no visible memory leaks or performance issues on my debian servers).

I hear constantly people saying to stick with 1.8.x but that&#039;s like sticking with g++ 2.x which the linux kernel tried to do for a long time. In the end it&#039;s better to move on (I mean the author of ruby has, so why not the rails developers?).  It is inevitable that rails should in the end run on 1.9.x since that is where ruby is going ...</description>
		<content:encoded><![CDATA[<p>I&#8217;m using rails 2.3.3 happily with ruby 1.9.1.<br />
With rails 2.3.4 there are the weird session errors ( ^&#8230;bla bla in application.html.erb)<br />
it had all to do with a timing attack they wanted to prevent (basically if your not running a site for the FBI, I believe it&#8217;s not important. Anyway I&#8217;d use ssl on that kind of thing which puts the security burdon on the apache daemon anyways. I mean ok it is good to secure the sessions but 2.3.4 just plain broke due to this timing attack patch!</p>
<p>When 2.3.5 came out with the patch that worked for 1.9.1 I tried using that but unfortunately every app became ultra slow. So some good advice is to stick to 2.3.3 with 1.9.1 you get speedy apps and speedy database access!</p>
<p>To install it:<br />
gem install rails -v 2.3.3</p>
<p>Anyway that&#8217;s what I&#8217;m sticking with until 2.3.6 with 1.9.x works equally fast/stable.</p>
<p>As for stability I can vouch that 1.9.1 with 2.3.3 is indeed stable (was running for months with no visible memory leaks or performance issues on my debian servers).</p>
<p>I hear constantly people saying to stick with 1.8.x but that&#8217;s like sticking with g++ 2.x which the linux kernel tried to do for a long time. In the end it&#8217;s better to move on (I mean the author of ruby has, so why not the rails developers?).  It is inevitable that rails should in the end run on 1.9.x since that is where ruby is going &#8230;</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: Sho</title>
		<link>http://fukamachi.org/wp/2009/11/29/rails-2-3-5-still-broken-on-1-9-1/comment-page-1/#comment-43276</link>
		<dc:creator>Sho</dc:creator>
		<pubDate>Mon, 30 Nov 2009 07:33:42 +0000</pubDate>
		<guid isPermaLink="false">http://fukamachi.org/wp/?p=1978#comment-43276</guid>
		<description>Ah well. I&#039;m sure it will be fixed eventually. For now, the &quot;place these two monkeypatches in your init folder&quot; hack is painless enough.</description>
		<content:encoded><![CDATA[<p>Ah well. I&#8217;m sure it will be fixed eventually. For now, the &#8220;place these two monkeypatches in your init folder&#8221; hack is painless enough.</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: Wincent Colaiuta</title>
		<link>http://fukamachi.org/wp/2009/11/29/rails-2-3-5-still-broken-on-1-9-1/comment-page-1/#comment-43274</link>
		<dc:creator>Wincent Colaiuta</dc:creator>
		<pubDate>Mon, 30 Nov 2009 07:27:55 +0000</pubDate>
		<guid isPermaLink="false">http://fukamachi.org/wp/?p=1978#comment-43274</guid>
		<description>I think it&#039;s because they speak English and they simply don&#039;t care. You can rest assured that if it did affect them or their userbases then something would be done about it.</description>
		<content:encoded><![CDATA[<p>I think it&#8217;s because they speak English and they simply don&#8217;t care. You can rest assured that if it did affect them or their userbases then something would be done about it.</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: Sho</title>
		<link>http://fukamachi.org/wp/2009/11/29/rails-2-3-5-still-broken-on-1-9-1/comment-page-1/#comment-43269</link>
		<dc:creator>Sho</dc:creator>
		<pubDate>Mon, 30 Nov 2009 05:22:29 +0000</pubDate>
		<guid isPermaLink="false">http://fukamachi.org/wp/?p=1978#comment-43269</guid>
		<description>See, this pisses me off:

&lt;blockquote&gt;Improved compatibility with Ruby 1.9

There were a few small bugs preventing full compatibility with Ruby 1.9. However, we wouldn’t be surprised you were already running Rails 2.3.X successfully before these bugs were fixed (they were small).&lt;/blockquote&gt;

Were a few bugs, eh. Past tense. Guess I, and everyone else on that ticket, are just experiencing a group hallucination!</description>
		<content:encoded><![CDATA[<p>See, this pisses me off:</p>
<blockquote><p>Improved compatibility with Ruby 1.9</p>
<p>There were a few small bugs preventing full compatibility with Ruby 1.9. However, we wouldn’t be surprised you were already running Rails 2.3.X successfully before these bugs were fixed (they were small).</p></blockquote>
<p>Were a few bugs, eh. Past tense. Guess I, and everyone else on that ticket, are just experiencing a group hallucination!</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: Sho</title>
		<link>http://fukamachi.org/wp/2009/11/29/rails-2-3-5-still-broken-on-1-9-1/comment-page-1/#comment-43265</link>
		<dc:creator>Sho</dc:creator>
		<pubDate>Mon, 30 Nov 2009 01:43:19 +0000</pubDate>
		<guid isPermaLink="false">http://fukamachi.org/wp/?p=1978#comment-43265</guid>
		<description>As I understand it, it will display pages correctly if there are no partials containing non-ascii characters, yes. And it&#039;s not like you have to be doing crazy foreign shit either - one of my apps dares to use a few unicode arrow symbols in various ways, with the same problem.

But I don&#039;t think submitting data to the DB will work, ie through a form, without the &quot;fix params&quot; patch referenced in the ticket, since Rails will try and fail to pass the incorrectly encoded UTF8 to the database. Well that was my experience anyway, didn&#039;t test it all that thoroughly.

2.3.5 does seems to have fixed one 1.9-related bug, to do with string comparisons, so at least most authentication will work now. I just can&#039;t understand why they didn&#039;t roll in the other fixes as well. In fact if I was not such a charitable person I would be tempted to think they are taking a perverse pleasure in &lt;em&gt;not&lt;/em&gt; rolling out the patches, which have been extensively used and verified by myself and others on the ticket for over *6 months*. Certainly precious few other reasons spring to mind.</description>
		<content:encoded><![CDATA[<p>As I understand it, it will display pages correctly if there are no partials containing non-ascii characters, yes. And it&#8217;s not like you have to be doing crazy foreign shit either &#8211; one of my apps dares to use a few unicode arrow symbols in various ways, with the same problem.</p>
<p>But I don&#8217;t think submitting data to the DB will work, ie through a form, without the &#8220;fix params&#8221; patch referenced in the ticket, since Rails will try and fail to pass the incorrectly encoded UTF8 to the database. Well that was my experience anyway, didn&#8217;t test it all that thoroughly.</p>
<p>2.3.5 does seems to have fixed one 1.9-related bug, to do with string comparisons, so at least most authentication will work now. I just can&#8217;t understand why they didn&#8217;t roll in the other fixes as well. In fact if I was not such a charitable person I would be tempted to think they are taking a perverse pleasure in <em>not</em> rolling out the patches, which have been extensively used and verified by myself and others on the ticket for over *6 months*. Certainly precious few other reasons spring to mind.</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: Wincent Colaiuta</title>
		<link>http://fukamachi.org/wp/2009/11/29/rails-2-3-5-still-broken-on-1-9-1/comment-page-1/#comment-43255</link>
		<dc:creator>Wincent Colaiuta</dc:creator>
		<pubDate>Sun, 29 Nov 2009 18:31:58 +0000</pubDate>
		<guid isPermaLink="false">http://fukamachi.org/wp/?p=1978#comment-43255</guid>
		<description>I was going to pen a bit of a rant myself about 2.3.5, actually, based on the fact that the reason they did the release was to get a security fix out, but if you do a &quot;git log v2.3.4..v2.3.5&quot; you&#039;ll see that alongside the security fix and some (welcome, I guess) bug fixes, there is a lot of other churn which isn&#039;t really justified in my opinion. But in the end couldn&#039;t muster the motivation to do so (at least me existing specs and Cucumber features continued to work).

I agree it would be nice if Rails just worked with 1.9. As I understand it, it basically does if you live in a monolingual world where only US English is spoken.</description>
		<content:encoded><![CDATA[<p>I was going to pen a bit of a rant myself about 2.3.5, actually, based on the fact that the reason they did the release was to get a security fix out, but if you do a &#8220;git log v2.3.4..v2.3.5&#8243; you&#8217;ll see that alongside the security fix and some (welcome, I guess) bug fixes, there is a lot of other churn which isn&#8217;t really justified in my opinion. But in the end couldn&#8217;t muster the motivation to do so (at least me existing specs and Cucumber features continued to work).</p>
<p>I agree it would be nice if Rails just worked with 1.9. As I understand it, it basically does if you live in a monolingual world where only US English is spoken.</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: Sho</title>
		<link>http://fukamachi.org/wp/2009/11/29/rails-2-3-5-still-broken-on-1-9-1/comment-page-1/#comment-43251</link>
		<dc:creator>Sho</dc:creator>
		<pubDate>Sun, 29 Nov 2009 17:45:13 +0000</pubDate>
		<guid isPermaLink="false">http://fukamachi.org/wp/?p=1978#comment-43251</guid>
		<description>Yeah, well me neither really. But a blog post and/or pointed comment on a Lighthouse ticket every now and again is well within my &quot;I care about this enough to spend the time writing it&quot; threshold. I&#039;ve spent much more time replying to your comments : )

It really will be good when we can trust 1.9 enough that even conservative actors such as yourself have no qualms about deploying on it - that will only happen when the mainstream sees it as a genuine option and has time to pound out the bugs. My goal is simply to bring that day forward a little bit. I like 1.9, I want to be able to standardise on it, it&#039;s as simple as that really.

Rails is a chaotic mess where the only thing that seems to work in getting things fixed is making sufficient noise and nuisance of one&#039;s self that the committers come to see it as a profitable use of their time to type the one-line command that merges the six-month-old well-tested patches that do the fixing. This post, along with all others by myself on the topic, are simply calculated to be that noise. I am also attempting to save time for hapless fellow developers wondering why their &quot;Ruby 1.9 compatible&quot; rails applications are not, in point of fact, compatible with Ruby 1.9. Call it civic duty.</description>
		<content:encoded><![CDATA[<p>Yeah, well me neither really. But a blog post and/or pointed comment on a Lighthouse ticket every now and again is well within my &#8220;I care about this enough to spend the time writing it&#8221; threshold. I&#8217;ve spent much more time replying to your comments : )</p>
<p>It really will be good when we can trust 1.9 enough that even conservative actors such as yourself have no qualms about deploying on it &#8211; that will only happen when the mainstream sees it as a genuine option and has time to pound out the bugs. My goal is simply to bring that day forward a little bit. I like 1.9, I want to be able to standardise on it, it&#8217;s as simple as that really.</p>
<p>Rails is a chaotic mess where the only thing that seems to work in getting things fixed is making sufficient noise and nuisance of one&#8217;s self that the committers come to see it as a profitable use of their time to type the one-line command that merges the six-month-old well-tested patches that do the fixing. This post, along with all others by myself on the topic, are simply calculated to be that noise. I am also attempting to save time for hapless fellow developers wondering why their &#8220;Ruby 1.9 compatible&#8221; rails applications are not, in point of fact, compatible with Ruby 1.9. Call it civic duty.</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: Wincent Colaiuta</title>
		<link>http://fukamachi.org/wp/2009/11/29/rails-2-3-5-still-broken-on-1-9-1/comment-page-1/#comment-43250</link>
		<dc:creator>Wincent Colaiuta</dc:creator>
		<pubDate>Sun, 29 Nov 2009 17:30:40 +0000</pubDate>
		<guid isPermaLink="false">http://fukamachi.org/wp/?p=1978#comment-43250</guid>
		<description>I actually think the reason people are so cache-crazy in the Rails ecosystem is that there are lot of pathologically coded apps out there which perform dozens (and dozens) of database queries per request.

In any case, agree with you that it&#039;s a good thing that somebody is pressuring for 1.9 support; I&#039;m just saying that I&#039;m not gonna be that person...</description>
		<content:encoded><![CDATA[<p>I actually think the reason people are so cache-crazy in the Rails ecosystem is that there are lot of pathologically coded apps out there which perform dozens (and dozens) of database queries per request.</p>
<p>In any case, agree with you that it&#8217;s a good thing that somebody is pressuring for 1.9 support; I&#8217;m just saying that I&#8217;m not gonna be that person&#8230;</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: Sho</title>
		<link>http://fukamachi.org/wp/2009/11/29/rails-2-3-5-still-broken-on-1-9-1/comment-page-1/#comment-43247</link>
		<dc:creator>Sho</dc:creator>
		<pubDate>Sun, 29 Nov 2009 16:22:57 +0000</pubDate>
		<guid isPermaLink="false">http://fukamachi.org/wp/?p=1978#comment-43247</guid>
		<description>I know where you&#039;re coming from and I agree to some extent, but I still consider it worth pursuing and writing about. Ruby 1.9 is stable and completely usable; it&#039;s simply a pity that the most common framework doesn&#039;t work on it. It looks bad, too, for the language, and considering the amount of time I&#039;ve invested in Ruby in general and Rails in particular, that&#039;s something I am significantly invested in. 

Ruby 1.9 is faster and uses less memory than Ruby 1.8 - these are both significant steps towards freeing Ruby from its (unfortunately well-deserved) reputation of being slow as cold molasses and an utter pig for memory. It&#039;s worth pushing those holding the commit rights to take 1.9 compatibility a little more seriously IMO.

You might not have an issue with speed - probably your application is cached to the hilt. Remember that such caching strategies are a significant amount of work to implement and are not necessary at all for many other languages - PHP generally runs fast enough that you don&#039;t even need to start thinking about caching until you&#039;re seeing a &lt;em&gt;lot&lt;/em&gt; of hits. Even python, usually the language Ruby is most directly contrasted against, kicks its ass in terms of speed. The fact that caching is such an important subject in the world of Rails is good in some ways but also speaks volumes about the underlying speed problems. And that&#039;s before we even start talking about memory usage. 

Ruby 1.9 is not the be all and end all, and caching is still important. But it&#039;s a step forward. I would like to see Rails take that step.

I&#039;m not using 1.9 in production for web servers yet but I&#039;m using it for image processing daemons, etc. The work is all done; I have had no problems at all using most of the common libraries, and I&#039;ve seen significant speedups and efficiency gains, usually around the order of 50%. That might not sound like much but if you&#039;re processing 50,000 images then it&#039;s the difference between 3 days processing time and 2 - and memory usage is greatly reduced, obsoleting hacks like REE. I have never had any stability issue whatsoever. The only problem is Rails, so I&#039;ll keep bugging the core team until they commit those damn patches and we can all forget about it and move on!</description>
		<content:encoded><![CDATA[<p>I know where you&#8217;re coming from and I agree to some extent, but I still consider it worth pursuing and writing about. Ruby 1.9 is stable and completely usable; it&#8217;s simply a pity that the most common framework doesn&#8217;t work on it. It looks bad, too, for the language, and considering the amount of time I&#8217;ve invested in Ruby in general and Rails in particular, that&#8217;s something I am significantly invested in. </p>
<p>Ruby 1.9 is faster and uses less memory than Ruby 1.8 &#8211; these are both significant steps towards freeing Ruby from its (unfortunately well-deserved) reputation of being slow as cold molasses and an utter pig for memory. It&#8217;s worth pushing those holding the commit rights to take 1.9 compatibility a little more seriously IMO.</p>
<p>You might not have an issue with speed &#8211; probably your application is cached to the hilt. Remember that such caching strategies are a significant amount of work to implement and are not necessary at all for many other languages &#8211; PHP generally runs fast enough that you don&#8217;t even need to start thinking about caching until you&#8217;re seeing a <em>lot</em> of hits. Even python, usually the language Ruby is most directly contrasted against, kicks its ass in terms of speed. The fact that caching is such an important subject in the world of Rails is good in some ways but also speaks volumes about the underlying speed problems. And that&#8217;s before we even start talking about memory usage. </p>
<p>Ruby 1.9 is not the be all and end all, and caching is still important. But it&#8217;s a step forward. I would like to see Rails take that step.</p>
<p>I&#8217;m not using 1.9 in production for web servers yet but I&#8217;m using it for image processing daemons, etc. The work is all done; I have had no problems at all using most of the common libraries, and I&#8217;ve seen significant speedups and efficiency gains, usually around the order of 50%. That might not sound like much but if you&#8217;re processing 50,000 images then it&#8217;s the difference between 3 days processing time and 2 &#8211; and memory usage is greatly reduced, obsoleting hacks like REE. I have never had any stability issue whatsoever. The only problem is Rails, so I&#8217;ll keep bugging the core team until they commit those damn patches and we can all forget about it and move on!</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: Wincent Colaiuta</title>
		<link>http://fukamachi.org/wp/2009/11/29/rails-2-3-5-still-broken-on-1-9-1/comment-page-1/#comment-43244</link>
		<dc:creator>Wincent Colaiuta</dc:creator>
		<pubDate>Sun, 29 Nov 2009 15:59:57 +0000</pubDate>
		<guid isPermaLink="false">http://fukamachi.org/wp/?p=1978#comment-43244</guid>
		<description>I&#039;m sticking with 1.8 for the forseeable future. Given that application performance is not a problem for me and application stability &lt;em&gt;is&lt;/em&gt; something that I care about, there&#039;s simply no reason to even consider moving to 1.9 at this stage.

Let the brave, the reckless, and those who insist on updating to the &quot;newest and shiniest&quot; bang their heads against this particular brick wall, while I get on with my life elsewhere.</description>
		<content:encoded><![CDATA[<p>I&#8217;m sticking with 1.8 for the forseeable future. Given that application performance is not a problem for me and application stability <em>is</em> something that I care about, there&#8217;s simply no reason to even consider moving to 1.9 at this stage.</p>
<p>Let the brave, the reckless, and those who insist on updating to the &#8220;newest and shiniest&#8221; bang their heads against this particular brick wall, while I get on with my life elsewhere.</p>
]]></content:encoded>
	</item>
</channel>
</rss>
