Robby on Rails comments

"Ruby eye for the anti-newbie guy" by Skivakantie Tsjechie

Sat, 11 Oct 2008 05:42:58 -0500

It’s too simple, which makes it nice though :-)

"Dialogue versus Debate" by boyer

Wed, 08 Oct 2008 23:02:13 -0500

timer software

"Dialogue versus Debate" by goodpack

Wed, 08 Oct 2008 23:01:44 -0500

泰拳搬屋搜尋引擎最佳化物理治療坐骨神經痛物理治療師飲水機新娘化妝服務廣告禮品物理治療| 坐骨神經痛物理治療師 Water Filter|

"The Daily Stand Up" by goodpack

Wed, 08 Oct 2008 23:01:32 -0500

泰拳搬屋搜尋引擎最佳化物理治療坐骨神經痛物理治療師飲水機新娘化妝服務廣告禮品物理治療| 坐骨神經痛物理治療師 Water Filter|

"The Daily Stand Up" by sadfasf

Wed, 08 Oct 2008 23:01:00 -0500

timer software

"The Daily Stand Up" by sadfasf

Wed, 08 Oct 2008 23:00:39 -0500

Interior Design 室內設計裝修飲水機 home Design office design decoration 飲水機| 泰拳迷你倉迷你倉泰拳搬屋文具物理治療| 迷你倉

"Review: Highrise, part 2" by Jeff

Wed, 08 Oct 2008 20:33:20 -0500

In regards to the contact form integration, just got my PHP form to work. You need to trick Highrise into thinking it’s getting a forward from your email address.

- I added a second set of variables to customize the message being sent to Highrise, and to change the ‘from’ address to be the email address corresponding to the dropbox address. - Have the form send a second email using these new variables to your dropbox address using “Fwd: ...” as the subject (this is the key), and with your own email as the from as stated above.

Highrise will receive the email thinking it was forwarded from you, and as long as you have a proper “From: Name email@email.com>” in the message body somewhere, it should add the person just fine.

I’m not entirely sure how to include details like “Company” and “Phone Number”, assuming these were part of the original form fields. Guess this is where the API comes in.

P.S. We use Highrise, Basecamp and Backpack religiously in our company.

"Subscribe to Basecamp RSS Feeds in Google Reader" by rick

Wed, 08 Oct 2008 10:25:31 -0500

no, but bloglines has had it for years.

"MySQL is just a toy" by Sebastian

Tue, 07 Oct 2008 16:18:45 -0500

I also appreciate postresql, but in fact mysql has many users.

"Review: Braintree " by nospam@matthewbmedia.com

Mon, 06 Oct 2008 22:02:14 -0500

If you are intending on doing your own integration work, then I can’t recommend using Paysimple. Their API is in flux, being worked on, and their documentation is outdated – their error reporting is not correct, in fact, it’s the worst error reporting I’ve seen from any gateway I’ve used so far.

It has been quite a pain to do a custom PHP integration.

Their developers are really nice, and helpful – but I would recommend coming back in 1 year if you are looking at paysimple.

However if you want to use their “hosted” order checkout page, then you won’t run into these issues.

"Installing Ruby on Rails and PostgreSQL on OS X, Third Edition" by Josh Kim

Sun, 05 Oct 2008 20:54:52 -0500

I guess I’m waiting for the 4th version of this. I think the official gem now is “ruby-pg” at http://rubyforge.org/projects/ruby-pg.

Everything is perfectly installed, except for the PostgreSQL gem. I cannot get my first rake db:migrate to work properly. It keeps telling me to install the gem “activerecord-postgresql-adapter”, which I know doesn’t exist, and it’s just a general activerecord error. I’ve tried not only the “ruby-pg” gem, but also (the now older) “postgres” as well as the “pg” (I don’t know what this one is).

Are you still using the “postgres” gem? Are things working out?

I was looking forward to rocking out in PostgreSQL land this weekend, but spending hours and hours scouring the Internet… and nothing. I might think it’s because I’m running 8.3.4, the latest version. Anyone else having my problem?

"Lesson Learned: Git Ref Naming" by pedro mg

Sat, 04 Oct 2008 15:41:54 -0500

git version 1.6.0.2.GIT

same behaviour here.

"Lesson Learned: Git Ref Naming" by pedro mg

Sat, 04 Oct 2008 10:03:02 -0500

Robby, wich versions are you using ?

"Installing Ruby on Rails and PostgreSQL on OS X, Third Edition" by nofxx

Sat, 04 Oct 2008 06:36:02 -0500

I`ve got some info on my brain:

http://wiki.nofxx.com/postgre

With postgis, how to start manually, the gem problem. Mac and Archlinux.

"Ruby 1.8.7 on MacPorts causing some problems" by gourmet food

Fri, 03 Oct 2008 11:31:26 -0500

I did not work in the end. But I know why: not enough coffee!

I copied all the line “file:///Users/Shared/dports and create that directory”

You should at a line break to your instructions my friend.

sudo port install ruby  rb-rubygems rb-termios rb-mysql swig
--->  Fetching ruby
--->  Attempting to fetch ruby-1.8.6-p114.tar.gz from http://www.ibiblio.org/pub/languages/ruby/1.8
[...]

thanks!

"Ruby 1.8.7 on MacPorts causing some problems" by gourmet food

Fri, 03 Oct 2008 11:06:59 -0500

not sure it works here… it gives me the following warning:

Warning: /opt/local/etc/macports/sources.conf specifies invalid source 'file:///Users/Shared/dports and create that directory', ignored.

just started it with the following command:

 sudo port install ruby@1.8.6-p114 rb-rubygems rb-termios rb-mysql swig

because without the @ it tried to install 1.8.7 anyway

"Q&A: ActiveRecord Observers and You" by grosser

Thu, 02 Oct 2008 06:55:11 -0500

My new observer did not work although tested, until i found out i forgot to add it to config…

Now i autoload my observers…

http://pragmatig.wordpress.com/2008/10/02/why-your-pretty-tested-observer-might-not-work/

"Review: Braintree " by Mark Percival

Wed, 01 Oct 2008 12:51:49 -0500

Yeah I got the same response from Briantree. $1mil plus in charge card revenue or else a minimum applies. I understand the need for minimums, but it’s a little steep if your just starting out and can’t guarantee that kind of volume. That being said I’d love to use the service, it gets great reviews, but at this point I’ll have to look elsewhere.

"Review: Braintree " by Mike Mangino

Wed, 01 Oct 2008 09:39:07 -0500

I just talked to the CEO of Braintree, Bryan Johnson. It sounds like they have some new plans that don’t require the minimum volume requirement. I’ve been using them for a while and am extremely happy.

Mike (Author of the Braintree gateway for ActiveMerchant)

"Review: Highrise, part 2" by Advocaat Amsterdam

Mon, 29 Sep 2008 18:21:16 -0500

Come to think about it, I will discuss it again as they improved much since the time i put it under their attention. Anybody has any experience implementing this in their corporate workflow?

"Review: Highrise, part 2" by Advocaat Amsterdam

Mon, 29 Sep 2008 18:16:55 -0500

I like almost all the work from 37 signals, i did discuss using highrise at work but atm we still continue using our old CRM which i think is just bloated with stuff that I don’t need. Thanks for sharing! - Advocaat

"Flash Message Conductor" by ActsAsFlinn

Thu, 18 Sep 2008 19:59:05 -0500

Nice stuff here. I’ve done something like this for a while, keying the flash message and classing the flash to show message context. Here’s a simple example:

http://snippy.actsasflinn.com/snippets/14

In the past I’ve even shown context icons along with the message.

"Starting MySQL after upgrading to OS X Leopard" by Darren

Tue, 16 Sep 2008 18:08:23 -0500

@Ash Do this:

$ cd /usr/local/mysql $ sudo chown -R mysql data/ $ sudo echo $ sudo ./bin/mysqld_safe &

That should fix chown problems and start mysql. Write /usr/local/mysql/: in your PATH variable in your .bash_login or profile if you don’t want to write /usr/local/mysql everytime.

"Ruby 1.8.7 on MacPorts causing some problems" by Chip Castle

Tue, 16 Sep 2008 15:07:36 -0500

This is exactly what I am looking for today, but I am running into a fetch problem on OS X 10.4.11:

chip-castles-powerbook-g4-17:/opt/local/var/macports/sources/rsync.macports.org/release/ports/lang/ruby deploy$ sudo port install rb-rubygems ruby -> Fetching ruby -> Attempting to fetch ruby-1.8.6-p114.tar.gz from http://www.ibiblio.org/pub/languages/ruby/1.8 -> Attempting to fetch ruby-1.8.6-p114.tar.gz from http://mirrors.sunsite.dk/ruby/1.8 -> Attempting to fetch ruby-1.8.6-p114.tar.gz from ftp://xyz.lcs.mit.edu/pub/ruby/1.8 -> Attempting to fetch ruby-1.8.6-p114.tar.gz from ftp://ftp.iij.ad.jp/pub/lang/ruby/1.8 -> Attempting to fetch ruby-1.8.6-p114.tar.gz from ftp://ftp.ruby-lang.org/pub/ruby/1.8 -> Attempting to fetch ruby-1.8.6-p114.tar.gz from ftp://ftp.fu-berlin.de/unix/languages/ruby/1.8 -> Attempting to fetch ruby-1.8.6-p114.tar.gz from ftp://ftp.easynet.be/ruby/ruby/1.8 -> Attempting to fetch ruby-1.8.6-p114.tar.gz from ftp://ftp.ntua.gr/pub/lang/ruby/1.8 -> Attempting to fetch ruby-1.8.6-p114.tar.gz from ftp://ftp.chg.ru/pub/lang/ruby/1.8 -> Attempting to fetch ruby-1.8.6-p114.tar.gz from ftp://ftp.kr.FreeBSD.org/pub/ruby/1.8 -> Attempting to fetch ruby-1.8.6-p114.tar.gz from ftp://ftp.iDaemons.org/pub/mirror/ftp.ruby-lang.org/ruby/1.8 -> Attempting to fetch ruby-1.8.6-p114.tar.gz from http://svn.macports.org/repository/macports/distfiles/ruby -> Attempting to fetch ruby-1.8.6-p114.tar.gz from http://svn.macports.org/repository/macports/distfiles/general/ -> Attempting to fetch ruby-1.8.6-p114.tar.gz from http://svn.macports.org/repository/macports/downloads/ruby Error: Target org.macports.fetch returned: fetch failed Error: The following dependencies failed to build: ruby Error: Status 1 encountered during processing.

Any suggestions for fixing this fetch issue?

Thanks!

"Request: Free OSX FTP Program" by Christian

Tue, 16 Sep 2008 13:23:42 -0500

If you develop in Firefox, then FireFTP (It’s an addon) might be a good option.

"the Argon Express" by Ameliyat

Mon, 15 Sep 2008 06:31:16 -0500

The final sentence in this article is: A video is coming soon too! Where can we find it? Regards

"the Argon Express" by Diyet Listesi

Mon, 15 Sep 2008 06:30:38 -0500

Very intesting post and pictures, I like the “Breakfast on the train”! Thanks for sharing.

"the Argon Express" by Prefabrik Ev

Mon, 15 Sep 2008 06:29:54 -0500

Very good article regards grom prefabrik

"git-svn is a gateway drug" by Jason Smith

Sat, 13 Sep 2008 13:38:48 -0500

@Robby

Great Post! Exactly what I am looking for. Thanks!

"MySQL is just a toy" by http://paulgresham.com

Sat, 13 Sep 2008 01:50:32 -0500

MySQL is a thin veneer above a set of decent technologies. Sleepycat (berkeley db) and Innodb were actually written by people that did know what they were doing. When MySQL tried to implement ACID, they simply couldn’t do it, they’re were a bunch of script kiddies that have risen off the backs of a talented few.

Anyhow Innodb and Sleepycat are now owned by Oracle. Whereas the MySQL ‘brand’ and it’s veneer (some installers and an SQL parser) were bought by Sun for US$1Bn. Smell something wrong here? I certainly no longer own any shares in Sun.

I still don’t understand, why use a system that is guaranteed to fail in such a way as to cause you hard to trace support issues? Job for life? Are you those types of people that just patch things up by hacking the database, think no more of it and go to fix the next error on some other system sitting on top of MySQL? If ‘good enough’ is the benchmark by which you develop software, go ahead and use MySQL, otherwise use a proper database.

"Parsing a RSS Feed" by naruto

Fri, 12 Sep 2008 21:01:05 -0500

how to parser RSS with Ajax on Rails?

"MySQL is just a toy" by planetmcd

Fri, 12 Sep 2008 08:46:15 -0500

I prefer Postgres by leaps and bounds, but if MySQL works for you, good for you.

I suppose the real issue here is whether MySQL is your choice because its popular or it fits your functionality level, and whether that matters.

If I need a lightweight database I tend to go towards SQLite. If I need robust features straight away, I gravitate towards Postgres. MySQL can handle either cases though.

I think the larger concern for MySQL users should be the dual license nature. What do you do when you want your product to go public? Do you need to buy a license. Currently not, when its a database backing a client server application but what if Sun changes their mind? Do you think your small start up, university, or non profit has the legal staff to challenge Sun’s? This is what scares me away from MySQL. With the availability of Postgres, I have a robust alternative that in many capacities meets or exceeds the functionality of MySQL, without the license issue.

"MySQL is just a toy" by Joe Grossberg

Fri, 12 Sep 2008 08:00:30 -0500

When will PostgreSQL evangelists learn that “good enough” is good enough for 99% of web apps (i.e. low- to medium-traffic, simple data, basic CRUD)?

People calling MySQL a “toy” is a counter-productive argument because it seems to imply that, if it works for you, your hard work must amount to a “toy” project.

"MySQL is just a toy" by Dick Davies

Fri, 12 Sep 2008 03:49:24 -0500

a million mysql users: I’ve never had a problem with it. postgresql users: Yes, that’s why you’re still a mysql user.

"Installing Ruby on Rails and PostgreSQL on OS X, Third Edition" by dan

Wed, 10 Sep 2008 12:23:52 -0500

quite very nice tutorial, but i still don’t want to use macports, is any problem if i wont use it?

"Working in Portland coffee shops and cafes reviews, part 1" by mykle

Tue, 09 Sep 2008 10:45:51 -0500

I agree with you—it would be totally sweet to maintain a dictionary of places for laptoppers to laptop in Portland, OR. (or elsewhere too, but Portland is where i live.)

"Get to Know a Gem: Rak" by zblmw

Sun, 07 Sep 2008 20:10:19 -0500

珠宝联盟网

中国珠宝联盟网(zblmw.com)是一家服务于中国大陆及全球华人社群的领先在线珠宝媒体及增值资讯服务提供商。中国珠宝网站拥有多家地区性网站，以服务大中华地区与海外华人以及珠宝企业为己任，通过为广大网民和政府企业用户提供网络媒体及娱乐、在线用户付费增值/无线增值服务和电子政务解决方案等在内的一系列服务。

专业珠宝门户——中国珠宝网站预计2008年在全球范围内注册用户超过500万，日浏览量能最高突破8000万次，将成为中国大陆及全球华人社群中最受推崇的行业互联网品牌。

高效的整合营销服务——凭借领先的技术和优质的服务，中国珠宝网站会深受广大网民的欢迎并能享有极高的声誉。

http://www.zblmw.com

"The new RubyURL API" by Korepetycje

Sun, 07 Sep 2008 09:01:20 -0500

Very nice!

"Google Chrome: discuss" by Paz

Sat, 06 Sep 2008 01:11:16 -0500

I downloaded it after reading the comic and I feel like I’ve been lied to. After about ten minutes of using it, one of the tabs went into a waiting mode (there was a video on the page) and I was unable to click on any of the other tabs. The entire browser was frozen.

Now I clearly remember the part of the comic that went into how “one tab won’t affect the others” and how one can freeze up and everything else is independent.

Yeah, well, it’s not true. The architecture is not the way they say it is, or that just couldn’t happen. Whatever. Maybe in a few years I’ll try it again and they’ll have figured it out. As a computer scientist, I would appreciate not being lied to about programming archiecture in the future though.

In other words: thanks but no thanks, Chrome.

"Google Chrome: discuss" by gschultens@sympatico.ca

Fri, 05 Sep 2008 19:15:04 -0500

I keep getting the “google installer is trying to access the Internet” message from my ZoneAlarm firewall. This concerns me. Quite likely it’s benign, but I can’t help get suspicious when an application keeps trying to access the Intenet on a regular basis.

"Parsing a RSS Feed" by Andrew

Fri, 05 Sep 2008 19:09:21 -0500

Can I get a link for that erection cream?

"Review: Braintree " by Chad Gardner

Fri, 05 Sep 2008 16:14:26 -0500

@ Ryan As Marlena mentioned previously in the comments, you should take a look at PaySimple—we provide a SaaS platform that caters to small businesses. Our all-in-one, customer-centric system is perfect for companies that aren’t pulling in $1.2 million but still want a level playing field and the same payment processing advantages of larger companies. If you have any questions, feel free to email me at cgardner (at) paysimple (dot) com

"Advanced Mathematics and Programming" by stephan

Fri, 05 Sep 2008 07:08:35 -0500

“Should an education be a prerequisite in pursuing a career in programming?” probably not, but it does help a lot.

both mathematics and programing are such broad subjects.

understand the boolean operation is definitly math. and programing without ‘if then’ is quite limited.

Anyone doing programing will learn math concepts one way or another. is it mandatory to learn it at school before doing programing ? i believe no and it’s one very peculiar strength of programing.

it’s really hard to do other scientific work without prior education. Like make a bridge or a car.

but our tools are made in the same stuff as what we make: some data in memory.

That’s why computer science have in incredible ability to learn by doing. So prior education is not mandatory, and passion can be the engine to learn on the field what other did at school.

The fact that programing is such broad spectrum (between {=$C$1*10} until PCI bus driver development) gives room to all levels to start and to progress. So programing is a really pervasive and democratic science giving room to everyone. Of course some masterpiece needs masters, that’s obvious, but there is room for everyone.

"Google Chrome: discuss" by xl

Thu, 04 Sep 2008 17:25:43 -0500

I like most of the features it provides, except one – The Google Installer (googleupdate.exe). It is configured to load on startup. I thought that was fine because I could disable it.

Well, I was wrong. It will load itself automatically and repeatedly even if you remove it from the startup or kill it in windows task manager, and tries to access the internet throughout the day!

"Ubiquity meets RubyURL" by kod pocztowy

Thu, 04 Sep 2008 07:34:37 -0500

@Robby – can you repair the link?

"Google Chrome: discuss" by movie buff

Wed, 03 Sep 2008 17:00:21 -0500

according to my very unscientific observations, it would seem that Chrome is a lot faster than FireFox and IE… though i do miss the “previously closed tabs” feature in Firefox

"Ubiquity meets RubyURL" by Patcito

Wed, 03 Sep 2008 12:47:58 -0500

The link to github doesn’t work.

"The new RubyURL API" by Matt Mower

Wed, 03 Sep 2008 03:33:37 -0500

Hi Robby. All clear now, thanks ;-)

"Google Chrome: discuss" by Bheeshmar

Tue, 02 Sep 2008 23:25:00 -0500

Wow: right-click context menu has “Inspect element” which brings up a very nice DOM/CSS/Javascript inspector. You can drill down to the component you are interested in and it will highlight on the web page.

So far, so good. I like being able to drag tabs out. Comes with Gears and Flash plug-ins.

Google is trying to protect their web-based business by not leaving the access to their domain in the hands of their rivals. They make their money from searches, but people search from browsers.

"Flash Message Conductor" by Allison Beckwith

Tue, 02 Sep 2008 13:38:04 -0500

Hi Nikolay,

We are using p tags because it is the appropriate element for the type of content it contains. A paragraph is a block level element for holding text, like the sentences of a flash message. The paragraphs are contained inside a div, which is a “general purpose” block level element used for dividing up content and creating structure. The general rule of thumb here at PA is that p tags are used for text elements, and divs are used for structural elements.

"Google Chrome: discuss" by Robby Russell

Tue, 02 Sep 2008 12:15:46 -0500

Sarat,

Given that it’ll be open source and allow for plugins, I don’t see why Adblock wouldn’t be able to be ported to it. Also, if they were to clutter the UI with ads… then people would just fork it and provide a non-ad based one. I really suspect that their aim is to help improve the browser-landscape… not stifle it with the failures of the past. (think.. Opera)

"Google Chrome: discuss" by Sarat Pediredla

Tue, 02 Sep 2008 11:09:15 -0500

Great, another vehicle for Google to run “contextual, on-demand” (stick a few other keywords here) ads directly in my browser.

Is it just me or does no one realise they are only releasing this browser so they can circumvent AdBlock in Firefox by directly serving ads on your browser?

"Google Chrome: discuss" by Jason Watkins

Tue, 02 Sep 2008 10:51:40 -0500

I think it’s good… the browser market needs some shaking up. If chrome grabs significant market share and gives priority to standards processes it’ll increase pressure on IE not to stagnate and hold back innovation on the web.

"Google Chrome: discuss" by Peter Fitzgibbons

Tue, 02 Sep 2008 08:54:50 -0500

Quite a nod to WebKit. Apple continues to induce change. Is Google the 800-lb Gorilla that the browser-battle has been awaiting ?

"Google Chrome: discuss" by Robby Russell

Tue, 02 Sep 2008 08:53:21 -0500

It seems that web browsers have become a major annoyance these days for people. If I had a nickel for every time Safari and/or Firefox hung on me after using it for a day or two of heavy usage. I don’t remember them being so unstable in the past… but we’re also pushing browsers a lot harder with the applications that we’re using today.

On a day-to-day basis, I’m using Safari for my web-browsing and use Firefox for development as I (echoing Isaac)... couldn’t image using anything besides Firebug. With the potential to keep those contexts separate between tabs… I’m quite eager to play with Google Chrome.

"Google Chrome: discuss" by Jan Wikholm

Tue, 02 Sep 2008 06:54:05 -0500

This is quite a nice surprise for me, since I am also one of those people who is developing web apps with a heavy-duty load of JavaScript (loves jQuery)

"Google Chrome: discuss" by Ed Spencer

Tue, 02 Sep 2008 06:14:03 -0500

This looks fraking incredible. Most of my development time has shifted from Ruby on the server to JavaScript on the browser (with the awe inspiring Ext JS library), so this development comes at just the right time!

I can’t wait

"The new RubyURL API" by kod pocztowy

Tue, 02 Sep 2008 03:01:50 -0500

I am impresed! :)

"Flash Message Conductor" by George

Tue, 02 Sep 2008 02:21:26 -0500

Thanks Robby, really useful yet simple.

"Google Chrome: discuss" by Isaac

Mon, 01 Sep 2008 23:24:32 -0500

I think one of the primary things I’m looking forward to in the tab-is-a-separate-process design is that loading a page with high latency on external objects, especially JavaScript, won’t block the entire browser. It will be really nice to have the option of killing just blocked tabs instead of the entire application. (I notice this mostly on pages with external advertisements that come with JS that hangs for whatever reason.) Same with plugins like Flash.

I don’t recall seeing it mentioned, but assuming that their browser processes are or map to OS processes, it will make the browser (Through the OS’s process scheduling) able to utilize multiple cores.

The only thing that might be missing as compared to Firefox would be the large base of extensions such as Firebug, Greasemonkey, etc. Can’t imagine web development life without Firebug.

"Google Chrome: discuss" by Robert

Mon, 01 Sep 2008 22:49:48 -0500

You got to be kidding me. After using google web DEcelerator and finding that you not only have no support for when it locks up your computer, you do not even follow the google web accelerator group.

I suppose as in web DEcelerator, all web pages go through your servers where you copy everybody’s information including bank accounts and passwords and then display out of date web pages.

I am so done with google. It was bad enough that you kept copies of my emails and allowed people to download files I has sent in my emails, but this is just about as much as I can take. I wouldn’t use google to collect my spam..

"Flash Message Conductor" by Brett

Mon, 01 Sep 2008 21:49:43 -0500

Nikolay Kolev,

I’m not sure that a list element would be the best to render flash messages. I suppose it depends on how you’re showing them, but these are messages and I’d argue that a paragraph tag is very appropriate given the content. Unless you’re confusing this with the output of error_messages_for, which this plugin doesn’t address and I’m sure many of us would be interested in hearing how Planet Argon handles those in their views.

"Flash Message Conductor" by Brett

Mon, 01 Sep 2008 21:47:22 -0500

Robby,

I like the simplicity here and honestly, this is much better than what our team is currently using. I’d really like to see more simple plugins get released.

"The new RubyURL API" by Robby Russell

Mon, 01 Sep 2008 11:19:03 -0500

@Matt,

I’m not trying to provide a redirect with this, just aiming to provide people with a short url to use wherever. RubyURL itself does the redirect… the API just provides a non-web browser interface to generating RubyURLs.

"The new RubyURL API" by Matt Mower

Mon, 01 Sep 2008 09:44:09 -0500

Maybe I missed something but, since you’re passing it in with the request, why do you need to return the website URL?

If you don’t then couldn’t you ditch XML and JSON and return an appropriate HTTP status code & Location: header?

"Flash Message Conductor" by grosser

Mon, 01 Sep 2008 06:21:11 -0500

nice plugin, simple but useful, will definetly go into my next project

"Flash Message Conductor" by Nikolay Kolev

Sun, 31 Aug 2008 19:51:02 -0500

Why are you using P tags instead of DIV or UL with LI ones?

"The new RubyURL API" by Ray

Sun, 31 Aug 2008 15:24:12 -0500

Yes! JSON support would be fantastic!

"The new RubyURL API" by Robby Russell

Sun, 31 Aug 2008 14:39:14 -0500

James,

I’m hoping to get some time to update the API and provide JSON support soon. Volunteers are welcome to further that cause. ;-)

"The new RubyURL API" by James Cox

Sun, 31 Aug 2008 14:05:51 -0500

xml? really?

"Git: Push it! (real good)" by Ken Liu

Sat, 30 Aug 2008 23:47:10 -0500

I prefer “git in my belly”. (Austin Powers reference).

"Git: Push it! (real good)" by Korepetycje

Sat, 30 Aug 2008 11:42:20 -0500

Very nice wideo!

"Flash Message Conductor" by Korepetycje

Sat, 30 Aug 2008 06:43:36 -0500

usefull plugin! Thank you!

"Git: Push it! (real good)" by weepy ;...(

Sat, 30 Aug 2008 04:32:13 -0500

Bill Withers had a song about Subversion way before this :

http://www.youtube.com/watch?v=sBw6UZP0aJ0

:D:D

"Flash Message Conductor" by Robby Russell

Sat, 30 Aug 2008 00:32:47 -0500

@Lar Van Der Jagt,

We actually specify the class on the paragraph tag because we sometimes have multiple types of flash messages to render. For example, we might show a notice and a message at the same time (with different colors to highlight them individually). It’s fairly rare that we have this happen… but it has.

"Flash Message Conductor" by Robby Russell

Sat, 30 Aug 2008 00:30:58 -0500

James,

Doh! I thought that I covered that. Just added a patch for that. :-)

"Flash Message Conductor" by James Chan

Fri, 29 Aug 2008 23:53:52 -0500

It’s a great plugin. It would be even better if it outputs nothing instead of an empty div tag when there’s no error or message set.

"Flash Message Conductor" by Lar Van Der Jagt

Fri, 29 Aug 2008 20:41:25 -0500

I have a personal plugin where I store things like this as well. I’ve taken a slightly different approach that attempts to combine the flash with error_messages_for. I’m not sure I like how it works anymore, but I agree that simple plugins to store this kind of code is a good way to go.

You can also use multiple class definitions on the parent div, so the class would look something like class=”flash message” or class=”flash error”. That lets you style all flash messages one way, with slight differences for each type of message.

"Flash Message Conductor" by Jason

Fri, 29 Aug 2008 18:18:46 -0500

While not the most advanced plugin that I have seen, it does save time when working across several projects. I’d guesstimate that our team copy & pastes about 10-15 various components from projects with one of them being flash helpers similar to yours. You’re right, it’s a good idea to consider breaking things down into smaller pieces and making them easy to reuse.

Thanks for the idea!

"ShortURL on Github" by Sukanta

Tue, 26 Aug 2008 02:01:29 -0500

Join My Side: http://EjobsJustFree.com

"Parsing a RSS Feed" by Dimitri Mallis

Sat, 23 Aug 2008 16:00:56 -0500

thanks for that quickie :)

"Those that Tend the Store need Dialogue" by William

Sat, 23 Aug 2008 02:28:25 -0500

The article is about using communication to correct misinterpretations of a project’s requirements, goals, etc. (The article calls such misinterpretations “misguided perceptions”).

The article is not about developers harboring negative emotions for a project or client. The article never mentions “negative perceptions.” Such a phrase is found only in the comments.

Cheers, —William

"Installing Ruby on Rails and PostgreSQL on OS X, Third Edition" by Nick G

Thu, 21 Aug 2008 18:29:24 -0500

I didnt get the error worked just as explained !

Thanks !

free Rapidshare premium account

"The Daily Stand Up" by Nick G

Thu, 21 Aug 2008 18:19:31 -0500

Great ! I love rails action !

free Rapidshare premium account

"RSpec: It Should Behave Like" by John Yerhot

Thu, 21 Aug 2008 15:27:52 -0500

I’m pretty new to Rspec, but this post and everyone’s comments/thoughts are great. Thanks!

"RSpec: It Should Behave Like" by Ryan Bates

Thu, 21 Aug 2008 13:18:30 -0500

I prefer to put shared spec behavior in a controller macros file. Then you can include that in your spec helper:

config.include(ControllerMacros, :type => :controller)

Now all the methods in ControllerMacros are available to use in any controller specs.

"RSpec: It Should Behave Like" by Chuck Remes

Thu, 21 Aug 2008 10:54:21 -0500

This has been one of the most useful blog postings ever (for me). While I was already using #shared_examples_for to great effect, I was not creating helper methods for my #before blocks. I am currently refactoring a beast of a spec to use these techniques.

Thank you for the post and the comments!

"RSpec: It Should Behave Like" by ander

Thu, 21 Aug 2008 01:39:33 -0500

What I was trying to say is that shared controller specs have an easy access to the controller by just saying “controller”, whereas shared model specs have to do some custom stuff to get to the model.

Some thoughts: http://exceptionisarule.blogspot.com/2008/08/rspec-shared-model-specs.html

"RSpec: It Should Behave Like" by ander

Thu, 21 Aug 2008 01:36:16 -0500

I was trying to point out that shared controller specs have easy access to the controller by just saying “controller”, whereas shared model specs apparently have to do some custom stuff to get to the model.

Here are some thoughts: http://exceptionisarule.blogspot.com/2008/08/rspec-shared-model-specs.html

"Things.app syncs with the iPhone!" by Fred Brunel

Wed, 20 Aug 2008 20:01:48 -0500

I’ve been using Things for couple of days and it’s pretty neat. The iPhone app works prefectly.

However, I thing the iPhone version lacks two things: (1) a button for adding items to the inbox whatever the page you’re looking at, (2) tags

"RSpec: It Should Behave Like" by Mark Wilden

Wed, 20 Aug 2008 18:33:44 -0500

One caveat is that if a shared spec fails, you don’t see it in the call stack.

David Chelimsky suggested using—backtrace, but I never got around to trying it.

"RSpec: It Should Behave Like" by Robby Russell

Wed, 20 Aug 2008 10:07:28 -0500

@Bryan,

Thanks for sharing your thoughts on using a method to do this as well. I’ll have to play around a bit and see how readable I can make this, which is one of my goals when writing specs.

@ander:

Yep, you can use it_should_behave_like in model specs as well.

@ed:

Great tutorial, I’ll have to give that a whirl in the near future.

"RSpec: It Should Behave Like" by Glenn

Wed, 20 Aug 2008 06:41:18 -0500

I think this is one of the most underused aspects of Rspec, and it’s so powerful. I only discovered it because a colleague refactored some tests when I was on holiday, so the more you can do to get the news out there the better ;)

"RSpec: It Should Behave Like" by Ed Spencer

Wed, 20 Aug 2008 05:59:32 -0500

Nice to see this feature getting some exposure, in fact you inspired me to write up how I handle shared example groups when testing CRUD actions. Shared Example Groups are an awesome but underused feature of the RSpec library.

One minor point though – you’re currently putting the expectation that Dohicky.should_receive( :paginate ).and_return( Array.new ) in your before(:each) block. It might just be a matter of one style vs another but I would probably use stub! there and put the expectation into an it “should” do … end block, as you’re currently making the same expectation many times over.

I think you’re also missing the :shared => true at the top of your last code example.

"RSpec: It Should Behave Like" by Dylan

Wed, 20 Aug 2008 03:57:25 -0500

I agree with Bryan, and personally think that shared examples can be a good substitute for always having to use the before block for everything. For example, using them for before_filters.

Once you test the before_filter in isolation, you should be free to stub it out in other tests, so that those tests are then being tested in isolation. It reads well too.

In regard to using them for the logged_in example, I’d personally opt for showing that within the actual request. Here’s an example of how I would usually do this fwiw:

shared_examples_for "Green Dohicky In Context" do
  before 
    @mock_green_dohicky = mock_model(Dohicky, {:color => 'green'})
    @mock_green_dohicky.should_receive(:aoeu).and_return(:stnh)
end

describe DohickiesController do
  describe "responding to GET /new" 
    describe "as an admin user" do
      it_should_behave_like "Green Dohicky In Context" 

      before do
        Dohicky.stub!(:paginate).and_return(Array.new)
      end

      def do_get
         logged_in_as_admin { get :new }
      end

      it "should return an array from a call to Dohicky.paginate" do
        before_get { Dohicky.should_receive(:paginate).and_return(Array.new) }
      end

      it "should do something with the green dohicky" do
         during_get { @mock_green_dohicky.should do_something }
      end

      it "should do something else with the green dohicky" do
        ...
      end
    end

   describe "as a non logged in user" do
     it_should_behave_like "Green Dohicky In Context" 
       ...
     end
   end
 end

But yeah, would usually test the before_filters specific to that controller in that test, and push the global shared_examples out to spec_helper or another module.

Oh… and I also use this awesomeness (that should be a word) for cleaning up some of the scope fun… ie: making the request in the right place: http://gist.github.com/6343

What do you think ? That’s my “seems-to-be-ever-changing” controller approach at the moment ;)

"RSpec: It Should Behave Like" by Nick

Wed, 20 Aug 2008 01:40:27 -0500

We generally add a login_user method to spec_helper. Then we can use that anywhere we need it. Similarly for login_admin_user.

"RSpec: It Should Behave Like" by ander

Wed, 20 Aug 2008 01:17:35 -0500

Conveniently there’s the controller instance you can get to in (shared) controller specs. Is there something similar built in for model specs? Or do I have to define e.g. a get_instance method in the actual model spec (which includes the it_should_behave_like ..) to get an instance of the model in the shared spec?

Cheers.

"RSpec: It Should Behave Like" by grosser

Wed, 20 Aug 2008 00:58:54 -0500

I normally also go with just defining a method, it not ‘as’ readable but you can pass arguments and they are easy to find via method-lookup(eclipse).

"RSpec: It Should Behave Like" by Lar Van Der Jagt

Wed, 20 Aug 2008 00:50:03 -0500

Glad to see you posting some more, any further RSpec tips are especially appreciated!

"RSpec: It Should Behave Like" by Bryan Helmkamp

Tue, 19 Aug 2008 23:28:49 -0500

I’m not a big fan of this usage of shared examples. Expanded, here’s what the code in your example says:

“Admin::DohickiesController#new should behave like an admin user is signed in.”

The problem is it’s taking what’s context and putting it in the place of an outcome (as in, the piece after the should).

That said, I agree that duplicating the stub for login_required all over is not ideal because you might refactor to a different login system later. I usually handle this by defining a method in my specs like “build_mock_user_and_login”.

If all I care about is being logged in, I can call that from the before(:each) or inside each example. The method returns a mock_model, so I can also capture the result of it and use that object in the remainder of the example. One nice thing about this approach is that if I refactor away from login_required, I only have one place to update for all the specs that aren’t specifically concerned with login.

You do have to repeat the call to build_mock_user_and_login, but that duplication isn’t really any worse than duplicating “it_should_behave_like ‘an admin user is signed in’”. Both are essentially declarative.

Cheers,

-Bryan