From martindemello at gmail.com  Wed Aug  1 09:56:08 2007
From: martindemello at gmail.com (Martin DeMello)
Date: Wed, 1 Aug 2007 19:26:08 +0530
Subject: merb speed
In-Reply-To: <46AF2D26.2030501@anl.gov>
References: <f93a6bcc0707310524l54fae232kfddb0adf121f3624@mail.gmail.com>
	<46AF2D26.2030501@anl.gov>
Message-ID: <f93a6bcc0708010656o38cc50d8o5288a92ab4d46237@mail.gmail.com>

Thamks, grabbed a copy :)

martin

On 7/31/07, Randall Potter <rpotter at anl.gov> wrote:
> Watch this:  http://peepcode.com/products/benchmarking-with-httperf  and
> you will no longer be an idiot about benchmarking. :)
>
> --R
>
> Martin DeMello wrote:
> > Hi,
> >
> > We have a rails app which is essentially a web UI to a unix
> > filesystem, and which has been performing rather slowly. I've been
> > looking at rewriting it in merb, so I did a quick spike where I ported
> > over all the rails code necessary to log in as a user, and generate
> > the same file list page that rails was serving up. However, the
> > numbers appear quite bad (twice as slow as rails, for a naive
> > benchmark), leading me to believe that I'm doing something wrong. I'll
> > try to get permission to post some code, but on general principles are
> > there any pitfalls to look for when converting rails code to merb?
> >
> > Also, I'm trying to do some sort of rough client-side benchmarking
> > using the following code - is there anything obviously wrong with it?
> > I'm getting numbers of over 10 seconds per page load when I have 50
> > users repeatedly requesting a dynamic page in a loop:
> >
> > client = HTTPAccess2::Client.new
> > client.set_cookie_store("cookie.dat")
> > client.post(url, {:user_login => login, :user_password => pwd})
> > client_time = []
> > all_client_avg = []
> > all_server_avg = []
> > $total_client_avg = []
> > $total_server_avg = []
> > num_iterations.times do |i|
> >       folders = YAML.load(client.get_content("#{url}/folder/all_user_folders"))
> >       folders.each do |folder|
> >               client_time << Benchmark.realtime do
> >                       response_size = client.post("#{url}/resources/set_current_folder",
> > {:folder => $folder}).content.length
> >               end
> >       end
> >
> >       client_avg = (client_time.inject(0) {|sum, element| sum + element}) /
> > client_time.length
> > end
> >
> > I'm rather new to both web development and benchmarking, so please
> > feel free to talk to me as if I were an idiot :)
> >
> > martin
> > _______________________________________________
> > Merb-devel mailing list
> > Merb-devel at rubyforge.org
> > http://rubyforge.org/mailman/listinfo/merb-devel
> >
>
> --
> "When something is too hard it means that you're not cheating enough." --DHH
>
> _______________________________________________
> Merb-devel mailing list
> Merb-devel at rubyforge.org
> http://rubyforge.org/mailman/listinfo/merb-devel
>

From cocoa.guy at gmail.com  Wed Aug  1 17:52:41 2007
From: cocoa.guy at gmail.com (Cocoa Guy)
Date: Wed, 1 Aug 2007 14:52:41 -0700
Subject: Deploying with Merb
In-Reply-To: <fa49c1d70707311508na03ab1bjcaf83e5bbcd2e23@mail.gmail.com>
References: <fa49c1d70707311508na03ab1bjcaf83e5bbcd2e23@mail.gmail.com>
Message-ID: <fa49c1d70708011452n7ab7917bi4ae30e6ba902287d@mail.gmail.com>

I ended up using an apache rewrite rule to redirect file uploads to
merb. It took me enough time to get this working, that I'm sharing it
in case others find it useful:

RewriteEngine On
RewriteCond %{HTTP:Content-Type} multipart/form-data
#RewriteCond %{REQUEST_URI} your-regex
RewriteCond %{REQUEST_METHOD} POST
RewriteRule ^/(.*) http://127.0.0.1:4000/$1 [P]

With the commented out line, this will let merb handle all multipart
post requests. You can un-comment the 3rd line and add your own
expression to further limit what is handed to Merb.

Cheers,
-pete

From mattaimonetti at gmail.com  Thu Aug  2 03:38:09 2007
From: mattaimonetti at gmail.com (Matt Aimonetti)
Date: Thu, 2 Aug 2007 00:38:09 -0700
Subject: Deploying with Merb
In-Reply-To: <fa49c1d70708011452n7ab7917bi4ae30e6ba902287d@mail.gmail.com>
References: <fa49c1d70707311508na03ab1bjcaf83e5bbcd2e23@mail.gmail.com>
	<fa49c1d70708011452n7ab7917bi4ae30e6ba902287d@mail.gmail.com>
Message-ID: <f58d8dcb0708020038jc293d10v142b799b8918c16c@mail.gmail.com>

Thanks,

 Did anyone tried to do something like that with nginx ?

-Matt

On 8/1/07, Cocoa Guy <cocoa.guy at gmail.com> wrote:
>
> I ended up using an apache rewrite rule to redirect file uploads to
> merb. It took me enough time to get this working, that I'm sharing it
> in case others find it useful:
>
> RewriteEngine On
> RewriteCond %{HTTP:Content-Type} multipart/form-data
> #RewriteCond %{REQUEST_URI} your-regex
> RewriteCond %{REQUEST_METHOD} POST
> RewriteRule ^/(.*) http://127.0.0.1:4000/$1 [P]
>
> With the commented out line, this will let merb handle all multipart
> post requests. You can un-comment the 3rd line and add your own
> expression to further limit what is handed to Merb.
>
> Cheers,
> -pete
> _______________________________________________
> Merb-devel mailing list
> Merb-devel at rubyforge.org
> http://rubyforge.org/mailman/listinfo/merb-devel
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://rubyforge.org/pipermail/merb-devel/attachments/20070802/4b1f6cd0/attachment.html 

From ry at tinyclouds.org  Thu Aug  2 16:34:37 2007
From: ry at tinyclouds.org (ry dahl)
Date: Thu, 2 Aug 2007 22:34:37 +0200
Subject: controller exceptions
In-Reply-To: <E559214A-A2A8-45B5-BDA0-C29E7958054F@engineyard.com>
References: <21ee31950707311057o173344d7v72506524087067f1@mail.gmail.com>
	<E559214A-A2A8-45B5-BDA0-C29E7958054F@engineyard.com>
Message-ID: <21ee31950708021334g57b60a29lffd18916c40885e1@mail.gmail.com>

>         Can you make a small proof of concept patch that we can benchmark a
> bit and then decide?

Attached is a patch which implements the controller exception api.
It's hacky and not meant for production.

In merb_exceptions.rb is the definition of
Merb::ControllerExceptions::Base. That is an abstract base class.
Derived from Base are a bunch of other classes:
class NoContent < Base; STATUS = 204; end
class ResetContent < Base; STATUS = 205; end
class PartialContent < Base; STATUS = 206; end
class MultipleChoices < Base; STATUS = 300; end

To use this, one would derive a class from one of these and write a
method :merb_action (i didn't use :action because of name conflicts)
which should work more or less like a typical controller action. The
patch contains specs showing it in action with render :layout
This should be enough to benchmark?

The erb file for MyException#merb_action is looked for at
app/views/exceptions/my_exception.erb


ry
-------------- next part --------------
A non-text attachment was scrubbed...
Name: controller_exceptions.diff
Type: application/octet-stream
Size: 9315 bytes
Desc: not available
Url : http://rubyforge.org/pipermail/merb-devel/attachments/20070802/e035470e/attachment.obj 

From ry at tinyclouds.org  Thu Aug  2 17:53:56 2007
From: ry at tinyclouds.org (ry dahl)
Date: Thu, 2 Aug 2007 23:53:56 +0200
Subject: random code questions
Message-ID: <21ee31950708021453x88e1467rbdb57279595f4cee@mail.gmail.com>

merb_dispatcher.rb:28
controller = klass.build(request.body, request.params, route, response)
Why not just use a typical constructor?

merb_dispatcher.rb:35
raise Merb::HTTPMethodNotAllowed.new(method, allowed)
Why not check for this after the action has been dispatched to the controller?

controller_mixin.rb:149
def query_parse(qs, d = '&;')
Why not offload query parsing onto Mongrel as much as possible?

From ez at engineyard.com  Thu Aug  2 18:08:04 2007
From: ez at engineyard.com (Ezra Zygmuntowicz)
Date: Thu, 2 Aug 2007 15:08:04 -0700
Subject: random code questions
In-Reply-To: <21ee31950708021453x88e1467rbdb57279595f4cee@mail.gmail.com>
References: <21ee31950708021453x88e1467rbdb57279595f4cee@mail.gmail.com>
Message-ID: <D05D4596-61B9-4618-A037-919011011ECF@engineyard.com>


On Aug 2, 2007, at 2:53 PM, ry dahl wrote:

> merb_dispatcher.rb:28
> controller = klass.build(request.body, request.params, route,  
> response)
> Why not just use a typical constructor?

For testability. I just recently separated this out. We need to be  
able to initialize and get a controller instance before we call  
parse_request to parse the request. This is so you can get an  
instance of your controller in your specs and setup mocks and whatnot  
before you call parse_request.  build just wraps up the same  
functionality into one method so it works exactly the same in  
Dispacther.handle. This is just to increase testability.


>
> merb_dispatcher.rb:35
> raise Merb::HTTPMethodNotAllowed.new(method, allowed)
> Why not check for this after the action has been dispatched to the  
> controller?

Because there is no valid controller/action to call if this gets  
raised. It's enforcing the rules of routing which happens before  
dispatching to the controller. I'm not opposed to changing the  
exception raised but this check does belong where it is IMHO.


>
> controller_mixin.rb:149
> def query_parse(qs, d = '&;')
> Why not offload query parsing onto Mongrel as much as possible?

Because mongrel does not have a method that parses nested parameters.  
so mongrel's query_parse will fail to parse ?foo[bar][baz]=nik but  
merb's query_parse will correctly handle it. I agree it's a bit  
obsfuscated though ;) Could possible be merged and refactored with  
normailize_params. I'll put it on the todo list.


Cheers-
-- Ezra Zygmuntowicz 
-- Founder & Ruby Hacker
-- ez at engineyard.com
-- Engine Yard, Serious Rails Hosting
-- (866) 518-YARD (9273)



From canadaduane at gmail.com  Sat Aug  4 15:36:00 2007
From: canadaduane at gmail.com (Duane Johnson)
Date: Sat, 4 Aug 2007 13:36:00 -0600
Subject: How do you run an app on "edge merb"?
Message-ID: <96C54309-DB41-43C7-A365-EE78F5E8B666@gmail.com>

I didn't quite see a direct answer to Matt's question regarding  
running merb from SVN:
	http://rubyforge.org/pipermail/merb-devel/2007-July/000132.html

In a merb application, is there something comparable to having rails  
in the vendor/ folder of a rails app?

Thanks,

Duane Johnson
(canadaduane)

From merb at dusty.name  Sat Aug  4 15:43:35 2007
From: merb at dusty.name (Dusty Doris)
Date: Sat, 4 Aug 2007 15:43:35 -0400
Subject: observers
Message-ID: <e9e3ad7b0708041243q242e5f01x224ab363cc4c6626@mail.gmail.com>

Just discovered something, thought I'd share in case anyone else was
trying to use active record observers.  In rails you define your
observers in the environment.rb and it instantiates them somewhere
during startup.

To use that in merb you have to define the observers and then also
instantiate them.  Add something like this to the end of the
merb_init.rb file.

ActiveRecord::Base.observers << :attachment_observer
ActiveRecord::Base.instantiate_observers

At least its working for me from merb -i right now.

From merb at dusty.name  Sat Aug  4 15:46:16 2007
From: merb at dusty.name (Dusty Doris)
Date: Sat, 4 Aug 2007 15:46:16 -0400
Subject: Fwd: How do you run an app on "edge merb"?
In-Reply-To: <e9e3ad7b0708041245o7b2e951coeef46c72f09ded59@mail.gmail.com>
References: <96C54309-DB41-43C7-A365-EE78F5E8B666@gmail.com>
	<e9e3ad7b0708041245o7b2e951coeef46c72f09ded59@mail.gmail.com>
Message-ID: <e9e3ad7b0708041246v51096b8bu228ee9df786963a2@mail.gmail.com>

forgot to cc the list


On 8/4/07, Duane Johnson <canadaduane at gmail.com> wrote:
> I didn't quite see a direct answer to Matt's question regarding
> running merb from SVN:
>         http://rubyforge.org/pipermail/merb-devel/2007-July/000132.html
>
> In a merb application, is there something comparable to having rails
> in the vendor/ folder of a rails app?
>
> Thanks,
>

I don't know if there is something like that, but you can just install edge.

  $ svn co http://svn.devjavu.com/merb/trunk merb
  $ cd merb
  $ rake install

That will install trunk as a gem on your system.

From ez at engineyard.com  Sat Aug  4 15:51:47 2007
From: ez at engineyard.com (Ezra Zygmuntowicz)
Date: Sat, 4 Aug 2007 12:51:47 -0700
Subject: How do you run an app on "edge merb"?
In-Reply-To: <e9e3ad7b0708041246v51096b8bu228ee9df786963a2@mail.gmail.com>
References: <96C54309-DB41-43C7-A365-EE78F5E8B666@gmail.com>
	<e9e3ad7b0708041245o7b2e951coeef46c72f09ded59@mail.gmail.com>
	<e9e3ad7b0708041246v51096b8bu228ee9df786963a2@mail.gmail.com>
Message-ID: <EFA6B9A1-644E-403D-970E-BB7C887C77C7@engineyard.com>


On Aug 4, 2007, at 12:46 PM, Dusty Doris wrote:

> forgot to cc the list
>
>
> On 8/4/07, Duane Johnson <canadaduane at gmail.com> wrote:
>> I didn't quite see a direct answer to Matt's question regarding
>> running merb from SVN:
>>         http://rubyforge.org/pipermail/merb-devel/2007-July/ 
>> 000132.html
>>
>> In a merb application, is there something comparable to having rails
>> in the vendor/ folder of a rails app?
>>
>> Thanks,
>>
>
> I don't know if there is something like that, but you can just  
> install edge.
>
>   $ svn co http://svn.devjavu.com/merb/trunk merb
>   $ cd merb
>   $ rake install
>
> That will install trunk as a gem on your system.


	Yes there are some built in rake tasks for freezing the framework:

rake merb:freeze           # freeze the merb framework from installed  
gem into dist/framework for portability
rake merb:freeze_from_svn  # freeze the merb framework from svn trunk
rake merb:unfreeze         # unfreeze this app from the framework and  
use system gem.

	So merb:freeze_from_svn will freeze the current trunk into your app.  
When you freeze merb into your app you need to use script/merb  
instead of plain old merb to start your server.


Cheers-
-- Ezra Zygmuntowicz 
-- Founder & Ruby Hacker
-- ez at engineyard.com
-- Engine Yard, Serious Rails Hosting
-- (866) 518-YARD (9273)



From mattaimonetti at gmail.com  Sun Aug  5 23:19:26 2007
From: mattaimonetti at gmail.com (Matt Aimonetti)
Date: Sun, 5 Aug 2007 20:19:26 -0700
Subject: new patch adding control_for :boolean to create a checkbox
Message-ID: <f58d8dcb0708052019k19e4845fk3bad3ad6cdf9304f@mail.gmail.com>

I'm sorry I didn't have time working on the tests/models generator but here
is a tiny patch I wrote to help you creating form check boxes.

http://merb.devjavu.com/projects/merb/ticket/114


-Matt
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://rubyforge.org/pipermail/merb-devel/attachments/20070805/e42cd930/attachment.html 

From canadaduane at gmail.com  Mon Aug  6 10:39:19 2007
From: canadaduane at gmail.com (Duane Johnson)
Date: Mon, 6 Aug 2007 08:39:19 -0600
Subject: Proposal: Pure Regex Router
Message-ID: <2895272E-FAE0-4FF9-A561-77569F03075B@gmail.com>

So this has been an idea ever since Rails came out.  Why not use a  
pure regular expression router in similar fashion to gsub?  I imagine  
it working something like this:

Merb::Router.prepare do |r|
   r.add %r[^/(.*)/(.*)/(\d+)$], :controller => '\1', :action =>  
'\2', :id => '\3'
   r.add %r[^/user-(\d+)$], :user_id => '\1', :controller =>  
'free_area/users', :action => 'index'
end

The flexibility is endless in this system (far more possibilities  
than Rails' system) and it lets people immediately grasp routes if  
then understand regular expressions (which they should, otherwise,  
this is a good opportunity to learn).

In addition, I propose we add more options to the router so that the  
protocol, port, and domain become available:

   r.add :path => %r[^/([^\/,?]*)$], :domain => /^(blog|news) 
\./, :controller => '\1', :action => 'index'

Once this is set up, we could potentially add a convenience layer on  
top so that we get something a little more like our current routes:

r.add ":controller/:action"
# Detection of a string causes r.add to create a regexp route  
matching the string, e.g.:
# r.add %r[^/([^/,?]+)/([^/,?]+)], :controller => '\1', :action => '\2'

Thoughts?

Duane Johnson
(canadaduane)

From marcus.roberts at gmail.com  Tue Aug  7 15:07:20 2007
From: marcus.roberts at gmail.com (Marcus Roberts)
Date: Tue, 7 Aug 2007 20:07:20 +0100
Subject: Upload progress and drb
Message-ID: <4e0e8ed70708071207x192c0ddbw67e9e510e7e8e16@mail.gmail.com>

Hi,

Sorry if this is a stupid question!


I used to have a site running on Merb 0.1.0.   When running as a cluster, it
would automatically start a drb process to co-ordinate the upload progress
monitoring.


Since uploading to 0.3.7 (a bit of a jump!), the drb process hasn't started
automatically.    I seem to start one if I run merb -s 32323 manually, but
when I do check on a file upload progress, I just get a nil back.


Before digging into the upload side of it, I suspect that the drb not
starting automatically might be the deeper cause of the problem.  Should drb
start automatically?


In my merb.yml  I have


:config: dist/conf/upload.conf
:drb_server_port: 32323

and in dist/conf/upload.conf I have


:drb: true
:path_info: /upload/applet_upload_file
:frequency: 3



Thanks for any help/pointers!
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://rubyforge.org/pipermail/merb-devel/attachments/20070807/9f01368d/attachment.html 

From scaudill at gmail.com  Tue Aug  7 15:43:01 2007
From: scaudill at gmail.com (Stephen Caudill)
Date: Tue, 7 Aug 2007 15:43:01 -0400
Subject: Upload progress and drb
In-Reply-To: <4e0e8ed70708071207x192c0ddbw67e9e510e7e8e16@mail.gmail.com>
References: <4e0e8ed70708071207x192c0ddbw67e9e510e7e8e16@mail.gmail.com>
Message-ID: <A5521544-AAC3-4260-B015-87F863D95FC6@gmail.com>

add

:start_drb: true

to your merb.yml and you should be good to go.

- v.dolo

On Aug 7, 2007, at 3:07 PM, Marcus Roberts wrote:

> Hi,
>
>
> Sorry if this is a stupid question!
>
>
> I used to have a site running on Merb 0.1.0.   When running as a  
> cluster, it would automatically start a drb process to co-ordinate  
> the upload progress monitoring.
>
>
> Since uploading to 0.3.7 (a bit of a jump!), the drb process hasn't  
> started automatically.    I seem to start one if I run merb -s  
> 32323 manually, but when I do check on a file upload progress, I  
> just get a nil back.
>
>
> Before digging into the upload side of it, I suspect that the drb  
> not starting automatically might be the deeper cause of the  
> problem.  Should drb start automatically?
>
>
> In my merb.yml  I have
>
>
> :config: dist/conf/upload.conf
> :drb_server_port: 32323
>
> and in dist/conf/upload.conf I have
>
>
> :drb: true
> :path_info: /upload/applet_upload_file
> :frequency: 3
>
>
>
> Thanks for any help/pointers!
>
>
>
>
>
>
>
>
> _______________________________________________
> Merb-devel mailing list
> Merb-devel at rubyforge.org
> http://rubyforge.org/mailman/listinfo/merb-devel


From canadaduane at gmail.com  Wed Aug  8 01:10:37 2007
From: canadaduane at gmail.com (Duane Johnson)
Date: Tue, 7 Aug 2007 23:10:37 -0600
Subject: Proposal: Pure Regex Router (and router wishlist)
In-Reply-To: <3D0820A7-5E51-470F-A33F-D46995BFC802@gmail.com>
References: <2895272E-FAE0-4FF9-A561-77569F03075B@gmail.com>
	<3D0820A7-5E51-470F-A33F-D46995BFC802@gmail.com>
Message-ID: <01138350-8B7B-4601-BF52-E6F9AB7B883E@gmail.com>


On Aug 7, 2007, at 2:41 AM, Luke Sutton wrote:

>> So this has been an idea ever since Rails came out.  Why not use a
>> pure regular expression router in similar fashion to gsub?  I imagine
>> it working something like this:
>>
>> Merb::Router.prepare do |r|
>>    r.add %r[^/(.*)/(.*)/(\d+)$], :controller => '\1', :action =>
>> '\2', :id => '\3'
>>    r.add %r[^/user-(\d+)$], :user_id => '\1', :controller =>
>> 'free_area/users', :action => 'index'
>> end
>
> Certainly very powerful, but I'm a little bit dubious of it's  
> usefulness ? is your intent to use regular expressions like this to  
> reduce redundancy in route declaration?
>
> In that case I think the routing DSL could be extended to make that  
> possible. That would be friendlier to noobs like yours truly.
>

I realize I'm fighting against long-held tradition.  Luckily, I'm so  
much against the current way of things as I am deterred by its  
inflexibility at times (or at least, the inflexibility of Rails  
Routes).  I feel like we should have a Pure Regex Router as the  
foundation, and then other router-adding styles built on top of that.

Some of the things I've tried to do that I couldn't:

- Use symbols other than "/" to denote a partition between segments.   
For example:
	/show-user-1
	/edit-movie-42

- Use specific characters to reduce the size of URLs, i.e. map url  
parts to other things.  For example, suppose I wanted to map the CRUD 
+ actions like this:
	i => index
	n => new
	c => create
	r => show
	e => edit
	u => update
	d => delete

	Such that the following URLs are valid:
	/users-i (maps to 'users' controller, 'index' action)
	/movies-u (maps to 'movies' controller, 'update' action)

	r.add %r[^/([^-]+)-([increud])$], :controller => '\1' do |match|
		map = {'i' => 'index', 'n' => 'new', 'c' => 'create'}  # etc.
		{:controller => match[1], :action => map[match[2]]}
	end

- Map segments to other segments.  For example, suppose you have  
several controllers within an admin "module", and let's say you  
decide later that these controllers should be externally accessible  
from the "superadmin" folder as well.  Currently, you would need to  
create a route for each controller:

	r.add '/superadmin/users/:action/:id', :controller => 'admin/users'
	r.add '/superadmin/movies/:action/:id', :controller => 'admin/movies'
	r.add '/superadmin/categories/:action/:id', :controller => 'admin/ 
categories'
	etc.

	Instead, we should be able to do something like this:

	r.add '/superadmin/:controller_segment/:action/:id', :controller =>  
'admin/:controller_segment'

- Type-agnostic routing.  For example, suppose I have a database of  
many types of objects that all share a universally unique identifier  
such as a GUID.  By looking at the URL only, I can't tell what the  
object is--so, first I need to do a database lookup to determine what  
type of object the referenced thing is and then route it to a  
controller/action pair.  For example:

	http://localhost:4000/555598

	r.add '^/(\d+)$' do |match|
		if obj = MyObject.find(match[1])
			{:controller => obj.controller, :action => 'index'}
		else
			false
		end
	end

	In the above example, the router can only determine where to route  
to after a database lookup.


>> In addition, I propose we add more options to the router so that the
>> protocol, port, and domain become available:
>>
>>    r.add :path => %r[^/([^\/,?]*)$], :domain => /^(blog|news)
>> \./, :controller => '\1', :action => 'index'
>
> Now this is a top idea! I can imagine many situations where this  
> would be handy. For example you might wish to route secure  
> connections differently.
>
> In addition to protocol, port and domain, it would be very useful  
> to have access to sub-domains in the route declarations.

Yes, it would be good.  Once we finish this discussion and everyone  
is happy with the ideas, I'll see if I can add incorporate them into  
Merb.

Duane

From canadaduane at gmail.com  Wed Aug  8 17:19:38 2007
From: canadaduane at gmail.com (Duane Johnson)
Date: Wed, 8 Aug 2007 15:19:38 -0600
Subject: Fwd: Proposal: Pure Regex Router (and router wishlist)
In-Reply-To: <f214044a0708081417g6cfb6223u78c40dd8afeb545c@mail.gmail.com>
References: <2895272E-FAE0-4FF9-A561-77569F03075B@gmail.com>
	<3D0820A7-5E51-470F-A33F-D46995BFC802@gmail.com>
	<01138350-8B7B-4601-BF52-E6F9AB7B883E@gmail.com>
	<2fff50390708072234v3997aca8ocbd7f957beb6a019@mail.gmail.com>
	<f214044a0708081417g6cfb6223u78c40dd8afeb545c@mail.gmail.com>
Message-ID: <f214044a0708081419k27d5d288n3fc82f06f0c6ad8a@mail.gmail.com>

Forgot to reply to list.

---------- Forwarded message ----------
From: Duane Johnson <canadaduane at gmail.com>
Date: Aug 8, 2007 3:17 PM
Subject: Re: Proposal: Pure Regex Router (and router wishlist)
To: Daniel N <has.sox at gmail.com>


I realize I'm fighting against long-held tradition.  Luckily, I'm so
> much against the current way of things as I am deterred by its
> inflexibility at times (or at least, the inflexibility of Rails
> Routes).  I feel like we should have a Pure Regex Router as the
> foundation, and then other router-adding styles built on top of that.
>
>
Correction: The above should read, "I'm *not* so much against the current
way of things, as I am deterred by its inflexibility at times."

Sorry about the harsh sounding original text.




On 8/7/07, Daniel N <has.sox at gmail.com> wrote:
>
>
>
> I'm not sure why you would want the above in any routes.  Merb supports
> restful routes so I don't see why you would want what you suggested there.
>
> You end up mixing the verb into the noun and get multiple url's for a
> resourse just to do crud.  That's what restful routes gets us away from.
> Personally I like the restful conventions and find that it does nice things
> to the way I think of stuff.
>


Agreed. I see that it was not a very convincing example, but the point was
to show how an idea outside the box is currently "impossible" without
rewriting the router.  My objective is to find a solution with the fewest
possible restrictions (a superset) so that the current best-practice ( i.e.
restful routes and rails-alike routes) becomes a subset of the solution.

 The other two examples, map segments to other segments, and type agnostic
> routing look interesting to me though.
>

 Just my 0.02
>

Thanks for your feedback on these.

Duane Johnson
(canadaduane)
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://rubyforge.org/pipermail/merb-devel/attachments/20070808/0fe29115/attachment.html 

From marcus.roberts at gmail.com  Thu Aug  9 08:55:17 2007
From: marcus.roberts at gmail.com (Marcus Roberts)
Date: Thu, 9 Aug 2007 13:55:17 +0100
Subject: File upload issue with Merb 0.3
Message-ID: <4e0e8ed70708090555v6604272s2d25e71dc7054b51@mail.gmail.com>

Hello,
I have written a Java applet that uses Merb to process file uploads.

This works perfectly on Merb 0.1.0.

When I upgraded to 0.3.7, all seems to work well, but I then started to see
problems - when people uploaded certain files, the ruby process would spin
out and start using lots of CPU, never coming back to life.  Other files
work just fine.

The problem is almost certainly in my Java file handling - when I upload the
same problem file through an HTML upload form, it works fine.

The one thing I did notice was that in the merb log, I would see the first
line of the request log entry, e.g.:

Request: REQUEST_URI:
/upload/applet_upload_file?token=98e4e4f3b55db4ea848b13ad7
01d4aaa,46bae8bc,9084,192,33,1  (2007-08-09 11:13:17)

but the next line, e.g. Params:{ ... never appeared - the process would just
hang.

I assume that Mongrel is doing the work of receiving the file (which is
about 8k) before handing it back to Merb, and this is where the problem is
occuring, but to help me track it down, does anyone know of any big changes
in the way file uploads worked in 0.1.0 and 0.3.7, or could someone give a
runthrough of the process when a file upload occurs, and where best to start
looking to debug.

Thanks

Marcus
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://rubyforge.org/pipermail/merb-devel/attachments/20070809/2fe9229e/attachment-0001.html 

From ry at tinyclouds.org  Fri Aug 10 13:20:09 2007
From: ry at tinyclouds.org (ry dahl)
Date: Fri, 10 Aug 2007 19:20:09 +0200
Subject: serving static files
Message-ID: <21ee31950708101020g6e9816eak32a01959a4aa906d@mail.gmail.com>

It seems that Merb is sending static files with Mongrel::DirHandler.
(mongrel_handler.rb:52)
    if get_or_head and @files.can_serve(path_info)
      # File exists as-is so serve it up
      MERB_LOGGER.info("Serving static file: #{path_info}")
      @files.process(request,response)

I haven't done benchmarks and I've hardly glanced at DirHandler's code
but the comment on DirHandler#send_file bothers me:

    # Sends the contents of a file back to the user. Not terribly
efficient since it's
    # opening and closing the file for each read.

Would it be better for Merb to use X-Sendfile in production mode?

ry

From ez at engineyard.com  Fri Aug 10 21:43:29 2007
From: ez at engineyard.com (Ezra Zygmuntowicz)
Date: Fri, 10 Aug 2007 18:43:29 -0700
Subject: serving static files
In-Reply-To: <21ee31950708101020g6e9816eak32a01959a4aa906d@mail.gmail.com>
References: <21ee31950708101020g6e9816eak32a01959a4aa906d@mail.gmail.com>
Message-ID: <45F446ED-D105-48AA-AA7D-69BDA11B0454@engineyard.com>


On Aug 10, 2007, at 10:20 AM, ry dahl wrote:

> It seems that Merb is sending static files with Mongrel::DirHandler.
> (mongrel_handler.rb:52)
>     if get_or_head and @files.can_serve(path_info)
>       # File exists as-is so serve it up
>       MERB_LOGGER.info("Serving static file: #{path_info}")
>       @files.process(request,response)
>
> I haven't done benchmarks and I've hardly glanced at DirHandler's code
> but the comment on DirHandler#send_file bothers me:
>
>     # Sends the contents of a file back to the user. Not terribly
> efficient since it's
>     # opening and closing the file for each read.
>
> Would it be better for Merb to use X-Sendfile in production mode?
>
> ry
>

Hey Ry-

	Merb only sends static files with the dir handler if you are running  
it by itself. When it is behind a proxy it doesn't serve any static  
files itself at all.

Cheers-
-- Ezra Zygmuntowicz 
-- Founder & Ruby Hacker
-- ez at engineyard.com
-- Engine Yard, Serious Rails Hosting
-- (866) 518-YARD (9273)



From ivey at gweezlebur.com  Sat Aug 11 22:37:46 2007
From: ivey at gweezlebur.com (Michael D. Ivey)
Date: Sat, 11 Aug 2007 21:37:46 -0500
Subject: [PATCH] [TINY] make scripts executable
Message-ID: <87AAB8E6-8FB5-489A-A205-938BF16406C9@gweezlebur.com>

Super simple patch to chmod 755 the scripts dir in the generator

-------------- next part --------------
A non-text attachment was scrubbed...
Name: merb-executable-scripts.diff
Type: application/octet-stream
Size: 510 bytes
Desc: not available
Url : http://rubyforge.org/pipermail/merb-devel/attachments/20070811/c1f6d24b/attachment.obj 

From ivey at gweezlebur.com  Sat Aug 11 22:14:12 2007
From: ivey at gweezlebur.com (Michael D. Ivey)
Date: Sat, 11 Aug 2007 21:14:12 -0500
Subject: [PATCH] merb-rakefile-more-dist-byebye.diff
Message-ID: <75EB8CFA-987A-4048-B6B8-90275CF6763F@gweezlebur.com>

In rev 410, "rake db:migrate" doesn't run because the Rake tasks are  
still looking for dist/

This patch removes all the references to dist/ that I could find.

Ez, consider this a full copyright transfer from me to you on this  
and future patches; I'll sign something official if you need it.


-------------- next part --------------
A non-text attachment was scrubbed...
Name: merb-rakefile-more-dist-byebye.diff
Type: application/octet-stream
Size: 8643 bytes
Desc: not available
Url : http://rubyforge.org/pipermail/merb-devel/attachments/20070811/bb3bf027/attachment.obj 
-------------- next part --------------



From ez at engineyard.com  Sun Aug 12 00:16:26 2007
From: ez at engineyard.com (Ezra Zygmuntowicz)
Date: Sat, 11 Aug 2007 21:16:26 -0700
Subject: [PATCH] merb-rakefile-more-dist-byebye.diff
In-Reply-To: <75EB8CFA-987A-4048-B6B8-90275CF6763F@gweezlebur.com>
References: <75EB8CFA-987A-4048-B6B8-90275CF6763F@gweezlebur.com>
Message-ID: <6400BD6D-22F1-4816-83AB-23DE4A7F3BF5@engineyard.com>


On Aug 11, 2007, at 7:14 PM, Michael D. Ivey wrote:

> In rev 410, "rake db:migrate" doesn't run because the Rake tasks  
> are still looking for dist/
>
> This patch removes all the references to dist/ that I could find.
>
> Ez, consider this a full copyright transfer from me to you on this  
> and future patches; I'll sign something official if you need it.
>
>
> <merb-rakefile-more-dist-byebye.diff>
>


Hey Michael-

	Thanks for the patches!  Can you do me a favor and submit these to  
the trac? http://merb.devjavu.com/

Cheers-

-- Ezra Zygmuntowicz 
-- Founder & Ruby Hacker
-- ez at engineyard.com
-- Engine Yard, Serious Rails Hosting
-- (866) 518-YARD (9273)



From ivey at gweezlebur.com  Sun Aug 12 08:14:05 2007
From: ivey at gweezlebur.com (Michael D. Ivey)
Date: Sun, 12 Aug 2007 07:14:05 -0500
Subject: [PATCH] merb-rakefile-more-dist-byebye.diff
In-Reply-To: <6400BD6D-22F1-4816-83AB-23DE4A7F3BF5@engineyard.com>
References: <75EB8CFA-987A-4048-B6B8-90275CF6763F@gweezlebur.com>
	<6400BD6D-22F1-4816-83AB-23DE4A7F3BF5@engineyard.com>
Message-ID: <D035D666-98AA-43F0-B931-B6D07AF2640D@gweezlebur.com>

On Aug 11, 2007, at 11:16 PM, Ezra Zygmuntowicz wrote:
> 	Thanks for the patches!  Can you do me a favor and submit these to  
> the trac? http://merb.devjavu.com/

Will do.  Didn't know there was one. :)


From luke.sutton at gmail.com  Mon Aug 13 07:42:30 2007
From: luke.sutton at gmail.com (Luke Sutton)
Date: Mon, 13 Aug 2007 21:12:30 +0930
Subject: Problems running merb from trunk
Message-ID: <B4E9B7CB-BAAE-4FE1-B325-1BA6077A2718@gmail.com>

I have no problems loading any static requests, but any that are  
routed to a controller blow up. I'm getting a 'Bad controller' error.  
The reason I'm a bit flummoxed is because the stack trace makes no  
sense.

It terminates on this line:

resolve_controller 	68  	/opt/local/lib/ruby/gems/1.8/gems/merb-0.3.7/ 
lib/merb/merb_dispatcher.rb

As you can see, this is in the merb-0.3.7 gem, but earlier in the  
stack trace I can see this:

run 	290  	/Users/lukesutton/Pending/merb_test/framework/lib/merb/ 
server.rb

So, even though I'm running merb from the framework dir, for some  
reason it's calling code in the gem and blowing up. This makes no  
sense to me :)

I'm running this using the script/merb script.

You guys have any clues to what I'm doing wrong?

-- 
luke



From dscataglini at dancingbeargroup.com  Tue Aug 14 13:22:11 2007
From: dscataglini at dancingbeargroup.com (Diego Scataglini)
Date: Tue, 14 Aug 2007 13:22:11 -0400
Subject: Does anybody know why activeRecord is not built on top of ruby-dbi?
Message-ID: <b0644b490708141022o3ab283ddjac9a2e90e950ceb0@mail.gmail.com>

I am curious about the opinionated opinion behind the decision.Was it
performance? Complexity? Databases support or something else.

I recently used Ruby:dbi and liked all the metadata that comes with the
data.
Performance looked good but I didn't benchmark it so I don't know which one
execute faster.

If I am not wrong Ruby:dbi offers true parameters binding which would be
nice to have.

Let me know if anybody knows why.

Thanks

Diego Scataglini
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://rubyforge.org/pipermail/merb-devel/attachments/20070814/7e7b388c/attachment.html 

From ry at tinyclouds.org  Fri Aug 17 13:08:21 2007
From: ry at tinyclouds.org (ry dahl)
Date: Fri, 17 Aug 2007 19:08:21 +0200
Subject: Cookie Sessions in Merb?
Message-ID: <21ee31950708171008s767efbc7wc0c91b270d76d020@mail.gmail.com>

I like the cookie sessions that Rails edge has - they make sense,
they're fast, easy.

For those that don't know: the default session mechanism is to give
developers a Hash called session. To store this object between
requests it Marshals the session object and sends the object itself
(now as a bitstream) back to clients to store in their cookie file. To
prevent users from tampering with this data, a cryptographic digest
(using HMAC with SHA-1) is sent along with the bitstream).

The major problem with this scheme is that the Marshalled data is not
at all encrypted! Users are free to unmarshal and examine what
websites are storing in their sessions. They cannot change the data,
but nevertheless this is a security problem!

The answers given by the rails developers sound rather na?ve:
> This is getting into very difficult crypto. Basically, depending on the cipher, there
> may be cases where you can change one part of the text without changing the other.
Not that I know much about cryptography, but I'm almost certain that
modifying AES encrypted Marshaled data is quite tamper proof. If not,
one could at least send a digest along with it.

There has been some mention of the extra AES operation slowing down
speed (but it's been said it's not much of a slowdown), I think having
secure cookies would be well worth the cost.

Is there interest in this type of session for Merb?
Related: I've just wrote a CookieJar class - perhaps it would be
useful for Merb?
http://rubyforge.org/pipermail/mongrel-users/2007-August/003890.html

ry

From ez at engineyard.com  Fri Aug 17 13:38:43 2007
From: ez at engineyard.com (Ezra Zygmuntowicz)
Date: Fri, 17 Aug 2007 10:38:43 -0700
Subject: Cookie Sessions in Merb?
In-Reply-To: <21ee31950708171008s767efbc7wc0c91b270d76d020@mail.gmail.com>
References: <21ee31950708171008s767efbc7wc0c91b270d76d020@mail.gmail.com>
Message-ID: <45E6D9AC-84A8-4EBD-A1A2-B012204A55E7@engineyard.com>

Yeah I'd be interested in a cookie store. But th security concerns are  
valid so we would need to investigate thouroughly

Ezra

Sent from my iPhone

On Aug 17, 2007, at 10:08 AM, "ry dahl" <ry at tinyclouds.org> wrote:

> I like the cookie sessions that Rails edge has - they make sense,
> they're fast, easy.
>
> For those that don't know: the default session mechanism is to give
> developers a Hash called session. To store this object between
> requests it Marshals the session object and sends the object itself
> (now as a bitstream) back to clients to store in their cookie file. To
> prevent users from tampering with this data, a cryptographic digest
> (using HMAC with SHA-1) is sent along with the bitstream).
>
> The major problem with this scheme is that the Marshalled data is not
> at all encrypted! Users are free to unmarshal and examine what
> websites are storing in their sessions. They cannot change the data,
> but nevertheless this is a security problem!
>
> The answers given by the rails developers sound rather na?ve:
>> This is getting into very difficult crypto. Basically, depending on  
>> the cipher, there
>> may be cases where you can change one part of the text without  
>> changing the other.
> Not that I know much about cryptography, but I'm almost certain that
> modifying AES encrypted Marshaled data is quite tamper proof. If not,
> one could at least send a digest along with it.
>
> There has been some mention of the extra AES operation slowing down
> speed (but it's been said it's not much of a slowdown), I think having
> secure cookies would be well worth the cost.
>
> Is there interest in this type of session for Merb?
> Related: I've just wrote a CookieJar class - perhaps it would be
> useful for Merb?
> http://rubyforge.org/pipermail/mongrel-users/2007-August/003890.html
>
> ry
> _______________________________________________
> Merb-devel mailing list
> Merb-devel at rubyforge.org
> http://rubyforge.org/mailman/listinfo/merb-devel

From mattaimonetti at gmail.com  Mon Aug 20 21:12:21 2007
From: mattaimonetti at gmail.com (Matt Aimonetti)
Date: Mon, 20 Aug 2007 18:12:21 -0700
Subject: plugins loaded after the models?
Message-ID: <f58d8dcb0708201812l44a5ad78xa68e5148534c332c@mail.gmail.com>

Hi,

 I was trying to port attachment_fu to Merb and I realized that adding

has_attachment(
    :content_type => :image,
[...]
)

in a model would break the app.

I realized (let me know if I'm wrong) that the plugins are loaded after the
models.  Because of that I now have to load my plugin from my model to be
able to use it directly.

Did I miss something or is there a specific reason why things are loaded the
way they are?

Thanks,

-Matt
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://rubyforge.org/pipermail/merb-devel/attachments/20070820/05a0b379/attachment.html 

From dscataglini at dancingbeargroup.com  Thu Aug 23 13:25:44 2007
From: dscataglini at dancingbeargroup.com (Diego Scataglini)
Date: Thu, 23 Aug 2007 13:25:44 -0400
Subject: Mailer delivery execution question (drb or not)
Message-ID: <b0644b490708231025s70982968hf9fee3f8d08e66fc@mail.gmail.com>

When I send an email with Mailer does it get processed in a separate
detached thread (a-la background-drb) or not?
If not the follow up would be:
- Is this something that would be a future feature?
- Does anybody know of a way to get background-drb working with merb?
(it's probably an obvious thing but I am pretty new to the project)

Thanks
Diego

From ivey at gweezlebur.com  Sat Aug 25 21:40:19 2007
From: ivey at gweezlebur.com (Michael D. Ivey)
Date: Sat, 25 Aug 2007 20:40:19 -0500
Subject: class reloading
Message-ID: <AB596A59-4964-4C96-B808-ACF9DF03D0BF@gweezlebur.com>

Ezra, what's the plan for class reloading in dev mode?  I was  
thinking of tackling that as a project, if no one else is working on it.

From canadaduane at gmail.com  Sat Aug 25 23:52:29 2007
From: canadaduane at gmail.com (Duane Johnson)
Date: Sat, 25 Aug 2007 21:52:29 -0600
Subject: Routes patch w/ specs
Message-ID: <B7972D7A-AA2D-4811-B1B5-9C5C462324CE@gmail.com>

The new routes system is available as a patch, with specs.  There are  
a few specs that fail (6) having to do with mail and content  
negotiation.  I could use some help figuring these ones out.  Lots of  
things have changed, mostly in the dispatcher and controller classes.

Let me know what you think, and then let's get this patch applied.

http://merb.devjavu.com/projects/merb/ticket/131

Duane

From canadaduane at gmail.com  Mon Aug 27 10:49:21 2007
From: canadaduane at gmail.com (Duane Johnson)
Date: Mon, 27 Aug 2007 08:49:21 -0600
Subject: New Router blog post
Message-ID: <C1DCE5C3-BFC1-441C-AF4E-F8BEE8A3E815@gmail.com>

FYI, I blogged about the new system here: http://blog.inquirylabs.com/ 
2007/08/27/new-router-in-merb/

Duane Johnson
(canadaduane)

From mattaimonetti at gmail.com  Tue Aug 28 18:37:47 2007
From: mattaimonetti at gmail.com (Matt Aimonetti)
Date: Tue, 28 Aug 2007 18:37:47 -0400
Subject: what's the view format in trunk?
Message-ID: <f58d8dcb0708281537o6610f93am5c3068ad1ac3f5d0@mail.gmail.com>

dumb question but what's the view format in trunk?

I have a simple router:

Merb::Router.prepare do |r|
  # default route, usually you don't want to change this
  r.default_routes

  # change this for your home page to be avaiable at /
  r.add '/', :controller => 'upload', :action =>'new'
end

I have a simple controller which should render the view:

  def new
    @photo = Photo.new

    respond_to do |format|
      format.html { render }
    end
  end

and in my app/views folder I have an upload folder and inside it a
new.html.erb file  but it doesn't get picked up.


No template matched
/Users/matta/rails_projects/my_app/trunk/upload_manager/app/views/upload/new.{}

I tried adding a .herb file and .html but it didn't help.

Thanks,

-Matt
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://rubyforge.org/pipermail/merb-devel/attachments/20070828/2ee8e56c/attachment.html 

From boss at topfunky.com  Tue Aug 28 23:45:25 2007
From: boss at topfunky.com (Geoffrey Grosenbach)
Date: Tue, 28 Aug 2007 20:45:25 -0700
Subject: Non-Erubis Templates
Message-ID: <15754f7e0708282045w78ca2c72wdb9351e2763bde53@mail.gmail.com>

Trunk Issue: Because of the use of 'autoload', template handlers other
than Erubis are not loaded automatically (Haml, XMLBuilder). Either
this should be fixed, or the documentation should be updated to
instruct people how to use non-Erb template engines.

Apparently the solution is to do something like this in merb_init.rb:

  ::Merb::AbstractController.register_engine ::Merb::Template::Haml, %w[ haml ]

So the question is:

* Is this the permanent situation for the future?
* Can I update the docs to reflect this requirement?

Geoffrey Grosenbach
http://peepcode.com

From B.Candler at pobox.com  Wed Aug 29 01:46:16 2007
From: B.Candler at pobox.com (Brian Candler)
Date: Wed, 29 Aug 2007 06:46:16 +0100
Subject: Non-Erubis Templates
In-Reply-To: <20070829053848.GB26518@uk.tiscali.com>
References: <15754f7e0708282045w78ca2c72wdb9351e2763bde53@mail.gmail.com>
	<20070829053848.GB26518@uk.tiscali.com>
Message-ID: <20070829054616.GA27128@uk.tiscali.com>

On Wed, Aug 29, 2007 at 06:38:48AM +0100, Brian Candler wrote:
> > * Can I update the docs to reflect this requirement?
> 
> It's now documented the merb_init.rb which is installed by merb -g. But I
> agree, History.txt should have info about upgrading existing projects.
> 
> Please see:
> http://merb.devjavu.com/projects/merb/ticket/130
> http://merb.devjavu.com/projects/merb/ticket/133

While you're at it, please document the other changed needed to
merb_init.rb, which is

-conn_options = YAML::load(Erubis::Eruby.new(IO.read("#{DIST_ROOT}/conf/database.yml")).result)
+conn_options = YAML::load(Erubis::Eruby.new(IO.read("#{DIST_ROOT}/conf/database.yml")).result(binding))

See http://merb.devjavu.com/projects/merb/ticket/118

Again, this should be correct for newly generated projects.

From B.Candler at pobox.com  Wed Aug 29 01:35:43 2007
From: B.Candler at pobox.com (Brian Candler)
Date: Wed, 29 Aug 2007 06:35:43 +0100
Subject: what's the view format in trunk?
In-Reply-To: <f58d8dcb0708281537o6610f93am5c3068ad1ac3f5d0@mail.gmail.com>
References: <f58d8dcb0708281537o6610f93am5c3068ad1ac3f5d0@mail.gmail.com>
Message-ID: <20070829053543.GA26518@uk.tiscali.com>

On Tue, Aug 28, 2007 at 06:37:47PM -0400, Matt Aimonetti wrote:
>    dumb question but what's the view format in trunk?
>    I have a simple router:
>    Merb::Router.prepare do |r|
>      # default route, usually you don't want to change this
>      r.default_routes
> 
>      # change this for your home page to be avaiable at /
>      r.add '/', :controller => 'upload', :action =>'new'
>    end
>    I have a simple controller which should render the view:
>      def new
>        @photo = Photo.new
>        respond_to do |format|
>          format.html { render }
>        end
>      end
>    and in my app/views folder I have an upload folder and inside it a
>    new.html.erb file  but it doesn't get picked up.
>    No template matched
>    /Users/matta/rails_projects/my_app/trunk/upload_manager/app/views/uplo
>    ad/new.{}
>    I tried adding a .herb file and .html but it didn't help.

You now need to explicitly load the ERB handler, since it's set to autoload.
Just add the following line to conf/merb_init.rb:

Merb::Template::Erubis

If you start a fresh project you should get a suitable merb_init.rb

Regards,

Brian.

From B.Candler at pobox.com  Wed Aug 29 01:38:48 2007
From: B.Candler at pobox.com (Brian Candler)
Date: Wed, 29 Aug 2007 06:38:48 +0100
Subject: Non-Erubis Templates
In-Reply-To: <15754f7e0708282045w78ca2c72wdb9351e2763bde53@mail.gmail.com>
References: <15754f7e0708282045w78ca2c72wdb9351e2763bde53@mail.gmail.com>
Message-ID: <20070829053848.GB26518@uk.tiscali.com>

On Tue, Aug 28, 2007 at 08:45:25PM -0700, Geoffrey Grosenbach wrote:
> Trunk Issue: Because of the use of 'autoload', template handlers other
> than Erubis are not loaded automatically (Haml, XMLBuilder). Either
> this should be fixed, or the documentation should be updated to
> instruct people how to use non-Erb template engines.
> 
> Apparently the solution is to do something like this in merb_init.rb:
> 
>   ::Merb::AbstractController.register_engine ::Merb::Template::Haml, %w[ haml ]

You just need to reference the class in merb_init.rb to trigger the
autoload, e.g.

Merb::Template::Haml

> So the question is:
> 
> * Is this the permanent situation for the future?
> * Can I update the docs to reflect this requirement?

It's now documented the merb_init.rb which is installed by merb -g. But I
agree, History.txt should have info about upgrading existing projects.

Please see:
http://merb.devjavu.com/projects/merb/ticket/130
http://merb.devjavu.com/projects/merb/ticket/133

From B.Candler at pobox.com  Wed Aug 29 03:36:30 2007
From: B.Candler at pobox.com (Brian Candler)
Date: Wed, 29 Aug 2007 08:36:30 +0100
Subject: Non-Erubis Templates
In-Reply-To: <20070829053848.GB26518@uk.tiscali.com>
References: <15754f7e0708282045w78ca2c72wdb9351e2763bde53@mail.gmail.com>
	<20070829053848.GB26518@uk.tiscali.com>
Message-ID: <20070829073629.GA336@uk.tiscali.com>

On Wed, Aug 29, 2007 at 06:38:48AM +0100, Brian Candler wrote:
> You just need to reference the class in merb_init.rb to trigger the
> autoload, e.g.
> 
> Merb::Template::Haml
> 
> > So the question is:
> > 
> > * Is this the permanent situation for the future?

On reflection, it doesn't seem to make much sense to set up autoloading, but
then to have to explicitly trigger it. Based on the Merb KISS principle, I
think it would be cleaner to just put

require 'merb/template/haml'

in merb_init.rb, and get rid of the autoload magic (although there's only a
tiny bit of that, in lib/merb/template.rb)

Does anyone else share this opinion?

Perhaps examples/skeleton/conf/merb_init.rb should be changed to say

--------------------------------------------------------------------------
# Uncomment the template handers which your application needs to load
require 'merb/template/erubis'		# .herb .jerb .erb .rhtml .html.erb .js.erb
# require 'merb/template/haml'		# .haml
# require 'merb/template/markaby'	# .mab
# require 'merb/template/xml_builder'	# .rxml .xerb .builder
--------------------------------------------------------------------------

and then it doesn't matter whether the autoload stuff is there or not.

Regards,

Brian.

From luke.sutton at gmail.com  Wed Aug 29 03:49:47 2007
From: luke.sutton at gmail.com (Luke Sutton)
Date: Wed, 29 Aug 2007 17:19:47 +0930
Subject: Non-Erubis Templates
In-Reply-To: <20070829073629.GA336@uk.tiscali.com>
References: <15754f7e0708282045w78ca2c72wdb9351e2763bde53@mail.gmail.com>
	<20070829053848.GB26518@uk.tiscali.com>
	<20070829073629.GA336@uk.tiscali.com>
Message-ID: <379B42F2-09BC-4928-B15F-5470578E763D@gmail.com>


On 29/08/2007, at 5:06 PM, Brian Candler wrote:

> On Wed, Aug 29, 2007 at 06:38:48AM +0100, Brian Candler wrote:
>> You just need to reference the class in merb_init.rb to trigger the
>> autoload, e.g.
>>
>> Merb::Template::Haml
>>
>>> So the question is:
>>>
>>> * Is this the permanent situation for the future?
>
> On reflection, it doesn't seem to make much sense to set up  
> autoloading, but
> then to have to explicitly trigger it. Based on the Merb KISS  
> principle, I
> think it would be cleaner to just put
>
> require 'merb/template/haml'
>
> in merb_init.rb, and get rid of the autoload magic (although  
> there's only a
> tiny bit of that, in lib/merb/template.rb)
>
> Does anyone else share this opinion?

I agree with that. Merb's mantra should be 'less magic'.

That said I'm not so sure about requiring vs. having an actual method  
for loading the template engine. Then in the future the templating  
code can me moved about without disturbing users. Also requiring like  
that means the code will be pulled in via a gem right? A method for  
loading the template classes would allow users to stash Merb in the  
framework directory without having to change the path used in the  
require.

how about something like:

templating :erubis

Just a thought :)

-- 
Luke


From ivey at gweezlebur.com  Wed Aug 29 10:07:28 2007
From: ivey at gweezlebur.com (Michael D. Ivey)
Date: Wed, 29 Aug 2007 09:07:28 -0500
Subject: Non-Erubis Templates
In-Reply-To: <20070829073629.GA336@uk.tiscali.com>
References: <15754f7e0708282045w78ca2c72wdb9351e2763bde53@mail.gmail.com>
	<20070829053848.GB26518@uk.tiscali.com>
	<20070829073629.GA336@uk.tiscali.com>
Message-ID: <7682FD46-9A2B-4F5B-B3AC-1DD1C44550AD@gweezlebur.com>

On Aug 29, 2007, at 2:36 AM, Brian Candler wrote:
> On reflection, it doesn't seem to make much sense to set up  
> autoloading, but
> then to have to explicitly trigger it. Based on the Merb KISS  
> principle, I

What about

------------------------------------------------------------------------ 
--
# Uncomment the template handers which your application needs to load
Merb::Template::Erubis         
# .herb .jerb .erb .rhtml .html.erb .js.erb
# Merb::Template::Haml        # .haml
# Merb::Template:::Markaby    # .mab
# Merb::Template::XMLBuilder  # .rxml .xerb .builder
------------------------------------------------------------------------ 
--

To me, this is even clearer what you're getting.  The more I look at  
the autoload, the
less it feels like magic.



Although, Luke's "templating :erubis" suggestion (possibly using the  
autoload mechanism in the
background) seems even better, now that I see it.

It could also be smart enough to know that :erb means :erubis, :rxml  
means :xmlbuilder, etc:

templating :erb, :rxml

From ry at tinyclouds.org  Wed Aug 29 10:19:50 2007
From: ry at tinyclouds.org (ry dahl)
Date: Wed, 29 Aug 2007 16:19:50 +0200
Subject: Non-Erubis Templates
In-Reply-To: <20070829073629.GA336@uk.tiscali.com>
References: <15754f7e0708282045w78ca2c72wdb9351e2763bde53@mail.gmail.com>
	<20070829053848.GB26518@uk.tiscali.com>
	<20070829073629.GA336@uk.tiscali.com>
Message-ID: <21ee31950708290719q289bdcbaqf345b9e4168de826@mail.gmail.com>

Hi,

Why not just autoload Haml (or whatever) when it encounters a .haml template?
Perhaps it doesn't do that now but surely this wouldn't be difficult.

ry

From B.Candler at pobox.com  Wed Aug 29 10:30:37 2007
From: B.Candler at pobox.com (Brian Candler)
Date: Wed, 29 Aug 2007 15:30:37 +0100
Subject: Non-Erubis Templates
In-Reply-To: <7682FD46-9A2B-4F5B-B3AC-1DD1C44550AD@gweezlebur.com>
References: <15754f7e0708282045w78ca2c72wdb9351e2763bde53@mail.gmail.com>
	<20070829053848.GB26518@uk.tiscali.com>
	<20070829073629.GA336@uk.tiscali.com>
	<7682FD46-9A2B-4F5B-B3AC-1DD1C44550AD@gweezlebur.com>
Message-ID: <20070829143037.GA20492@uk.tiscali.com>

On Wed, Aug 29, 2007 at 09:07:28AM -0500, Michael D. Ivey wrote:
> On Aug 29, 2007, at 2:36 AM, Brian Candler wrote:
> > On reflection, it doesn't seem to make much sense to set up  
> > autoloading, but
> > then to have to explicitly trigger it. Based on the Merb KISS  
> > principle, I
> 
> What about
> 
> ------------------------------------------------------------------------ 
> --
> # Uncomment the template handers which your application needs to load
> Merb::Template::Erubis         
> # .herb .jerb .erb .rhtml .html.erb .js.erb
> # Merb::Template::Haml        # .haml
> # Merb::Template:::Markaby    # .mab
> # Merb::Template::XMLBuilder  # .rxml .xerb .builder
> ------------------------------------------------------------------------ 
> --
> 
> To me, this is even clearer what you're getting.

I actually wrote that :-)

> The more I look at the autoload, the less it feels like magic.

I could understand the benefit of this approach if the extensions were
pre-registered: e.g.

module Merb
  module Template
    autoload :Erubis, 'merb/template/erubis'
    register_template :Erubis, '.erb', '.rhtml'
  end
end

Then, Merb would search for .erb / .rhtml files, and if it found one and
needed to process it, the template handling code would be loaded at that
time.

But as it is, you get a chicken-and-egg situation: the template extensions
aren't registered until the class is used, but the class won't be used until
a template matches one of its extensions. So you have to explicitly 'touch'
the class to get it loaded. In that case, I argue you might as well just
load the class explicitly.

As a side point: the current approach also restricts you to the templating
modules which are defined in lib/merb/template.rb: namely,

module Merb
  module Template
    autoload :Erubis,  'merb/template/erubis'
    autoload :Haml,  'merb/template/haml'
    autoload :Markaby,  'merb/template/markaby'
    autoload :XMLBuilder,  'merb/template/xml_builder'
  end
end

Of course, a plugin can have its own init.rb which sets up autoload for its
own templating class, and then you can explicitly 'touch' the class in
merb_init.rb. But why jump through these hoops?

Regards,

Brian.

From B.Candler at pobox.com  Wed Aug 29 10:36:30 2007
From: B.Candler at pobox.com (Brian Candler)
Date: Wed, 29 Aug 2007 15:36:30 +0100
Subject: Non-Erubis Templates
In-Reply-To: <21ee31950708290719q289bdcbaqf345b9e4168de826@mail.gmail.com>
References: <15754f7e0708282045w78ca2c72wdb9351e2763bde53@mail.gmail.com>
	<20070829053848.GB26518@uk.tiscali.com>
	<20070829073629.GA336@uk.tiscali.com>
	<21ee31950708290719q289bdcbaqf345b9e4168de826@mail.gmail.com>
Message-ID: <20070829143630.GA21166@uk.tiscali.com>

On Wed, Aug 29, 2007 at 04:19:50PM +0200, ry dahl wrote:
> Why not just autoload Haml (or whatever) when it encounters a .haml template?
> Perhaps it doesn't do that now but surely this wouldn't be difficult.

Merb doesn't search for filenames ending in .haml unless the Haml extension
has been loaded.

Old Merb (0.3.7) simply loaded all the template extensions in regardless.
Therefore it would always check for .haml, and process it accordingly.

Ezra changed this in trunk r401:
http://merb.devjavu.com/projects/merb/changeset/401#file18

This doesn't seem to have been fully thought-through, because not loading
the Haml template library means that the code that registers '.haml' as a
known extension, and associates it with Merb::Template::Haml, is not run.

It would make sense if register_engine were called to register .haml outside
of Merb::Template::Haml (but register_engine would have to be modified to be
passed a symbol, not a class, because referencing the class would cause it
to be autoloaded anyway)

Regards,

Brian.

From ry at tinyclouds.org  Wed Aug 29 12:13:27 2007
From: ry at tinyclouds.org (ry dahl)
Date: Wed, 29 Aug 2007 18:13:27 +0200
Subject: Non-Erubis Templates
In-Reply-To: <21ee31950708290719q289bdcbaqf345b9e4168de826@mail.gmail.com>
References: <15754f7e0708282045w78ca2c72wdb9351e2763bde53@mail.gmail.com>
	<20070829053848.GB26518@uk.tiscali.com>
	<20070829073629.GA336@uk.tiscali.com>
	<21ee31950708290719q289bdcbaqf345b9e4168de826@mail.gmail.com>
Message-ID: <21ee31950708290913t78ebb24ftfcefa0d1360f061c@mail.gmail.com>

I've posted a little patch for this behavior
http://merb.devjavu.com/projects/merb/ticket/152

On 8/29/07, ry dahl <ry at tinyclouds.org> wrote:
> Hi,
>
> Why not just autoload Haml (or whatever) when it encounters a .haml template?
> Perhaps it doesn't do that now but surely this wouldn't be difficult.
>
> ry
>

From canadaduane at gmail.com  Wed Aug 29 12:46:26 2007
From: canadaduane at gmail.com (Duane Johnson)
Date: Wed, 29 Aug 2007 10:46:26 -0600
Subject: What makes Rails' routes non-thread-safe?
Message-ID: <5517B2BB-FB07-4E74-8B9E-63FE927C4BF6@gmail.com>

I was reading here and there about the differences between Merb and  
Rails [1] and became curious about statements such as this:

> It is threadsafe with configurable Mutex Locks (Routing is also  
> threadsafe)

Having mucked with the routes system, I had this awful feeling that I  
don't understand threading well enough to know if my contribution is  
"thread safe".  Can someone help me to understand what makes Merb's  
routes threadsafe while Rails' is not?  For example, does the late- 
bound procs I've added to the new routing system make it non-threadsafe?

Thanks,

Duane Johnson
(canadaduane)

[1] http://www.infoq.com/news/2007/08/performance-ruby-on-merb
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://rubyforge.org/pipermail/merb-devel/attachments/20070829/faca0c23/attachment.html 

From ez at engineyard.com  Wed Aug 29 14:52:38 2007
From: ez at engineyard.com (Ezra Zygmuntowicz)
Date: Wed, 29 Aug 2007 11:52:38 -0700
Subject: Non-Erubis Templates
In-Reply-To: <21ee31950708290913t78ebb24ftfcefa0d1360f061c@mail.gmail.com>
References: <15754f7e0708282045w78ca2c72wdb9351e2763bde53@mail.gmail.com>
	<20070829053848.GB26518@uk.tiscali.com>
	<20070829073629.GA336@uk.tiscali.com>
	<21ee31950708290719q289bdcbaqf345b9e4168de826@mail.gmail.com>
	<21ee31950708290913t78ebb24ftfcefa0d1360f061c@mail.gmail.com>
Message-ID: <2037C863-427B-40E5-BF07-F1F1557701D7@engineyard.com>


On Aug 29, 2007, at 9:13 AM, ry dahl wrote:

> I've posted a little patch for this behavior
> http://merb.devjavu.com/projects/merb/ticket/152
>
> On 8/29/07, ry dahl <ry at tinyclouds.org> wrote:
>> Hi,
>>
>> Why not just autoload Haml (or whatever) when it encounters  
>> a .haml template?
>> Perhaps it doesn't do that now but surely this wouldn't be difficult.
>>
>> ry
>>
> _______________________________________________
> Merb-devel mailing list
> Merb-devel at rubyforge.org
> http://rubyforge.org/mailman/listinfo/merb-devel


	Yeah the autoload of template engines was not thought thru enough on  
my part. I wanted to be able to not load haml, markaby and builder if  
your app doesnt use them since its just overhead and more dependencies.

	But the current situation does need to be changed. Whatever people  
think is best is fine with me.

Cheers-

-- Ezra Zygmuntowicz 
-- Founder & Ruby Hacker
-- ez at engineyard.com
-- Engine Yard, Serious Rails Hosting
-- (866) 518-YARD (9273)



From ivey at gweezlebur.com  Wed Aug 29 16:07:07 2007
From: ivey at gweezlebur.com (Michael D. Ivey)
Date: Wed, 29 Aug 2007 15:07:07 -0500
Subject: Non-Erubis Templates
In-Reply-To: <2037C863-427B-40E5-BF07-F1F1557701D7@engineyard.com>
References: <15754f7e0708282045w78ca2c72wdb9351e2763bde53@mail.gmail.com>
	<20070829053848.GB26518@uk.tiscali.com>
	<20070829073629.GA336@uk.tiscali.com>
	<21ee31950708290719q289bdcbaqf345b9e4168de826@mail.gmail.com>
	<21ee31950708290913t78ebb24ftfcefa0d1360f061c@mail.gmail.com>
	<2037C863-427B-40E5-BF07-F1F1557701D7@engineyard.com>
Message-ID: <A2A86427-9D84-4456-8881-8EA8CE3FA6A0@gweezlebur.com>

On Aug 29, 2007, at 1:52 PM, Ezra Zygmuntowicz wrote:
> On Aug 29, 2007, at 9:13 AM, ry dahl wrote:
>
>> I've posted a little patch for this behavior
>> http://merb.devjavu.com/projects/merb/ticket/152
>
> 	Yeah the autoload of template engines was not thought thru enough on
> my part. I wanted to be able to not load haml, markaby and builder if
> your app doesnt use them since its just overhead and more  
> dependencies.
>
> 	But the current situation does need to be changed. Whatever people
> think is best is fine with me.

+1 for #152.  It matches the described behavior of "You can use any  
of these
with no extra config" and it also gets the "Don't load it if it isn't  
needed.

From ez at engineyard.com  Wed Aug 29 16:48:00 2007
From: ez at engineyard.com (Ezra Zygmuntowicz)
Date: Wed, 29 Aug 2007 13:48:00 -0700
Subject: What makes Rails' routes non-thread-safe?
In-Reply-To: <5517B2BB-FB07-4E74-8B9E-63FE927C4BF6@gmail.com>
References: <5517B2BB-FB07-4E74-8B9E-63FE927C4BF6@gmail.com>
Message-ID: <E1B41642-B1C7-46E3-99E7-F9E7E4BBDBE8@engineyard.com>


On Aug 29, 2007, at 9:46 AM, Duane Johnson wrote:

> I was reading here and there about the differences between Merb and  
> Rails [1] and became curious about statements such as this:
>
>> It is threadsafe with configurable Mutex Locks (Routing is also  
>> threadsafe)
>
> Having mucked with the routes system, I had this awful feeling that  
> I don't understand threading well enough to know if my contribution  
> is "thread safe".  Can someone help me to understand what makes  
> Merb's routes threadsafe while Rails' is not?  For example, does  
> the late-bound procs I've added to the new routing system make it  
> non-threadsafe?
>
> Thanks,
>
> Duane Johnson
> (canadaduane)
>

Duane-

	Mostly the fact that is uses a bunch of alias_method_chain and  
method redefinition. I didn't dig into to it much further then that.  
I'll review your new routes to see if I think anything is not safe  
but I think we're fine.

Cheers-
-- Ezra Zygmuntowicz 
-- Founder & Ruby Hacker
-- ez at engineyard.com
-- Engine Yard, Serious Rails Hosting
-- (866) 518-YARD (9273)



From merb at dusty.name  Thu Aug 30 01:02:32 2007
From: merb at dusty.name (Dusty Doris)
Date: Thu, 30 Aug 2007 01:02:32 -0400
Subject: active record alternatives
Message-ID: <e9e3ad7b0708292202g432ed4c2mb12bf59c114bcc07@mail.gmail.com>

Just wanted to get the opinion from merb users on active record
alternatives that you are either using or watching.  I've been reading
a bit about sequel and data mapper.  Both look pretty nice.  Anyone
actively using those yet?  Any others?

If anyone cares to provide their thoughts, I'd love to hear them.

BTW - is this too off-topic for this list?  If so, just let me know
and I'll stop posting questions like this.

From ivey at gweezlebur.com  Thu Aug 30 02:01:35 2007
From: ivey at gweezlebur.com (Michael D. Ivey)
Date: Thu, 30 Aug 2007 01:01:35 -0500
Subject: Plugin Manager changes
Message-ID: <0C55FD28-7EA6-4CF8-A49E-8F83B944E7B3@gweezlebur.com>

I just committed some stuff in PluginManager to actually load the Gem- 
based plugins from the manifest.

Merb::PluginManager.load_plugins is what goes in merb_init to load  
them all.

For every plugin in the manifest, 'theplugin' is required, and then  
we try and require 'theplugin/merb_init' and, if that fails, we try  
'plugin/init'

I'll be working on this some more this week and next, to make sure it  
works as expected.  I'll also write some docs.

From ez at engineyard.com  Thu Aug 30 14:16:01 2007
From: ez at engineyard.com (Ezra Zygmuntowicz)
Date: Thu, 30 Aug 2007 11:16:01 -0700
Subject: active record alternatives
In-Reply-To: <e9e3ad7b0708292202g432ed4c2mb12bf59c114bcc07@mail.gmail.com>
References: <e9e3ad7b0708292202g432ed4c2mb12bf59c114bcc07@mail.gmail.com>
Message-ID: <1787183C-AA28-4CFE-9B4D-A1E654C797B1@engineyard.com>


On Aug 29, 2007, at 10:02 PM, Dusty Doris wrote:

> Just wanted to get the opinion from merb users on active record
> alternatives that you are either using or watching.  I've been reading
> a bit about sequel and data mapper.  Both look pretty nice.  Anyone
> actively using those yet?  Any others?
>
> If anyone cares to provide their thoughts, I'd love to hear them.
>
> BTW - is this too off-topic for this list?  If so, just let me know
> and I'll stop posting questions like this.
>


	BothSequel and Datamapper are viable alternatives. I will be pulling  
out anything ActiveRecord related from merb core and making it a  
plugin. So merb will be orm agnostic and you will install a plugin to  
get AR or Sequel or DataMapper when you generate a new app.

Cheers-
-- Ezra Zygmuntowicz 
-- Founder & Ruby Hacker
-- ez at engineyard.com
-- Engine Yard, Serious Rails Hosting
-- (866) 518-YARD (9273)



From B.Candler at pobox.com  Thu Aug 30 15:40:33 2007
From: B.Candler at pobox.com (Brian Candler)
Date: Thu, 30 Aug 2007 20:40:33 +0100
Subject: Why Regexp#freeze?
Message-ID: <20070830194033.GA15695@uk.tiscali.com>

I notice in recent commits that a number of these have been added:

    /\[(\d+)\]/.freeze

I thought that Regexps are immutable. Furthermore, Regexp literals don't
generate a new object each time they are encountered:

irb(main):001:0> 3.times { puts /\d+/.object_id }
-605707646
-605707646
-605707646
=> 3
irb(main):002:0>

Unlike strings:

irb(main):002:0> 3.times { puts "abc".freeze.object_id }
-605884514
-605884594
-605884674
=> 3

So, could someone please explain why .freeze is beneficial on regexp
literals?

Thanks,

Brian.

From mattaimonetti at gmail.com  Thu Aug 30 18:40:48 2007
From: mattaimonetti at gmail.com (Matt Aimonetti)
Date: Thu, 30 Aug 2007 18:40:48 -0400
Subject: problems with 'no Accept header' request
Message-ID: <f58d8dcb0708301540n436ca3c7m9ea426669e39b760@mail.gmail.com>

I'm trying to setup a Merb cluster + attachment_fu to handle uploads for a
client's Rails app.

I'm using Flash to upload the files and I realized that Merb was giving me
an error for each upload. It turns out that Flash doesn't seem to send an
Accept header and Merb's responder crashes because it's trying to split the
missing Accept header.

responder.rb line 37

        def initialize(accept_header, params={})
          MERB_LOGGER.info accept_header
          @accepts = Responder.parse(accept_header)
          @params = params
          @stack  = {}
        end

@accepts seems to be later on used in negotiate_content() l84,
negotiate_by_format() l92, negotiate_by_accept_header l103

What's the proper way of dealing with this problem?

Thanks,

-Matt
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://rubyforge.org/pipermail/merb-devel/attachments/20070830/b77f873a/attachment.html 

From canadaduane at gmail.com  Thu Aug 30 20:00:28 2007
From: canadaduane at gmail.com (Duane Johnson)
Date: Thu, 30 Aug 2007 18:00:28 -0600
Subject: New routing system has specs and examples
Message-ID: <989F1624-39BB-46C9-9814-3882A4978A38@gmail.com>

Check it out here:

http://pastie.textmate.org/92610

Let me know if you have any questions.

Duane Johnson
(canadaduane)
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://rubyforge.org/pipermail/merb-devel/attachments/20070830/c808d045/attachment.html 

From B.Candler at pobox.com  Fri Aug 31 01:48:10 2007
From: B.Candler at pobox.com (Brian Candler)
Date: Fri, 31 Aug 2007 06:48:10 +0100
Subject: problems with 'no Accept header' request
In-Reply-To: <f58d8dcb0708301540n436ca3c7m9ea426669e39b760@mail.gmail.com>
References: <f58d8dcb0708301540n436ca3c7m9ea426669e39b760@mail.gmail.com>
Message-ID: <20070831054810.GA12954@uk.tiscali.com>

On Thu, Aug 30, 2007 at 06:40:48PM -0400, Matt Aimonetti wrote:
> 
>    I'm trying to setup a Merb cluster + attachment_fu to handle uploads
>    for a client's Rails app.
>    I'm using Flash to upload the files and I realized that Merb was
>    giving me an error for each upload. It turns out that Flash doesn't
>    seem to send an Accept header and Merb's responder crashes because
>    it's trying to split the missing Accept header.
>    responder.rb line 37
>            def initialize(accept_header, params={})
>              MERB_LOGGER.info accept_header
>              @accepts = Responder.parse(accept_header)
>              @params = params
>              @stack  = {}
>            end
>    @accepts seems to be later on used in negotiate_content() l84,
>    negotiate_by_format() l92, negotiate_by_accept_header l103
>    What's the proper way of dealing with this problem?

RFC 2616:

"  If no Accept header field is present, then it is assumed that the
   client accepts all media types."

So I think you could make Responser.parse return [:all] in the event of nil
accept_header being passed in.

  return [:all] unless accept_header

B.

From B.Candler at pobox.com  Fri Aug 31 15:21:24 2007
From: B.Candler at pobox.com (Brian Candler)
Date: Fri, 31 Aug 2007 20:21:24 +0100
Subject: New routing system has specs and examples
In-Reply-To: <989F1624-39BB-46C9-9814-3882A4978A38@gmail.com>
References: <989F1624-39BB-46C9-9814-3882A4978A38@gmail.com>
Message-ID: <20070831192124.GB22846@uk.tiscali.com>

On Thu, Aug 30, 2007 at 06:00:28PM -0600, Duane Johnson wrote:
>    [1]http://pastie.textmate.org/92610
> 
>    Let me know if you have any questions.

This is very useful to understand what the new system is capable of, thank
you.

Could you explain if there some efficiency penalty in using the defer_to
syntax? For example,

r.match("/::/users/::").
        to(:controller => "users", :action => "[2]", :id => "[1]")

could potentially be rewritten as

r.match("/::/users/::").defer_to {
        {:controller => "users", :action => $2, :id => $1}
}

(I'm not sure $1, $2 etc are actually set at this point, but you get the
idea). Doing it this way would use more standard Ruby notation for captures
than "[1]" and "[2]", which is why I like it.

But does using the to(...) syntax allow you to compile the router into a
faster form?

Regards,

Brian.

From B.Candler at pobox.com  Fri Aug 31 16:51:50 2007
From: B.Candler at pobox.com (Brian Candler)
Date: Fri, 31 Aug 2007 21:51:50 +0100
Subject: Bootstrapping from SVN
Message-ID: <20070831205150.GA28101@uk.tiscali.com>

Is there a simple way to bootstrap merb directly from the Subversion
repository, rather than first installing it as a gem?

I can do the following if I already have an old gem lying around:

   merb -g testapp        # using an old gem
   cd testapp
   rake merb:freeze_from_svn

Unfortunately, that means I'm using an old version of the generator. So for
example, when I just tried this, the generator made a conf/ directory,
whereas the code in trunk expects config/.

Equally, I can't test any new features recently added into the generator
(such as the snazzy new screens committed in r464-r466), without first
building trunk into a gem, and gem installing it over my existing one.

I have tried:

svn co http://svn.devjavu.com/merb/trunk merb
cd merb
ruby -Ilib bin/merb -g /var/tmp/test123

but I get:

Couldn't find 'merb' generator
  /usr/lib/ruby/gems/1.8/gems/rubigen-1.0.3/lib/rubigen/lookup.rb:159:in `lookup'
  /usr/lib/ruby/gems/1.8/gems/rubigen-1.0.3/lib/rubigen/lookup.rb:164:in `instance'
  /usr/lib/ruby/gems/1.8/gems/rubigen-1.0.3/lib/rubigen/scripts/../scripts.rb:30:in `run'
  /v/build/merb-trunk/lib/merb/generators/merb_app/merb_app.rb:13:in `run'
  ./lib/merb/server.rb:216:in `run'
  bin/merb:6

Cheers,

Brian.

From canadaduane at gmail.com  Fri Aug 31 18:45:14 2007
From: canadaduane at gmail.com (Duane Johnson)
Date: Fri, 31 Aug 2007 16:45:14 -0600
Subject: New routing system has specs and examples
In-Reply-To: <20070831192124.GB22846@uk.tiscali.com>
References: <989F1624-39BB-46C9-9814-3882A4978A38@gmail.com>
	<20070831192124.GB22846@uk.tiscali.com>
Message-ID: <7206E91A-EF64-426A-BFE4-335D47834CFB@gmail.com>


On Aug 31, 2007, at 1:21 PM, Brian Candler wrote:

> On Thu, Aug 30, 2007 at 06:00:28PM -0600, Duane Johnson wrote:
>>    [1]http://pastie.textmate.org/92610
>>
>>    Let me know if you have any questions.
>
> This is very useful to understand what the new system is capable  
> of, thank
> you.
>
> Could you explain if there some efficiency penalty in using the  
> defer_to
> syntax? For example,
>
> r.match("/::/users/::").
>         to(:controller => "users", :action => "[2]", :id => "[1]")
>
> could potentially be rewritten as
>
> r.match("/::/users/::").defer_to {
>         {:controller => "users", :action => $2, :id => $1}
> }
>
> (I'm not sure $1, $2 etc are actually set at this point, but you  
> get the
> idea). Doing it this way would use more standard Ruby notation for  
> captures
> than "[1]" and "[2]", which is why I like it.
>
> But does using the to(...) syntax allow you to compile the router  
> into a
> faster form?
>

Yes, there is a slight penalty.  If you can match without deferring,  
your routes will be optimized in a big if/elsif statement.  With that  
said, however, the deferred block is evaluated AFTER all other  
conditions (e.g. only if the "/::/users/::" portion is a match), so  
you can get away with this slight performance decrease if you  
restrict your deferred matching to specific portions of the site.

With regard to your example above, I think this will work:

r.match("/:action/users/:id").defer_to { |request, params|
	# params = {:action => "some_action", :id => "some_id"}
	params.merge(:controller => "users")
}

I had some questions in IRC recently about the defer_to arguments, so  
perhaps I can repeat it here.

The defer_to block takes two arguments, which I prefer to call  
"request" and "params".  The simplest case of a deferred route block  
is as follows:

r.defer_to { |request, params| params }

It says, "This route is always a match. Return the params hash,  
unchanged".  You could add conditions, like so:

r.defer_to { |request, params| params if ExternalLibrary.matches? }

In the case where matches? returns true, the params hash will be  
returned, and the route will be evaluated as true.  In the false  
case, the if statement returns nil, the route match thus fails, and  
the routes system will continue checking for matches against other  
routes down the line.

Here's a more practical example of a type-agnostic route where we  
defer to a block at run-time to determine if the route is a match:

r.match("/object/:id").defer_to(:action => "index") do |request, params|
	thing = GenericObject.find_by_id(params[:id])
	if (thing)
		params.merge(:controller => thing.controller)
	end
end

As of revision 467 (a few minutes ago, as of this writing), the  
"params" hashes passed into the deferred blocks above now include the  
query_string's parsed params as well as the placeholders parsed out  
of the path. For example:

REQUEST: http://www.mysite.com/object/22392?backdoor=true
# Using the example above...
r.match("/object/:id").defer_to(:action => "index") do |request, params|
	# params ==> {:id => "22392", :backdoor => "true"}
	# ...
end


Regards,

Duane Johnson
(canadaduane)
http://blog.inquirylabs.com/

From canadaduane at gmail.com  Fri Aug 31 18:47:16 2007
From: canadaduane at gmail.com (Duane Johnson)
Date: Fri, 31 Aug 2007 16:47:16 -0600
Subject: Bootstrapping from SVN
In-Reply-To: <20070831205150.GA28101@uk.tiscali.com>
References: <20070831205150.GA28101@uk.tiscali.com>
Message-ID: <DA7FFEF1-27FA-4D01-BA5D-12C518955079@gmail.com>

The way I've seen it done is to run "rake install" in a checked out  
version of merb.  Then it just installs a new gem over top of  
whatever gem your system has.

Duane

On Aug 31, 2007, at 2:51 PM, Brian Candler wrote:

> Is there a simple way to bootstrap merb directly from the Subversion
> repository, rather than first installing it as a gem?
>
> I can do the following if I already have an old gem lying around:
>
>    merb -g testapp        # using an old gem
>    cd testapp
>    rake merb:freeze_from_svn
>
> Unfortunately, that means I'm using an old version of the  
> generator. So for
> example, when I just tried this, the generator made a conf/ directory,
> whereas the code in trunk expects config/.
>
> Equally, I can't test any new features recently added into the  
> generator
> (such as the snazzy new screens committed in r464-r466), without first
> building trunk into a gem, and gem installing it over my existing one.
>
> I have tried:
>
> svn co http://svn.devjavu.com/merb/trunk merb
> cd merb
> ruby -Ilib bin/merb -g /var/tmp/test123
>
> but I get:
>
> Couldn't find 'merb' generator
>   /usr/lib/ruby/gems/1.8/gems/rubigen-1.0.3/lib/rubigen/lookup.rb: 
> 159:in `lookup'
>   /usr/lib/ruby/gems/1.8/gems/rubigen-1.0.3/lib/rubigen/lookup.rb: 
> 164:in `instance'
>   /usr/lib/ruby/gems/1.8/gems/rubigen-1.0.3/lib/rubigen/scripts/../ 
> scripts.rb:30:in `run'
>   /v/build/merb-trunk/lib/merb/generators/merb_app/merb_app.rb: 
> 13:in `run'
>   ./lib/merb/server.rb:216:in `run'
>   bin/merb:6
>
> Cheers,
>
> Brian.
> _______________________________________________
> Merb-devel mailing list
> Merb-devel at rubyforge.org
> http://rubyforge.org/mailman/listinfo/merb-devel