From paul at pauldix.net  Wed Apr  9 09:30:58 2008
From: paul at pauldix.net (Paul Dix)
Date: Wed, 9 Apr 2008 09:30:58 -0400
Subject: [Cells-talk] putting a form in a cell
Message-ID: <c0a615890804090630m982a5acqb9d0af1f77ec1692@mail.gmail.com>

Hi all,
I'm playing around with cells and I'm trying to put a form in one.
It's dying on the forgery protection stuff. I suppose because there's
no inheritance chain from ActionController. Is putting a form in a
cell unsupported behavior?

Thanks,
Paul

From apotonick at gmail.com  Wed Apr  9 10:09:56 2008
From: apotonick at gmail.com (Nick Sutterer)
Date: Wed, 9 Apr 2008 16:09:56 +0200
Subject: [Cells-talk] putting a form in a cell
In-Reply-To: <c0a615890804090630m982a5acqb9d0af1f77ec1692@mail.gmail.com>
References: <c0a615890804090630m982a5acqb9d0af1f77ec1692@mail.gmail.com>
Message-ID: <e2cf7aef0804090709r166e5c9cwa73e6804907d6cf3@mail.gmail.com>

On Wed, Apr 9, 2008 at 3:30 PM, Paul Dix <paul at pauldix.net> wrote:
> Hi all,
>  I'm playing around with cells and I'm trying to put a form in one.
>  It's dying on the forgery protection stuff. I suppose because there's
>  no inheritance chain from ActionController. Is putting a form in a
>  cell unsupported behavior?
>
that's a known issue we're working on (in rails 2). putting a form in
a cell view is of course supported, since cells should move behaviour
from controllers to components.
sorry for that inconvinience. maybe i can take a look at this next week.

nick


>  Thanks,
>  Paul
>  _______________________________________________
>  Cells-talk mailing list
>  Cells-talk at rubyforge.org
>  http://rubyforge.org/mailman/listinfo/cells-talk
>

From mike.pence at gmail.com  Wed Apr  9 10:14:14 2008
From: mike.pence at gmail.com (Mike Pence)
Date: Wed, 9 Apr 2008 10:14:14 -0400
Subject: [Cells-talk] Let's put Cells on Github
Message-ID: <d8ee51740804090714j658945bemef622550d6ae7fb@mail.gmail.com>

Since git opens up a project significantly to more collaborators, I
would like to propose that we move Cells over to github.

From apotonick at gmail.com  Wed Apr  9 10:21:02 2008
From: apotonick at gmail.com (Nick Sutterer)
Date: Wed, 9 Apr 2008 16:21:02 +0200
Subject: [Cells-talk] Let's put Cells on Github
In-Reply-To: <d8ee51740804090714j658945bemef622550d6ae7fb@mail.gmail.com>
References: <d8ee51740804090714j658945bemef622550d6ae7fb@mail.gmail.com>
Message-ID: <e2cf7aef0804090721j232fe7b3l2ad712b487d015cf@mail.gmail.com>

mike- if you tell me git is good, it is good. we could still leave the
rubyforge project as is, but move svn, tracker, etc to git.

nick


On Wed, Apr 9, 2008 at 4:14 PM, Mike Pence <mike.pence at gmail.com> wrote:
> Since git opens up a project significantly to more collaborators, I
>  would like to propose that we move Cells over to github.
>  _______________________________________________
>  Cells-talk mailing list
>  Cells-talk at rubyforge.org
>  http://rubyforge.org/mailman/listinfo/cells-talk
>

From mike.pence at gmail.com  Wed Apr  9 10:27:25 2008
From: mike.pence at gmail.com (Mike Pence)
Date: Wed, 9 Apr 2008 10:27:25 -0400
Subject: [Cells-talk] Let's put Cells on Github
In-Reply-To: <e2cf7aef0804090721j232fe7b3l2ad712b487d015cf@mail.gmail.com>
References: <d8ee51740804090714j658945bemef622550d6ae7fb@mail.gmail.com>
	<e2cf7aef0804090721j232fe7b3l2ad712b487d015cf@mail.gmail.com>
Message-ID: <d8ee51740804090727k8b5b298g38e13288fc73847b@mail.gmail.com>

The sweet setup seems to be to use lighthouse for issue tracking,
github for file hosting and distributed version control. I will look
into it tonight, US Eastern time.

On Wed, Apr 9, 2008 at 10:21 AM, Nick Sutterer <apotonick at gmail.com> wrote:
> mike- if you tell me git is good, it is good. we could still leave the
>  rubyforge project as is, but move svn, tracker, etc to git.
>
>  nick
>
>
>
>  On Wed, Apr 9, 2008 at 4:14 PM, Mike Pence <mike.pence at gmail.com> wrote:
>  > Since git opens up a project significantly to more collaborators, I
>  >  would like to propose that we move Cells over to github.
>  >  _______________________________________________
>  >  Cells-talk mailing list
>  >  Cells-talk at rubyforge.org
>  >  http://rubyforge.org/mailman/listinfo/cells-talk
>  >
>

From paul at pauldix.net  Wed Apr  9 10:31:01 2008
From: paul at pauldix.net (Paul Dix)
Date: Wed, 9 Apr 2008 10:31:01 -0400
Subject: [Cells-talk] putting a form in a cell
In-Reply-To: <e2cf7aef0804090709r166e5c9cwa73e6804907d6cf3@mail.gmail.com>
References: <c0a615890804090630m982a5acqb9d0af1f77ec1692@mail.gmail.com>
	<e2cf7aef0804090709r166e5c9cwa73e6804907d6cf3@mail.gmail.com>
Message-ID: <c0a615890804090731h16119b05g75d500cd6a114b4b@mail.gmail.com>

So I was able to hack around it doing this in my cell controller:

  include ActionController::RequestForgeryProtection

  def allow_forgery_protection
    ActionController::Base.allow_forgery_protection
  end

  def request_forgery_protection_token
    form_authenticity_token
  end

Maybe that could be put into Cell::Base?

Cheers,
Paul


On Wed, Apr 9, 2008 at 10:09 AM, Nick Sutterer <apotonick at gmail.com> wrote:
> On Wed, Apr 9, 2008 at 3:30 PM, Paul Dix <paul at pauldix.net> wrote:
>
>
> > Hi all,
>  >  I'm playing around with cells and I'm trying to put a form in one.
>  >  It's dying on the forgery protection stuff. I suppose because there's
>  >  no inheritance chain from ActionController. Is putting a form in a
>  >  cell unsupported behavior?
>  >
>  that's a known issue we're working on (in rails 2). putting a form in
>  a cell view is of course supported, since cells should move behaviour
>  from controllers to components.
>  sorry for that inconvinience. maybe i can take a look at this next week.
>
>  nick
>
>
>
>
>  >  Thanks,
>  >  Paul
>  >  _______________________________________________
>  >  Cells-talk mailing list
>  >  Cells-talk at rubyforge.org
>  >  http://rubyforge.org/mailman/listinfo/cells-talk
>  >
>  _______________________________________________
>  Cells-talk mailing list
>  Cells-talk at rubyforge.org
>  http://rubyforge.org/mailman/listinfo/cells-talk
>

From mike.pence at gmail.com  Wed Apr  9 10:35:06 2008
From: mike.pence at gmail.com (Mike Pence)
Date: Wed, 9 Apr 2008 10:35:06 -0400
Subject: [Cells-talk] putting a form in a cell
In-Reply-To: <c0a615890804090731h16119b05g75d500cd6a114b4b@mail.gmail.com>
References: <c0a615890804090630m982a5acqb9d0af1f77ec1692@mail.gmail.com>
	<e2cf7aef0804090709r166e5c9cwa73e6804907d6cf3@mail.gmail.com>
	<c0a615890804090731h16119b05g75d500cd6a114b4b@mail.gmail.com>
Message-ID: <d8ee51740804090735t9560f6aod006f87d9df06ed1@mail.gmail.com>

You're the man, Paul.

On Wed, Apr 9, 2008 at 10:31 AM, Paul Dix <paul at pauldix.net> wrote:
> So I was able to hack around it doing this in my cell controller:
>
>   include ActionController::RequestForgeryProtection
>
>   def allow_forgery_protection
>     ActionController::Base.allow_forgery_protection
>   end
>
>   def request_forgery_protection_token
>     form_authenticity_token
>   end
>
>  Maybe that could be put into Cell::Base?
>
>  Cheers,
>  Paul
>
>
>
>
>  On Wed, Apr 9, 2008 at 10:09 AM, Nick Sutterer <apotonick at gmail.com> wrote:
>  > On Wed, Apr 9, 2008 at 3:30 PM, Paul Dix <paul at pauldix.net> wrote:
>  >
>  >
>  > > Hi all,
>  >  >  I'm playing around with cells and I'm trying to put a form in one.
>  >  >  It's dying on the forgery protection stuff. I suppose because there's
>  >  >  no inheritance chain from ActionController. Is putting a form in a
>  >  >  cell unsupported behavior?
>  >  >
>  >  that's a known issue we're working on (in rails 2). putting a form in
>  >  a cell view is of course supported, since cells should move behaviour
>  >  from controllers to components.
>  >  sorry for that inconvinience. maybe i can take a look at this next week.
>  >
>  >  nick
>  >
>  >
>  >
>  >
>  >  >  Thanks,
>  >  >  Paul
>  >  >  _______________________________________________
>  >  >  Cells-talk mailing list
>  >  >  Cells-talk at rubyforge.org
>  >  >  http://rubyforge.org/mailman/listinfo/cells-talk
>  >  >
>  >  _______________________________________________
>  >  Cells-talk mailing list
>  >  Cells-talk at rubyforge.org
>  >  http://rubyforge.org/mailman/listinfo/cells-talk
>  >
>  _______________________________________________
>  Cells-talk mailing list
>  Cells-talk at rubyforge.org
>  http://rubyforge.org/mailman/listinfo/cells-talk
>

From apotonick at gmail.com  Wed Apr  9 11:14:25 2008
From: apotonick at gmail.com (Nick Sutterer)
Date: Wed, 9 Apr 2008 17:14:25 +0200
Subject: [Cells-talk] putting a form in a cell
In-Reply-To: <d8ee51740804090735t9560f6aod006f87d9df06ed1@mail.gmail.com>
References: <c0a615890804090630m982a5acqb9d0af1f77ec1692@mail.gmail.com>
	<e2cf7aef0804090709r166e5c9cwa73e6804907d6cf3@mail.gmail.com>
	<c0a615890804090731h16119b05g75d500cd6a114b4b@mail.gmail.com>
	<d8ee51740804090735t9560f6aod006f87d9df06ed1@mail.gmail.com>
Message-ID: <e2cf7aef0804090814t4183c1afk38da5a6097464740@mail.gmail.com>

On Wed, Apr 9, 2008 at 4:35 PM, Mike Pence <mike.pence at gmail.com> wrote:
> You're the man, Paul.
>
>
you kick ass, thank you so much! i will create a test and patch it ASAP.

nick

>
>  On Wed, Apr 9, 2008 at 10:31 AM, Paul Dix <paul at pauldix.net> wrote:
>  > So I was able to hack around it doing this in my cell controller:
>  >
>  >   include ActionController::RequestForgeryProtection
>  >
>  >   def allow_forgery_protection
>  >     ActionController::Base.allow_forgery_protection
>  >   end
>  >
>  >   def request_forgery_protection_token
>  >     form_authenticity_token
>  >   end
>  >
>  >  Maybe that could be put into Cell::Base?
>  >
>  >  Cheers,
>  >  Paul
>  >
>  >
>  >
>  >
>  >  On Wed, Apr 9, 2008 at 10:09 AM, Nick Sutterer <apotonick at gmail.com> wrote:
>  >  > On Wed, Apr 9, 2008 at 3:30 PM, Paul Dix <paul at pauldix.net> wrote:
>  >  >
>  >  >
>  >  > > Hi all,
>  >  >  >  I'm playing around with cells and I'm trying to put a form in one.
>  >  >  >  It's dying on the forgery protection stuff. I suppose because there's
>  >  >  >  no inheritance chain from ActionController. Is putting a form in a
>  >  >  >  cell unsupported behavior?
>  >  >  >
>  >  >  that's a known issue we're working on (in rails 2). putting a form in
>  >  >  a cell view is of course supported, since cells should move behaviour
>  >  >  from controllers to components.
>  >  >  sorry for that inconvinience. maybe i can take a look at this next week.
>  >  >
>  >  >  nick
>  >  >
>  >  >
>  >  >
>  >  >
>  >  >  >  Thanks,
>  >  >  >  Paul
>  >  >  >  _______________________________________________
>  >  >  >  Cells-talk mailing list
>  >  >  >  Cells-talk at rubyforge.org
>  >  >  >  http://rubyforge.org/mailman/listinfo/cells-talk
>  >  >  >
>  >  >  _______________________________________________
>  >  >  Cells-talk mailing list
>  >  >  Cells-talk at rubyforge.org
>  >  >  http://rubyforge.org/mailman/listinfo/cells-talk
>  >  >
>  >  _______________________________________________
>  >  Cells-talk mailing list
>  >  Cells-talk at rubyforge.org
>  >  http://rubyforge.org/mailman/listinfo/cells-talk
>  >
>

From Peter.Bex at solide-ict.nl  Wed Apr  9 11:25:33 2008
From: Peter.Bex at solide-ict.nl (Peter Bex)
Date: Wed, 9 Apr 2008 17:25:33 +0200
Subject: [Cells-talk] Let's put Cells on Github
In-Reply-To: <d8ee51740804090727k8b5b298g38e13288fc73847b@mail.gmail.com>
References: <d8ee51740804090714j658945bemef622550d6ae7fb@mail.gmail.com>
	<e2cf7aef0804090721j232fe7b3l2ad712b487d015cf@mail.gmail.com>
	<d8ee51740804090727k8b5b298g38e13288fc73847b@mail.gmail.com>
Message-ID: <20080409152532.GD697@darko.solide-ict.nl>

On Wed, Apr 09, 2008 at 10:27:25AM -0400, Mike Pence wrote:
> The sweet setup seems to be to use lighthouse for issue tracking,
> github for file hosting and distributed version control. I will look
> into it tonight, US Eastern time.

Right, so we're leaving a perfectly good and proven hosting platform
(Rubyforge) for something else, because that somehow magically
"opens up a project significantly to more collaborators".
How exactly does it do that?

Who is going to pay for Lighthouse?  Their free service does not allow
you to make a project public, nor does it allow the project to grow
beyond 2 people.  How does this "open up a project significantly to
more collaborators"?

In short, "what's wrong with Rubyforge"?

Cheers,
Peter Bex
Solide ICT - http://www.solide-ict.nl

From Peter.Bex at solide-ict.nl  Wed Apr  9 11:25:57 2008
From: Peter.Bex at solide-ict.nl (Peter Bex)
Date: Wed, 9 Apr 2008 17:25:57 +0200
Subject: [Cells-talk] putting a form in a cell
In-Reply-To: <e2cf7aef0804090814t4183c1afk38da5a6097464740@mail.gmail.com>
References: <c0a615890804090630m982a5acqb9d0af1f77ec1692@mail.gmail.com>
	<e2cf7aef0804090709r166e5c9cwa73e6804907d6cf3@mail.gmail.com>
	<c0a615890804090731h16119b05g75d500cd6a114b4b@mail.gmail.com>
	<d8ee51740804090735t9560f6aod006f87d9df06ed1@mail.gmail.com>
	<e2cf7aef0804090814t4183c1afk38da5a6097464740@mail.gmail.com>
Message-ID: <20080409152557.GE697@darko.solide-ict.nl>

On Wed, Apr 09, 2008 at 05:14:25PM +0200, Nick Sutterer wrote:
> On Wed, Apr 9, 2008 at 4:35 PM, Mike Pence <mike.pence at gmail.com> wrote:
> > You're the man, Paul.
> >
> >
> you kick ass, thank you so much! i will create a test and patch it ASAP.

+1

This rocks!

Cheers,
Peter Bex
Solide ICT - http://www.solide-ict.nl

From mike.pence at gmail.com  Wed Apr  9 11:25:59 2008
From: mike.pence at gmail.com (Mike Pence)
Date: Wed, 9 Apr 2008 11:25:59 -0400
Subject: [Cells-talk] Let's put Cells on Github
In-Reply-To: <20080409152532.GD697@darko.solide-ict.nl>
References: <d8ee51740804090714j658945bemef622550d6ae7fb@mail.gmail.com>
	<e2cf7aef0804090721j232fe7b3l2ad712b487d015cf@mail.gmail.com>
	<d8ee51740804090727k8b5b298g38e13288fc73847b@mail.gmail.com>
	<20080409152532.GD697@darko.solide-ict.nl>
Message-ID: <d8ee51740804090825y7332125fnc19c413f5aa44eec@mail.gmail.com>

We can stay on RubyForge, and we can create a svn snapshot from the
git repository to keep that available to users.

The main advantage is git, which makes forking and merging code
trivial. Take some time and get to know git to understand the benefits
I am alluding to.

On Wed, Apr 9, 2008 at 11:25 AM, Peter Bex <Peter.Bex at solide-ict.nl> wrote:
> On Wed, Apr 09, 2008 at 10:27:25AM -0400, Mike Pence wrote:
>  > The sweet setup seems to be to use lighthouse for issue tracking,
>  > github for file hosting and distributed version control. I will look
>  > into it tonight, US Eastern time.
>
>  Right, so we're leaving a perfectly good and proven hosting platform
>  (Rubyforge) for something else, because that somehow magically
>  "opens up a project significantly to more collaborators".
>  How exactly does it do that?
>
>  Who is going to pay for Lighthouse?  Their free service does not allow
>  you to make a project public, nor does it allow the project to grow
>  beyond 2 people.  How does this "open up a project significantly to
>  more collaborators"?
>
>  In short, "what's wrong with Rubyforge"?
>
>  Cheers,
>  Peter Bex
>  Solide ICT - http://www.solide-ict.nl
>
>
> _______________________________________________
>  Cells-talk mailing list
>  Cells-talk at rubyforge.org
>  http://rubyforge.org/mailman/listinfo/cells-talk
>

From mike.pence at gmail.com  Wed Apr  9 11:30:30 2008
From: mike.pence at gmail.com (Mike Pence)
Date: Wed, 9 Apr 2008 11:30:30 -0400
Subject: [Cells-talk] Let's put Cells on Github
In-Reply-To: <d8ee51740804090825y7332125fnc19c413f5aa44eec@mail.gmail.com>
References: <d8ee51740804090714j658945bemef622550d6ae7fb@mail.gmail.com>
	<e2cf7aef0804090721j232fe7b3l2ad712b487d015cf@mail.gmail.com>
	<d8ee51740804090727k8b5b298g38e13288fc73847b@mail.gmail.com>
	<20080409152532.GD697@darko.solide-ict.nl>
	<d8ee51740804090825y7332125fnc19c413f5aa44eec@mail.gmail.com>
Message-ID: <d8ee51740804090830h60c8f547gad74189d9e4e7de5@mail.gmail.com>

http://git.or.cz/course/svn.html
http://railscasts.com/episodes/96
http://video.google.com/videoplay?docid=-3999952944619245780

From Peter.Bex at solide-ict.nl  Thu Apr 10 04:30:15 2008
From: Peter.Bex at solide-ict.nl (Peter Bex)
Date: Thu, 10 Apr 2008 10:30:15 +0200
Subject: [Cells-talk] Let's put Cells on Github
In-Reply-To: <d8ee51740804090825y7332125fnc19c413f5aa44eec@mail.gmail.com>
References: <d8ee51740804090714j658945bemef622550d6ae7fb@mail.gmail.com>
	<e2cf7aef0804090721j232fe7b3l2ad712b487d015cf@mail.gmail.com>
	<d8ee51740804090727k8b5b298g38e13288fc73847b@mail.gmail.com>
	<20080409152532.GD697@darko.solide-ict.nl>
	<d8ee51740804090825y7332125fnc19c413f5aa44eec@mail.gmail.com>
Message-ID: <20080410083015.GA482@darko.solide-ict.nl>

On Wed, Apr 09, 2008 at 11:25:59AM -0400, Mike Pence wrote:
> We can stay on RubyForge, and we can create a svn snapshot from the
> git repository to keep that available to users.
> 
> The main advantage is git, which makes forking and merging code
> trivial. Take some time and get to know git to understand the benefits
> I am alluding to.

*sighs*
I've seen my share of version control flamewars by now, and I've seen
most arguments already.  I understand and for the most part agree on the
advantages of a distributed versioning system versus a centralised one,
so I'm not going to argue with you about that.

In your post it looks like you're pretty much equating Git with DSCM,
which is a mistake.  Git is just *a* DSCM, not *the* DSCM.  It's not even
a great DSCM, whatever the current hype wants you to believe.  

Here are a couple of reasons not to use Git:

1. Interface

The number one complaint I keep hearing when talking to people about
git is the fact that its CLI is pretty horrible.  It installs a boatload
of irritating little scripts that are hard to learn.  Other DSCMs are
much less intrusive.

Current situation is better because svn, even though it's centralised,
*stays out of your way*.  It's just a simple tool that stays on the
background and just lets you do your work (in the end, we're working on
software, the goal is not to use the greatest SCM around)

Besides this, git does not seem to be as well-supported by external
tools as subversion.

2. Concepts

Git gets two things majorly wrong:
- It uses the CVS concept of tagging.  svn's approach to tags ("just make
   a directory and let convention decide it's a tag") was very refreshing
   and really opened my eyes as to how weird and unintuitive the CVS way
   of tagging is.
- It doesn't clean up garbage.  It has this braindead approach to storing
   data that in aborted transactions and a few other situations it
   creates a node in its version tree that does not get associated with any
   other node.  This is garbage, which you have to periodically clean up
   using a special command.  Other DSCMs do not have this design mistake.

3. Portability

Apparently Git doesn't function properly under Windows.  I've seen numerous
people complain about this particular thing when Rails announced their
switch to Git.  I couldn't care less about Windows, but the reason you
stated for switching to Git was to make our project more accessible to
other people.  If anything, it puts up a big barrier for people who are
still on Windows.

4. Code

Svn, darcs, mercurial, Bazaar, etc etc are all written in one language.
This makes it easier to understand and hack if something needs to be
customized.  Git just a random bunch of C programs, shell scripts, Perl
scripts and tcl scripts.

5. Hype

I strongly get the impression that the reason you're suggesting git is
that Rails core decided to go use it.  This is one of the biggest problems
in the IT industry.  Something new and shiny comes along and people
decide to switch to it en masse.  Then 6 months down the road something
newer, shinier comes along and it's time for a major overhaul yet again
because "oh yes, this hot new thing is so much better than that old thing".

Hype is a bad reason to dismiss a piece of software, but it's an even worse
reason to decide to use it.  Software should be chosen on merit, not on
how hot it is.  For example, other projects have done their research on
DSCMs.  You might want to look at their results before jumping on the git
bandwagon:

http://live.gnome.org/DistributedSCM
http://www.opensolaris.org/os/community/tools/scm
http://adiumx.com/pipermail/adium-devl_adiumx.com/2008-February/004542.html
http://www.mail-archive.com/classpath at gnu.org/msg14357.html
http://wiki.freebsd.org/VersionControl
http://www.ruby-forum.com/topic/87174

If anything, let's sit this out and see how well it works for Rails.
If we can watch them succeed or fail from the side line, we can make a
more informed decision about Git and github in particular.

About github itself:

Currently we are using all services from one place and one place only:
Rubyforge.  If we want to use Github, we'd switch over a couple of things.
Github seems to support just git hosting, which includes a repos viewer,
and it has a wiki (which is a good thing).  For other things you're
supposed to use Lighthouse (not sure what that offers), but that's a paid
service.  Rubyforge offers us a central spot for mailing lists, svn (with
web viewer), ticket tracking, file release management, the Trove software
map (which makes the project more findable) and web space (which isn't
too great because it only allows static HTML).

You can argue that we don't have to leave Rubyforge to get Git support
from github, but that would imply that we're scattering around the places
our project lives at.  That's confusing.

So right now I see no compelling reason to switch to another hoster.
Subversion is not broken, so why fix it? When Rubyforge starts
supporting a DSCM (hopefully not git!), I'm certainly in for re-evaluation.

In closing, I'm in favor of turning around your proposal.  Instead of
instantly switching to Git and dropping svn as main repos, we can
have a git mirror of the official svn repos.  If it turns out that git
works great and svn is dragging us down, we can still make the switch.

Cheers,
Peter Bex
Solide ICT - http://www.solide-ict.nl

From mike.pence at gmail.com  Thu Apr 10 09:20:11 2008
From: mike.pence at gmail.com (Mike Pence)
Date: Thu, 10 Apr 2008 09:20:11 -0400
Subject: [Cells-talk] Let's put Cells on Github
In-Reply-To: <20080410083015.GA482@darko.solide-ict.nl>
References: <d8ee51740804090714j658945bemef622550d6ae7fb@mail.gmail.com>
	<e2cf7aef0804090721j232fe7b3l2ad712b487d015cf@mail.gmail.com>
	<d8ee51740804090727k8b5b298g38e13288fc73847b@mail.gmail.com>
	<20080409152532.GD697@darko.solide-ict.nl>
	<d8ee51740804090825y7332125fnc19c413f5aa44eec@mail.gmail.com>
	<20080410083015.GA482@darko.solide-ict.nl>
Message-ID: <d8ee51740804100620v2c2e7189l5640dd2a9a5d6e1f@mail.gmail.com>

Peter,

That sounds like a reasonable solution. Your speculation about my
thinking process is way off, but the conclusion works for me.

I have worked with darcs on some projects and with svn since it first
came out -- CVS prior to that. I have also worked with several other
commercial VCS's at major US corporations. Git is new, so the tools
will grow with it in time, my main motivation is just to make it
trivial to fork and merge with Cells so that others can contribute and
keep their own parallel changes if their changes are not appropriate
for the main Cells code-base.

Yeah, all the cool kids are using it, but I know those guys well -- I
hang out in #caboose with them (and have for a couple of years) and
have met many of them personally. I respect their judgment and
understand their motivations.

I understand that the groupthink of the Rails community can be wrong
-- that is why I back Cells. At 41, I am not prone to follow whims of
fashion. But I think that Cells would benefit, as Rails will benefit,
from a widened contributor base, and git is a great tool for that.

Best,
Mike Pence

On Thu, Apr 10, 2008 at 4:30 AM, Peter Bex <Peter.Bex at solide-ict.nl> wrote:
> On Wed, Apr 09, 2008 at 11:25:59AM -0400, Mike Pence wrote:
>  > We can stay on RubyForge, and we can create a svn snapshot from the
>  > git repository to keep that available to users.
>  >
>  > The main advantage is git, which makes forking and merging code
>  > trivial. Take some time and get to know git to understand the benefits
>  > I am alluding to.
>
>  *sighs*
>  I've seen my share of version control flamewars by now, and I've seen
>  most arguments already.  I understand and for the most part agree on the
>  advantages of a distributed versioning system versus a centralised one,
>  so I'm not going to argue with you about that.
>
>  In your post it looks like you're pretty much equating Git with DSCM,
>  which is a mistake.  Git is just *a* DSCM, not *the* DSCM.  It's not even
>  a great DSCM, whatever the current hype wants you to believe.
>
>  Here are a couple of reasons not to use Git:
>
>  1. Interface
>
>  The number one complaint I keep hearing when talking to people about
>  git is the fact that its CLI is pretty horrible.  It installs a boatload
>  of irritating little scripts that are hard to learn.  Other DSCMs are
>  much less intrusive.
>
>  Current situation is better because svn, even though it's centralised,
>  *stays out of your way*.  It's just a simple tool that stays on the
>  background and just lets you do your work (in the end, we're working on
>  software, the goal is not to use the greatest SCM around)
>
>  Besides this, git does not seem to be as well-supported by external
>  tools as subversion.
>
>  2. Concepts
>
>  Git gets two things majorly wrong:
>  - It uses the CVS concept of tagging.  svn's approach to tags ("just make
>    a directory and let convention decide it's a tag") was very refreshing
>    and really opened my eyes as to how weird and unintuitive the CVS way
>    of tagging is.
>  - It doesn't clean up garbage.  It has this braindead approach to storing
>    data that in aborted transactions and a few other situations it
>    creates a node in its version tree that does not get associated with any
>    other node.  This is garbage, which you have to periodically clean up
>    using a special command.  Other DSCMs do not have this design mistake.
>
>  3. Portability
>
>  Apparently Git doesn't function properly under Windows.  I've seen numerous
>  people complain about this particular thing when Rails announced their
>  switch to Git.  I couldn't care less about Windows, but the reason you
>  stated for switching to Git was to make our project more accessible to
>  other people.  If anything, it puts up a big barrier for people who are
>  still on Windows.
>
>  4. Code
>
>  Svn, darcs, mercurial, Bazaar, etc etc are all written in one language.
>  This makes it easier to understand and hack if something needs to be
>  customized.  Git just a random bunch of C programs, shell scripts, Perl
>  scripts and tcl scripts.
>
>  5. Hype
>
>  I strongly get the impression that the reason you're suggesting git is
>  that Rails core decided to go use it.  This is one of the biggest problems
>  in the IT industry.  Something new and shiny comes along and people
>  decide to switch to it en masse.  Then 6 months down the road something
>  newer, shinier comes along and it's time for a major overhaul yet again
>  because "oh yes, this hot new thing is so much better than that old thing".
>
>  Hype is a bad reason to dismiss a piece of software, but it's an even worse
>  reason to decide to use it.  Software should be chosen on merit, not on
>  how hot it is.  For example, other projects have done their research on
>  DSCMs.  You might want to look at their results before jumping on the git
>  bandwagon:
>
>  http://live.gnome.org/DistributedSCM
>  http://www.opensolaris.org/os/community/tools/scm
>  http://adiumx.com/pipermail/adium-devl_adiumx.com/2008-February/004542.html
>  http://www.mail-archive.com/classpath at gnu.org/msg14357.html
>  http://wiki.freebsd.org/VersionControl
>  http://www.ruby-forum.com/topic/87174
>
>  If anything, let's sit this out and see how well it works for Rails.
>  If we can watch them succeed or fail from the side line, we can make a
>  more informed decision about Git and github in particular.
>
>  About github itself:
>
>  Currently we are using all services from one place and one place only:
>  Rubyforge.  If we want to use Github, we'd switch over a couple of things.
>  Github seems to support just git hosting, which includes a repos viewer,
>  and it has a wiki (which is a good thing).  For other things you're
>  supposed to use Lighthouse (not sure what that offers), but that's a paid
>  service.  Rubyforge offers us a central spot for mailing lists, svn (with
>  web viewer), ticket tracking, file release management, the Trove software
>  map (which makes the project more findable) and web space (which isn't
>  too great because it only allows static HTML).
>
>  You can argue that we don't have to leave Rubyforge to get Git support
>  from github, but that would imply that we're scattering around the places
>  our project lives at.  That's confusing.
>
>  So right now I see no compelling reason to switch to another hoster.
>  Subversion is not broken, so why fix it? When Rubyforge starts
>  supporting a DSCM (hopefully not git!), I'm certainly in for re-evaluation.
>
>  In closing, I'm in favor of turning around your proposal.  Instead of
>  instantly switching to Git and dropping svn as main repos, we can
>  have a git mirror of the official svn repos.  If it turns out that git
>  works great and svn is dragging us down, we can still make the switch.
>
>
>
>  Cheers,
>  Peter Bex
>  Solide ICT - http://www.solide-ict.nl
>  _______________________________________________
>  Cells-talk mailing list
>  Cells-talk at rubyforge.org
>  http://rubyforge.org/mailman/listinfo/cells-talk
>

From mike.pence at gmail.com  Thu Apr 10 12:11:35 2008
From: mike.pence at gmail.com (Mike Pence)
Date: Thu, 10 Apr 2008 12:11:35 -0400
Subject: [Cells-talk] Cells and RJS
Message-ID: <d8ee51740804100911u6cdc1134ma50cb3b8a6d2c7f1@mail.gmail.com>

So, I am still trying to figure out the best way to integrate Cells
and Ext.js. I am using them together, but more with Ext as a richer
markup for fancier components, rendering to and using content from
divs, not really using them together as richly as I could.

What is the best strategy, I wonder, for integrating Cells and RJS? I
have designated a controller as my Cells proxy dispatcher -- the
controller brokers requests to my cells, etc. But a cell could return
JavaScript in addition to markup and CSS. How could a Cell return both
JS and markup, with the JS being immediately evaluated, like one would
with "page << script" via RJS?

Best,
Mike

From Peter.Bex at solide-ict.nl  Fri Apr 11 03:11:41 2008
From: Peter.Bex at solide-ict.nl (Peter Bex)
Date: Fri, 11 Apr 2008 09:11:41 +0200
Subject: [Cells-talk] Cells and RJS
In-Reply-To: <d8ee51740804100911u6cdc1134ma50cb3b8a6d2c7f1@mail.gmail.com>
References: <d8ee51740804100911u6cdc1134ma50cb3b8a6d2c7f1@mail.gmail.com>
Message-ID: <20080411071141.GA1073@darko.solide-ict.nl>

On Thu, Apr 10, 2008 at 12:11:35PM -0400, Mike Pence wrote:
> What is the best strategy, I wonder, for integrating Cells and RJS? I
> have designated a controller as my Cells proxy dispatcher -- the
> controller brokers requests to my cells, etc. But a cell could return
> JavaScript in addition to markup and CSS. How could a Cell return both
> JS and markup, with the JS being immediately evaluated, like one would
> with "page << script" via RJS?

I'm not sure what's fundamentally different from regular RJS here.
Normally, you'd just do something like

page.update 'some-element', render_cell :foo, :mystate
page << 'bar();'

If you're simply returning HTML, you can just put some JS in a <script> tag.

Maybe I'm misunderstanding what you're asking for here?

Cheers,
Peter Bex
Solide ICT - http://www.solide-ict.nl

From mike.pence at gmail.com  Fri Apr 11 10:07:48 2008
From: mike.pence at gmail.com (Mike Pence)
Date: Fri, 11 Apr 2008 10:07:48 -0400
Subject: [Cells-talk] Cells and RJS
In-Reply-To: <20080411071141.GA1073@darko.solide-ict.nl>
References: <d8ee51740804100911u6cdc1134ma50cb3b8a6d2c7f1@mail.gmail.com>
	<20080411071141.GA1073@darko.solide-ict.nl>
Message-ID: <d8ee51740804110707r3ff4fd66o1debb0ce8c4815d5@mail.gmail.com>

Wrapping JS in a script tag turns out to be the best solution for me
to dynamically create new Ext objects.

On Fri, Apr 11, 2008 at 3:11 AM, Peter Bex <Peter.Bex at solide-ict.nl> wrote:
> On Thu, Apr 10, 2008 at 12:11:35PM -0400, Mike Pence wrote:
>  > What is the best strategy, I wonder, for integrating Cells and RJS? I
>  > have designated a controller as my Cells proxy dispatcher -- the
>  > controller brokers requests to my cells, etc. But a cell could return
>  > JavaScript in addition to markup and CSS. How could a Cell return both
>  > JS and markup, with the JS being immediately evaluated, like one would
>  > with "page << script" via RJS?
>
>  I'm not sure what's fundamentally different from regular RJS here.
>  Normally, you'd just do something like
>
>  page.update 'some-element', render_cell :foo, :mystate
>  page << 'bar();'
>
>  If you're simply returning HTML, you can just put some JS in a <script> tag.
>
>  Maybe I'm misunderstanding what you're asking for here?
>
>  Cheers,
>  Peter Bex
>  Solide ICT - http://www.solide-ict.nl
>
>
> _______________________________________________
>  Cells-talk mailing list
>  Cells-talk at rubyforge.org
>  http://rubyforge.org/mailman/listinfo/cells-talk
>

From wthomas989 at yahoo.com  Mon Apr 14 16:48:55 2008
From: wthomas989 at yahoo.com (William Thomas)
Date: Mon, 14 Apr 2008 13:48:55 -0700 (PDT)
Subject: [Cells-talk] Ajax in cells?
Message-ID: <512542.47401.qm@web58708.mail.re1.yahoo.com>

An HTML attachment was scrubbed...
URL: http://rubyforge.org/pipermail/cells-talk/attachments/20080414/4b8d8ed4/attachment-0001.html 

From paul at pauldix.net  Mon Apr 14 17:02:39 2008
From: paul at pauldix.net (Paul Dix)
Date: Mon, 14 Apr 2008 17:02:39 -0400
Subject: [Cells-talk] Ajax in cells?
In-Reply-To: <512542.47401.qm@web58708.mail.re1.yahoo.com>
References: <512542.47401.qm@web58708.mail.re1.yahoo.com>
Message-ID: <c0a615890804141402n2c0b9aebn8e3f32f6af939057@mail.gmail.com>

You need to have the action that handles the ajax in a regular
controller. A Cell controller bypasses the whole routing system (which
is what makes it fast) so the web server doesn't know how to route
your ajax request.

In my case, I made a cell for a generic comment system and the form
was for a new comment. So I can drop this anywhere in my app. I made a
simple comment controller that handled posts.

Not sure if that's the intended use, but I'm still playing around with
the whole thing...

Paul

On Mon, Apr 14, 2008 at 4:48 PM, William Thomas <wthomas989 at yahoo.com> wrote:
> I want to start by saying thanks for the Cells plugin, it has come in handy
> on my application a few times now.
>
> I'm trying to write a generic city/state selection form that can be re-used
> in on my web app. I did some research and found out about the whole forgery
> protection issue, which I have got around using the proposed solution posted
> earlier " putting a form in a cell". Anyways- I'm not quite sure how the
> callback should live, I get errors putting it into the cell. I didn't really
> expect it to work there, but I thought it was worth a shot.
>
> How am I supposed to handle ajax callbacks - are they even possible?
>
> city_state.html.erb:
>
> <%=  javascript_include_tag 'prototype' %>
>
> <label>State:</label><br/>
> <%= collection_select( @model, 'state_id', @states , :id, :short,{ :prompt
> => true } )%>
>
> <%= observe_field "#{@model}_state_id",
> :url => { :action => "state_update_handler" } ,
> :with => "'#{@model}_state_id='+ escape(value)",
> :update => "city" %>
>
> <label>City</label><br />
>
> location_cell.rb:
> class LocationCell < Cell::Base
>
>   include ActionController::RequestForgeryProtection
>
>   def allow_forgery_protection
>     ActionController::Base.allow_forgery_protection
>   end
>
>   def request_forgery_protection_token
>     form_authenticity_token
>   end
>
>   def city_state
>     @model = @opts[:model]
>     @states = State.find(:all)
>   end
>
>   def state_update_handler
>     @cities = City.find_all_by_state_id(params[:state_id])
>   end
>
> end
>
>
>
> _______________________________________________
>  Cells-talk mailing list
>  Cells-talk at rubyforge.org
>  http://rubyforge.org/mailman/listinfo/cells-talk
>
>

From apotonick at gmail.com  Tue Apr 15 06:42:09 2008
From: apotonick at gmail.com (Nick Sutterer)
Date: Tue, 15 Apr 2008 12:42:09 +0200
Subject: [Cells-talk] Fwd:  Ajax in cells?
In-Reply-To: <e2cf7aef0804150340w11a0a3ffmd9a277c7fe9d31b8@mail.gmail.com>
References: <512542.47401.qm@web58708.mail.re1.yahoo.com>
	<c0a615890804141402n2c0b9aebn8e3f32f6af939057@mail.gmail.com>
	<e2cf7aef0804150340w11a0a3ffmd9a277c7fe9d31b8@mail.gmail.com>
Message-ID: <e2cf7aef0804150342g5853e741ib652a8be059709ff@mail.gmail.com>

whoops... here you go.


---------- Forwarded message ----------
From: Nick Sutterer <apotonick at gmail.com>
Date: Tue, Apr 15, 2008 at 12:40 PM
Subject: Re: [Cells-talk] Ajax in cells?
To: Paul Dix <paul at pauldix.net>


hey guys,


 On Mon, Apr 14, 2008 at 11:02 PM, Paul Dix <paul at pauldix.net> wrote:
 > You need to have the action that handles the ajax in a regular
 >  controller. A Cell controller bypasses the whole routing system (which
 >  is what makes it fast) so the web server doesn't know how to route
 >  your ajax request.
 >
 send the ajax request to a dedicated cells controller. this controller
 can route the request to the respective cell (by calling #render_cell
 on this cell and state) and then pass back the content.

 what about that?

 cheers,

 nick

From vossberg at exedra.de  Mon Apr 21 08:29:58 2008
From: vossberg at exedra.de (Lucas Vossberg)
Date: Mon, 21 Apr 2008 14:29:58 +0200
Subject: [Cells-talk] Cells, Ajax
Message-ID: <7.0.1.0.2.20080421142956.03bd11e8@exedra.de>

hi

I read the posts on this list about Cells and Ajax. but I'm still not able
to find a solution for my problem. so any help with this is appreciated.

I'm updating two DIVs with a single Ajax call to "update_info_blocks":

class PlayersController < ApplicationController
def update_info_blocks
   render :update do |page|
      page.replace_html 'block1', render_cell(:info, :block1, ...)
      page.replace_html 'block2', render_cell(:info, :block2, ...)
   end
end

how can I move this into the cell mini-controller? I'ld like to have

class PlayersController < ApplicationController
def update_info_blocks
   render_cell(:info, :blocks, ...)
end

and let the cell decide which blocks to update.
I have the feeling this can not work for some obvious reason?

thank you for your feedback
Lucas




From vossberg at exedra.de  Mon Apr 21 08:27:27 2008
From: vossberg at exedra.de (Lucas Vossberg)
Date: Mon, 21 Apr 2008 14:27:27 +0200
Subject: [Cells-talk] Cells, Ajax
Message-ID: <7.0.1.0.2.20080421140340.03c80a78@exedra.de>

hi

I read the posts on this list about Cells and Ajax. but I'm still not able
to find a solution for my problem. so any help with this is appreciated.

I'm updating two DIVs with a single Ajax call to "update_info_blocks":

class PlayersController < ApplicationController
def update_info_blocks
   render :update do |page|
      page.replace_html 'block1', render_cell(:info, :block1, ...)
      page.replace_html 'block2', render_cell(:info, :block2, ...)
   end
end

how can I move this into the cell mini-controller? I'ld like to have

class PlayersController < ApplicationController
def update_info_blocks
   render_cell(:info, :blocks, ...)
end

and let the cell decide which blocks to update.
I have the feeling this can not work for some obvious reason?

thank you for your feedback
Lucas




From vossberg at exedra.de  Mon Apr 21 08:35:49 2008
From: vossberg at exedra.de (Lucas Vossberg)
Date: Mon, 21 Apr 2008 14:35:49 +0200
Subject: [Cells-talk] Cells, Ajax
Message-ID: <7.0.1.0.2.20080421142956.03bd11e8@exedra.de>

hi

I read the posts on this list about Cells and Ajax. but I'm still not able
to find a solution for my problem. so any help with this is appreciated.

I'm updating two DIVs with a single Ajax call to "update_info_blocks":

class PlayersController < ApplicationController
def update_info_blocks
   render :update do |page|
      page.replace_html 'block1', render_cell(:info, :block1, ...)
      page.replace_html 'block2', render_cell(:info, :block2, ...)
   end
end

how can I move this into the cell mini-controller? I'ld like to have

class PlayersController < ApplicationController
def update_info_blocks
   render_cell(:info, :blocks, ...)
end

and let the cell decide which blocks to update.
I have the feeling this can not work for some obvious reason?

thank you for your feedback
Lucas




From vossberg at exedra.de  Mon Apr 21 16:57:51 2008
From: vossberg at exedra.de (Lucas Vossberg)
Date: Mon, 21 Apr 2008 22:57:51 +0200
Subject: [Cells-talk] Cells, Ajax
Message-ID: <7.0.1.0.2.20080421223835.03c80a78@exedra.de>

sorry for the double post above. this is a reply to my original question.

to be able to update two DIVs from a Cell-method as a response to an Ajax call
I'm now doing the following (hack?):

class PlayersController < ApplicationController
def replace_div(elems)
     render :update do |page|
       elems.each { |key, value| page.replace_html key, value }
     end
end

def update_info_blocks
   render_cell_to_string(:info, :blocks, ... )
end
end

class InfoCell < Cell::Base
def blocks
   case @opts[...]
   when 1
    @controller.replace_div( { 'block1' => 
render_view_for_state(:block1view) }
   when 2
    @controller.replace_div( { 'block2' => 
render_view_for_state(:block2view) }
   when 3
    @controller.replace_div( { 'block1' => 
render_view_for_state(:block1view), 'block2' => 
render_view_for_state(:block2view) }
   end
end

end


I believe it is not good to call the controller from the cell. but at 
least it allows me to move
a lot of code from the main controller to the cell classes. where 
everything can be arranged
much for flexible due to the nice features of the Cells.

or is there a better way of doing this?

again, thank you very much for your feedback and helping me to understand
Rails and Cells.

Lucas




>Date: Mon, 21 Apr 2008 14:35:49 +0200
>To: cells-talk at rubyforge.org
>From: Lucas Vossberg <vossberg at exedra.de>
>Subject: Cells, Ajax
>
>hi
>
>I read the posts on this list about Cells and Ajax. but I'm still not able
>to find a solution for my problem. so any help with this is appreciated.
>
>I'm updating two DIVs with a single Ajax call to "update_info_blocks":
>
>class PlayersController < ApplicationController
>def update_info_blocks
>   render :update do |page|
>      page.replace_html 'block1', render_cell(:info, :block1, ...)
>      page.replace_html 'block2', render_cell(:info, :block2, ...)
>   end
>end
>
>how can I move this into the cell mini-controller? I'ld like to have
>
>class PlayersController < ApplicationController
>def update_info_blocks
>   render_cell(:info, :blocks, ...)
>end
>
>and let the cell decide which blocks to update.
>I have the feeling this can not work for some obvious reason?
>
>thank you for your feedback
>Lucas
>