From transfire at gmail.com  Mon Sep  3 12:28:07 2007
From: transfire at gmail.com (Trans)
Date: Mon, 03 Sep 2007 16:28:07 -0000
Subject: [Facets] Two vs. one mailing list?
Message-ID: <1188836887.938215.62800@w3g2000hsg.googlegroups.com>

Hi--

Just wondering what arguments others might have for or against having
two mailing lists (one for development and the other for end-usage)
vs. having just this one list.

In particular, has the number of ticket tracking posts on this list
bother anyone? Or do you deem it actually a good thing, as an end-
user, to have a closer tie to th pulse of development?

Thanks,
T.


From o.renaud at laposte.net  Mon Sep  3 15:50:27 2007
From: o.renaud at laposte.net (Olivier Renaud)
Date: Mon, 3 Sep 2007 21:50:27 +0200
Subject: [Facets] Two vs. one mailing list?
In-Reply-To: <1188836887.938215.62800@w3g2000hsg.googlegroups.com>
References: <1188836887.938215.62800@w3g2000hsg.googlegroups.com>
Message-ID: <200709032150.27572.o.renaud@laposte.net>

Le lundi 03 septembre 2007 18:28, Trans a ?crit?:
> Hi--
>
> Just wondering what arguments others might have for or against having
> two mailing lists (one for development and the other for end-usage)
> vs. having just this one list.
>
> In particular, has the number of ticket tracking posts on this list
> bother anyone? Or do you deem it actually a good thing, as an end-
> user, to have a closer tie to th pulse of development?
>
> Thanks,
> T.

Hi,

I think the traffic of this list is low enough for not being split into two 
separate lists. The tickets posts are not a problem for me.

-- 
Olivier Renaud


From transfire at gmail.com  Wed Sep  5 21:05:51 2007
From: transfire at gmail.com (Trans)
Date: Thu, 06 Sep 2007 01:05:51 -0000
Subject: [Facets] 2.w00t!!!
Message-ID: <1189040751.101556.258970@19g2000hsx.googlegroups.com>

15:54 trans facets/trunk]% task/test
Loaded suite task/test
Started
.........................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................
Finished in 6.790911 seconds.

714 tests, 1785 assertions, 0 failures, 0 errors


From calamitates at gmail.com  Thu Sep  6 03:56:29 2007
From: calamitates at gmail.com (Calamitas)
Date: Thu, 6 Sep 2007 09:56:29 +0200
Subject: [Facets] 2.w00t!!!
In-Reply-To: <1189040751.101556.258970@19g2000hsx.googlegroups.com>
References: <1189040751.101556.258970@19g2000hsx.googlegroups.com>
Message-ID: <c4a653370709060056h7adbd4c4h54c4e656f4222ad0@mail.gmail.com>

Congratulations, Tom. You made it! We should celebrate! Well, next
time I get to the US that is...

Peter

PS: the link to the core API on http://facets.rubyforge.org/learn.html
is broken.

From transfire at gmail.com  Thu Sep  6 08:17:33 2007
From: transfire at gmail.com (TRANS)
Date: Thu, 6 Sep 2007 08:17:33 -0400
Subject: [Facets] 2.w00t!!!
In-Reply-To: <c4a653370709060056h7adbd4c4h54c4e656f4222ad0@mail.gmail.com>
References: <1189040751.101556.258970@19g2000hsx.googlegroups.com>
	<c4a653370709060056h7adbd4c4h54c4e656f4222ad0@mail.gmail.com>
Message-ID: <4b6f054f0709060517r4d994a67sf5b5e21352651cd0@mail.gmail.com>

On 9/6/07, Calamitas <calamitates at gmail.com> wrote:
> Congratulations, Tom. You made it! We should celebrate! Well, next
> time I get to the US that is...

Hey, I was just about to write you this morning -- you beat me to it! :)

First I wanted to apologize for not being in touch more -- I have been
quite busy trying to get all this Ruby code ready. Now that Facets 2.0
is nearly out, I will soon finally be ready to transition away from
Ruby --and to "Prime/Fly".

BTW, did you see Matz post on ruby-talk:

  http://groups.google.com/group/ruby-talk-google/browse_thread/thread/488b3e196a15c80b/c6ecc9647e1d6d08?hl=en#c6ecc9647e1d6d08

Takes me back to our first conversations on AOP. How odd(?), after all
this time, he's coming around to some of those original ideas.

On another front. I finally got around to creating a test case for
units.rb. Unfortunately I have to report that much of it is broken :(
I'd like to get Stick released sooner rather than later, so I was
hoping maybe you could find a little time to have a look. (And if so
maybe you could split off the currency into a separate require too?)
If you're too busy I understand. But if you can, it would a big time
saver for me... and well, how about a six-pack of your favorite
beverage? ;)

Also, while I've all but written-off the RCR process, I'm going to
take our Cuts RCR, update it with some new undertstanding I've gained
in my attempts to implement it (Facets 2.0 has a pretty good
implementation btw --much better than any of those I made before). My
"new" understanding is to return to the "old" roots, where by a Cut
is, pure and simple, a transparent subclass --no target or join
language. On top of that simple construct we build an Aspect class
that has the special language. A little further out, maybe generalize
it beyond Ruby and make a little paper out of it. Are there C.S.
journals that it could be submitted to I wonder?

> PS: the link to the core API on http://facets.rubyforge.org/learn.html
> is broken.

Ah, thanks.

So how is your new job? Are you really working on AOP for Ruby?

T.

From transfire at gmail.com  Thu Sep  6 09:57:18 2007
From: transfire at gmail.com (Trans)
Date: Thu, 06 Sep 2007 13:57:18 -0000
Subject: [Facets] 2.w00t!!!
In-Reply-To: <c4a653370709060056h7adbd4c4h54c4e656f4222ad0@mail.gmail.com>
References: <1189040751.101556.258970@19g2000hsx.googlegroups.com>
	<c4a653370709060056h7adbd4c4h54c4e656f4222ad0@mail.gmail.com>
Message-ID: <1189087038.666980.94010@k79g2000hse.googlegroups.com>

Ooops. Sorry all, my last post was meant to go to Peter directly. No
big deal, but just so you realize.

Thanks,
T.


From noreply at rubyforge.org  Fri Sep  7 07:01:56 2007
From: noreply at rubyforge.org (noreply at rubyforge.org)
Date: Fri,  7 Sep 2007 07:01:56 -0400 (EDT)
Subject: [Facets] [ facets-Patches-13750 ] Factorial Refactoring
Message-ID: <20070907110156.62DF55240A39@rubyforge.org>

Patches item #13750, was opened at 2007-09-07 06:01
You can respond by visiting: 
http://rubyforge.org/tracker/?func=detail&atid=3171&aid=13750&group_id=804

Category: None
Group: None
Status: Open
Resolution: None
Priority: 3
Submitted By: Drew Olson (dfg59)
Assigned to: Nobody (None)
Summary: Factorial Refactoring

Initial Comment:
Refactoring of the factorial code using inject.

----------------------------------------------------------------------

You can respond by visiting: 
http://rubyforge.org/tracker/?func=detail&atid=3171&aid=13750&group_id=804

From transfire at gmail.com  Fri Sep  7 07:52:59 2007
From: transfire at gmail.com (Trans)
Date: Fri, 07 Sep 2007 11:52:59 -0000
Subject: [Facets] Two vs. one mailing list?
In-Reply-To: <200709032150.27572.o.renaud@laposte.net>
References: <1188836887.938215.62800@w3g2000hsg.googlegroups.com>
	<200709032150.27572.o.renaud@laposte.net>
Message-ID: <1189165979.413869.252860@57g2000hsv.googlegroups.com>



On Sep 3, 12:50 pm, Olivier Renaud <o.ren... at laposte.net> wrote:

> I think the traffic of this list is low enough for not being split into two
> separate lists. The tickets posts are not a problem for me.

Hi Olivier,

Glad to hear your voice. And being the only one on the topic at hand,
doubly verifies that you are absolutely correct ;)

Which is, Yea!, less work for me anyway.

Thanks,
T.


From transfire at gmail.com  Fri Sep  7 15:21:12 2007
From: transfire at gmail.com (Trans)
Date: Fri, 07 Sep 2007 19:21:12 -0000
Subject: [Facets] [ facets-Patches-13750 ] Factorial Refactoring
In-Reply-To: <20070907110156.62DF55240A39@rubyforge.org>
References: <20070907110156.62DF55240A39@rubyforge.org>
Message-ID: <1189192872.288022.133340@w3g2000hsg.googlegroups.com>



On Sep 7, 4:01 am, <nore... at rubyforge.org> wrote:
> Patches item #13750, was opened at 2007-09-07 06:01
> You can respond by visiting:http://rubyforge.org/tracker/?func=detail&atid=3171&aid=13750&group_i...

> Refactoring of the factorial code using inject.

Don't know why it didn't come over to the mailing list. Anyway, I may
a reply to this on the tacker forum.

Thanks,
T.


From o.renaud at laposte.net  Fri Sep  7 18:44:32 2007
From: o.renaud at laposte.net (Olivier Renaud)
Date: Sat, 8 Sep 2007 00:44:32 +0200
Subject: [Facets] Two vs. one mailing list?
In-Reply-To: <1189165979.413869.252860@57g2000hsv.googlegroups.com>
References: <1188836887.938215.62800@w3g2000hsg.googlegroups.com>
	<200709032150.27572.o.renaud@laposte.net>
	<1189165979.413869.252860@57g2000hsv.googlegroups.com>
Message-ID: <200709080044.32229.o.renaud@laposte.net>

Le vendredi 07 septembre 2007 13:52, Trans a ?crit?:
> On Sep 3, 12:50 pm, Olivier Renaud <o.ren... at laposte.net> wrote:
> > I think the traffic of this list is low enough for not being split into
> > two separate lists. The tickets posts are not a problem for me.
>
> Hi Olivier,
>
> Glad to hear your voice. And being the only one on the topic at hand,
> doubly verifies that you are absolutely correct ;)
>
> Which is, Yea!, less work for me anyway.
>
> Thanks,
> T.

Thanks. Oh yeah, I like this list : my opinions are truly taken into 
account :D

-- 
Olivier Renaud


From transfire at gmail.com  Sun Sep  9 09:57:21 2007
From: transfire at gmail.com (Trans)
Date: Sun, 09 Sep 2007 13:57:21 -0000
Subject: [Facets] Lay me out 2.0
Message-ID: <1189346241.282037.147980@y42g2000hsy.googlegroups.com>

I'm just about ready to release version 2.0. I have only one last
major decision to make, and it is an important one. So I thought it a
good idea to put it out to the community.

As you know, Facets 1.8+ is laid out between facets/core/ and facets/
more/ subdirectories. There is also a facet/ shortcut directory that
simply contains files that redirect back to the first two. The idea
being that the extension method libs were in core/, the other libs in
more/, and facet/ was for convenience.

With Facets 2.0 there are two points that make things different:

1) A goal of 2.0 was to make all the libs directly available via
facets/, without the need for core/ or more/ or redirection.

2) The extensions are no longer stored one-file-per-method, but are
instead bundled with other closely related methods. However, Facets
will still provide per-method requires via redirection files.

So those two points change things a bit. To take them into account,
the current 2.0 layout looks like this:

  lib/facets/  <-- all libs
  lib/facet/  <-- method redirects

Unfortunately this layout almost completely breaks backward-
compatibility with 1.8.

Is there a remedy? I could offer backward compatibility if I did this
instead:

  lib/facets/core/  <-- method redirects
  lib/facets/more/ <-- additional libs
  lib/facets/xore/  <-- extension libs

Yes, the xore/ is a bit weird (do you have a better name?), but it
keeps with the flow. With this layout I can use RubyGems' libpath
specification parameter to add more/ and xore/ to the load path and
achieve my first goal while remaining backward compatible with 1.8.

The problem though, is that _manual_ installs have no means of
automatically adding to the load path. So those will need a special
file that would either have to be loaded at the start of ones app or
added to the RUBYOPT environment variable.

Is that extra hassle worth it?

And, with regard to require statements, how important is backward
compatibility to you?

Thanks,
T.


From o.renaud at laposte.net  Tue Sep 11 16:40:09 2007
From: o.renaud at laposte.net (Olivier Renaud)
Date: Tue, 11 Sep 2007 22:40:09 +0200
Subject: [Facets] Lay me out 2.0
In-Reply-To: <1189346241.282037.147980@y42g2000hsy.googlegroups.com>
References: <1189346241.282037.147980@y42g2000hsy.googlegroups.com>
Message-ID: <200709112240.10583.o.renaud@laposte.net>

Le dimanche 09 septembre 2007 15:57, Trans a ?crit?:
> I'm just about ready to release version 2.0. I have only one last
> major decision to make, and it is an important one. So I thought it a
> good idea to put it out to the community.
>
> As you know, Facets 1.8+ is laid out between facets/core/ and facets/
> more/ subdirectories. There is also a facet/ shortcut directory that
> simply contains files that redirect back to the first two. The idea
> being that the extension method libs were in core/, the other libs in
> more/, and facet/ was for convenience.
>
> With Facets 2.0 there are two points that make things different:
>
> 1) A goal of 2.0 was to make all the libs directly available via
> facets/, without the need for core/ or more/ or redirection.
>
> 2) The extensions are no longer stored one-file-per-method, but are
> instead bundled with other closely related methods. However, Facets
> will still provide per-method requires via redirection files.
>
> So those two points change things a bit. To take them into account,
> the current 2.0 layout looks like this:
>
>   lib/facets/  <-- all libs
>   lib/facet/  <-- method redirects
>
> Unfortunately this layout almost completely breaks backward-
> compatibility with 1.8.
>
> Is there a remedy? I could offer backward compatibility if I did this
> instead:
>
>   lib/facets/core/  <-- method redirects
>   lib/facets/more/ <-- additional libs
>   lib/facets/xore/  <-- extension libs
>
> Yes, the xore/ is a bit weird (do you have a better name?), but it
> keeps with the flow. With this layout I can use RubyGems' libpath
> specification parameter to add more/ and xore/ to the load path and
> achieve my first goal while remaining backward compatible with 1.8.
>
> The problem though, is that _manual_ installs have no means of
> automatically adding to the load path. So those will need a special
> file that would either have to be loaded at the start of ones app or
> added to the RUBYOPT environment variable.
>
> Is that extra hassle worth it?
>
> And, with regard to require statements, how important is backward
> compatibility to you?
>
> Thanks,
> T.

	Uh oh, dilemma ahead !
	I thought this 2.0 version would completely break backward compatibility. But 
if there is a way to be backward compatible, this possibility should be made 
feasible. However, the solution you are presenting does the opposite : it 
keeps the same old layout, and provides a hack for being used the way you 
first intended to. Not a big deal, but I think the better would be : let the 
new layout be the default, and provide a way to be backward compatible. I 
don't know if this is possible.
	Anyway, I say this, but I personnaly broke the API of PQueue when I worked on 
it ! Someone on rubytalk pointed out the incompatibility, so I think there's 
a real need for backward compatibility.

-- 
Olivier Renaud


From transfire at gmail.com  Sat Sep 15 07:43:44 2007
From: transfire at gmail.com (Trans)
Date: Sat, 15 Sep 2007 11:43:44 -0000
Subject: [Facets] Lay me out 2.0
In-Reply-To: <200709112240.10583.o.renaud@laposte.net>
References: <1189346241.282037.147980@y42g2000hsy.googlegroups.com>
	<200709112240.10583.o.renaud@laposte.net>
Message-ID: <1189856624.425676.88360@50g2000hsm.googlegroups.com>



On Sep 11, 1:40 pm, Olivier Renaud <o.ren... at laposte.net> wrote:

>         Uh oh, dilemma ahead !
>         I thought this 2.0 version would completely break backward compatibility. But
> if there is a way to be backward compatible, this possibility should be made
> feasible. However, the solution you are presenting does the opposite : it
> keeps the same old layout, and provides a hack for being used the way you
> first intended to. Not a big deal, but I think the better would be : let the
> new layout be the default, and provide a way to be backward compatible. I
> don't know if this is possible.
>         Anyway, I say this, but I personnaly broke the API of PQueue when I worked on
> it ! Someone on rubytalk pointed out the incompatibility, so I think there's
> a real need for backward compatibility.

Backward compatibility is important, of course. And except for the
libs that have been moved to their own projects, the vast majority of
Facets is backward compatible with 1.8 series. That's not to say there
won't be upgrade issues to address, but mostly 2.0 was a matter of
restructuring, not so much altering functionality.

I actually made a mistake in my analysis of the possibility for
backward compatibility of the require statements. Turns out, it's
simply not possible (short of hacking #require itself). So that's
that. Thankfully this is only a matter of require statements --so it's
not difficult to update a project.

All requires are now simply:

   require 'facets/{name}'

There is no longer more/ or core/, or even singular facet/. This is
achieved via RubyGem's libpath parameter (or a specialized installer
in the case of site_ruby installs --that was the tricky part ;)

Also note, that if you do:

  require 'facets'

You're loading the entire "core" library. (Look in lib/core/facets, to
see what that contains). This is quite different from the old  require
'facets'  which only loaded a dozen or so very basic extensions and
added facets/core and facets/more to the LOAD_PATH. I may make an
extra lib for the latter functionality, but it's not as important now
that everything is simply under facets/.

T.


From transfire at gmail.com  Sat Sep 15 07:58:43 2007
From: transfire at gmail.com (Trans)
Date: Sat, 15 Sep 2007 11:58:43 -0000
Subject: [Facets] 2.0 Beta Gem
Message-ID: <1189857523.903632.118090@g4g2000hsf.googlegroups.com>

If you'd like to try 2.0 out, you can get a beta gem here:

  $ wget http://facets.rubyforge.org/pkg/facets-2.0.0.gem
  $ gem install facets-2.0.0.gem

Usage is simple:

  require 'facets/{lib}'

Where {lib} is the name of a library, class, module or method
reference. Some examples:

  require 'facets/openobject'
  require 'facets/hash'
  require 'facets/string/indexable'
  require 'facets/integer/factorial'

You can also do

  require 'facets'

This loads all Facets' core extensions (look in lib/core/facets to see
what this includes), and is the recommended approach if you are using
numerous extensions. (Note that require 'facets' is not the same as in
1.8 --it no longer adds facets' lib locations to the load path.)

I know of only one feature I have to fix before release (adding #new
to Multiton), though surely a few others will pop up. Please let me
know if you come across anything while trying it out.

T.


From transfire at gmail.com  Sun Sep 16 13:17:09 2007
From: transfire at gmail.com (Trans)
Date: Sun, 16 Sep 2007 17:17:09 -0000
Subject: [Facets] 2.0 Beta Gem
In-Reply-To: <1189857523.903632.118090@g4g2000hsf.googlegroups.com>
References: <1189857523.903632.118090@g4g2000hsf.googlegroups.com>
Message-ID: <1189963029.126847.81370@k79g2000hse.googlegroups.com>



On Sep 15, 4:58 am, Trans <transf... at gmail.com> wrote:
> If you'd like to try 2.0 out, you can get a beta gem here:
>
>   $ wgethttp://facets.rubyforge.org/pkg/facets-2.0.0.gem
>   $ gem install facets-2.0.0.gem

Just updated the gem with a few minor fixes.

T.