From Daniel.Berger at qwest.com  Fri Feb  2 17:40:00 2007
From: Daniel.Berger at qwest.com (Daniel Berger)
Date: Fri, 02 Feb 2007 15:40:00 -0700
Subject: [Rake-devel] [PATCH] clean.rb, directories
Message-ID: <45C3BDC0.5070700@qwest.com>

Hi,

This is a patch for the clean task, so that it only deletes files, not 
directories.

--- clean.orig  Fri Feb  2 15:30:56 2007
+++ clean.rb    Fri Feb  2 15:35:18 2007
@@ -20,7 +20,10 @@

  desc "Remove any temporary products."
  task :clean do
-  CLEAN.each { |fn| rm_r fn rescue nil }
+  CLEAN.each { |fn|
+     next if File.directory?(fn)
+     rm_r fn rescue nil
+  }
  end

  CLOBBER = Rake::FileList.new

This addresses an earlier email where someone brought up the fact that 
"rake clean" deletes a directory named 'core', not just core files. 
Since I hit it myself recently, I figure I should submit a patch. :)

Regards,

Dan

From jim at weirichhouse.org  Sun Feb  4 21:51:34 2007
From: jim at weirichhouse.org (Jim Weirich)
Date: Sun, 04 Feb 2007 21:51:34 -0500
Subject: [Rake-devel] Incresing GemPackageTask Efficiency
In-Reply-To: <4b6f054f0701310725l1827d0d0q9f002bd2a03b2dcf@mail.gmail.com>
References: <4b6f054f0701310725l1827d0d0q9f002bd2a03b2dcf@mail.gmail.com>
Message-ID: <45C69BB6.6000509@weirichhouse.org>

TRANS wrote:
> Note the difference in 'platform' and more importantly 'files'. The
> GemPackageTask would work those out and build the GemSpec at runtime
> rather then definition time.

Thanks for the idea.  I see your point.  I'm going to ponder this for a bit.

-- Jim Weirich

From jim at weirichhouse.org  Sun Feb  4 21:52:58 2007
From: jim at weirichhouse.org (Jim Weirich)
Date: Sun, 04 Feb 2007 21:52:58 -0500
Subject: [Rake-devel] Observer mode: new option --observe
In-Reply-To: <45C0F961.4050009@rubychan.de>
References: <45C0F961.4050009@rubychan.de>
Message-ID: <45C69C0A.3010402@weirichhouse.org>

murphy wrote:
 > hi list!
 >
 > I've implemented a new feature to Rake that seems very useful to me.
 >
 > You can use the --observe (-O) option and an optional interval to start
 > Rake in an endless loop, like this:
 >
 > % rake -O compile

What an interesting idea.  Let me mull this over a bit, but I very well 
may work this into the system.  Thanks.

-- 
-- Jim Weirich      jim at weirichhouse.org     http://onestepback.org
-- In theory, practice and theory are the same.
-- In practice, they are different.




From jim at weirichhouse.org  Sun Feb  4 21:55:35 2007
From: jim at weirichhouse.org (Jim Weirich)
Date: Sun, 04 Feb 2007 21:55:35 -0500
Subject: [Rake-devel] [PATCH] clean.rb, directories
In-Reply-To: <45C3BDC0.5070700@qwest.com>
References: <45C3BDC0.5070700@qwest.com>
Message-ID: <45C69CA7.1070004@weirichhouse.org>

Daniel Berger wrote:
> Hi,
> 
> This is a patch for the clean task, so that it only deletes files, not 
> directories.
[...]
> This addresses an earlier email where someone brought up the fact that 
> "rake clean" deletes a directory named 'core', not just core files. 
> Since I hit it myself recently, I figure I should submit a patch. :)

Thanks for the suggestion.  However, prohibiting CLEAN from deleting 
directories feels like it is cutting off the arm to fix a hangnail. 
CLEAN should be able to delete any temporary files created during a 
build process, and that might include directories.

I would rather see a solution targeted explicitly at the "core" 
directory issue.

Thanks Daniel.

-- Jim Weirich


From transfire at gmail.com  Sun Feb  4 22:39:13 2007
From: transfire at gmail.com (TRANS)
Date: Sun, 4 Feb 2007 22:39:13 -0500
Subject: [Rake-devel] [PATCH] clean.rb, directories
In-Reply-To: <45C69CA7.1070004@weirichhouse.org>
References: <45C3BDC0.5070700@qwest.com> <45C69CA7.1070004@weirichhouse.org>
Message-ID: <4b6f054f0702041939h6c2b6096lfda66e2822bc57f2@mail.gmail.com>

On 2/4/07, Jim Weirich <jim at weirichhouse.org> wrote:
> Daniel Berger wrote:
> > Hi,
> >
> > This is a patch for the clean task, so that it only deletes files, not
> > directories.
> [...]
> > This addresses an earlier email where someone brought up the fact that
> > "rake clean" deletes a directory named 'core', not just core files.
> > Since I hit it myself recently, I figure I should submit a patch. :)
>
> Thanks for the suggestion.  However, prohibiting CLEAN from deleting
> directories feels like it is cutting off the arm to fix a hangnail.
> CLEAN should be able to delete any temporary files created during a
> build process, and that might include directories.
>
> I would rather see a solution targeted explicitly at the "core"
> directory issue.

Perhaps it would be a good time to bring this up. In setting up my
rake tasks I've sort of took the opposite approach to CLEAN and
created IGNORE. When building, say a package, I just have it include
everything but what's in IGNORE, so there's no need to run 'clean'
first.

T.

From transfire at gmail.com  Sun Feb  4 22:59:14 2007
From: transfire at gmail.com (TRANS)
Date: Sun, 4 Feb 2007 22:59:14 -0500
Subject: [Rake-devel] Eleminating the need for repackage and rerdoc ?
Message-ID: <4b6f054f0702041959oeff9755t3adaa7698213df38@mail.gmail.com>

The PackageTask and the RdocTask both create re-renditions of the main
task in order to force the creation of packages or rdocs that already
exist. Personally, I just did

  Rake::Task['rerdoc'].comment = nil
  task :rdoc => [:clobber_rdoc]

Perhaps that's considered dangerous for some reason? If so why not
just ask the user if they're sure if they want to contiue or not?

Another possibility. What about using a --force option on the command
line to do a re-task.

Example:

  $ rake rdoc
  rdocs at doc/ already exist. Use --force to rerdoc.

  $ rake rdoc --force
   ...

Not sure, but it does seem kind of a waste to have two tasks that
essentailly do the very same thing.

Food for thought,
T.

From djberg96 at gmail.com  Sun Feb  4 23:23:41 2007
From: djberg96 at gmail.com (Daniel Berger)
Date: Sun, 4 Feb 2007 21:23:41 -0700
Subject: [Rake-devel] [PATCH] clean.rb, directories
In-Reply-To: <4b6f054f0702041939h6c2b6096lfda66e2822bc57f2@mail.gmail.com>
References: <45C3BDC0.5070700@qwest.com> <45C69CA7.1070004@weirichhouse.org>
	<4b6f054f0702041939h6c2b6096lfda66e2822bc57f2@mail.gmail.com>
Message-ID: <6037b70c0702042023n4b8a6cefoa3890fd38833a295@mail.gmail.com>

On 2/4/07, TRANS <transfire at gmail.com> wrote:
> On 2/4/07, Jim Weirich <jim at weirichhouse.org> wrote:
> > Daniel Berger wrote:
> > > Hi,
> > >
> > > This is a patch for the clean task, so that it only deletes files, not
> > > directories.
> > [...]
> > > This addresses an earlier email where someone brought up the fact that
> > > "rake clean" deletes a directory named 'core', not just core files.
> > > Since I hit it myself recently, I figure I should submit a patch. :)
> >
> > Thanks for the suggestion.  However, prohibiting CLEAN from deleting
> > directories feels like it is cutting off the arm to fix a hangnail.
> > CLEAN should be able to delete any temporary files created during a
> > build process, and that might include directories.
> >
> > I would rather see a solution targeted explicitly at the "core"
> > directory issue.
>
> Perhaps it would be a good time to bring this up. In setting up my
> rake tasks I've sort of took the opposite approach to CLEAN and
> created IGNORE. When building, say a package, I just have it include
> everything but what's in IGNORE, so there's no need to run 'clean'
> first.

That seems like a reasonable suggestion, and I don't think it's
mutually exclusive to dealing with directories named "core". We can do
both.

But, to deal with just the 'core' issue for now:

--- clean.orig  Sun Jun 06 00:51:00 2004
+++ clean.rb    Sun Feb 04 21:22:10 2007
@@ -20,7 +20,11 @@

 desc "Remove any temporary products."
 task :clean do
-  CLEAN.each { |fn| rm_r fn rescue nil }
+  CLEAN.each { |fn|
+    unless fn.downcase == 'core' && File.directory?(fn)
+      rm_r fn rescue nil
+    end
+  }
 end

 CLOBBER = Rake::FileList.new

Regards,

Dan

From transfire at gmail.com  Mon Feb  5 15:09:19 2007
From: transfire at gmail.com (TRANS)
Date: Mon, 5 Feb 2007 15:09:19 -0500
Subject: [Rake-devel] backwards dependency?
Message-ID: <4b6f054f0702051209s113e64bfi6ad8c688385c8814@mail.gmail.com>

I'm confused. I'm trying to alter the dependencies of a premade rake
task, and while it accepts it, it's actully running the dependcy after
the task rather than before it. With this:

  Rake::GemPackageTask.new(spec) do |pkg|
    ...
  end

  # Don't show the clobber task, just always clobber when
  # creating packages. Maybe in future ask to make sure.
  Rake::Task['clobber_package'].comment = nil
  Rake::Task['repackage'].comment = nil
  task :package => [:clobber_package]

Shouldn't it invoke clobber_package BEFORE package? I ran it with --
trace and it's clearly doing the opposite. What's up?

Thanks,
T.

From transfire at gmail.com  Tue Feb  6 16:04:07 2007
From: transfire at gmail.com (TRANS)
Date: Tue, 6 Feb 2007 16:04:07 -0500
Subject: [Rake-devel] backwards dependency?
In-Reply-To: <4b6f054f0702051209s113e64bfi6ad8c688385c8814@mail.gmail.com>
References: <4b6f054f0702051209s113e64bfi6ad8c688385c8814@mail.gmail.com>
Message-ID: <4b6f054f0702061304y59220dfbr467343c3a45d9852@mail.gmail.com>

>   Rake::GemPackageTask.new(spec) do |pkg|
>     ...
>   end
>
>   task :package => [:clobber_package]

No one can tell me why this isn't working?

T.

From jim at weirichhouse.org  Sat Feb 10 14:40:26 2007
From: jim at weirichhouse.org (Jim Weirich)
Date: Sat, 10 Feb 2007 14:40:26 -0500
Subject: [Rake-devel] backwards dependency?
In-Reply-To: <4b6f054f0702051209s113e64bfi6ad8c688385c8814@mail.gmail.com>
References: <4b6f054f0702051209s113e64bfi6ad8c688385c8814@mail.gmail.com>
Message-ID: <45CE1FAA.5000704@weirichhouse.org>

TRANS wrote:
 > I'm confused. I'm trying to alter the dependencies of a premade rake
 > task, and while it accepts it, it's actully running the dependcy after
 > the task rather than before it. With this:
 >
 >   Rake::GemPackageTask.new(spec) do |pkg|
 >     ...
 >   end
 >
 >   # Don't show the clobber task, just always clobber when
 >   # creating packages. Maybe in future ask to make sure.
 >   Rake::Task['clobber_package'].comment = nil
 >   Rake::Task['repackage'].comment = nil
 >   task :package => [:clobber_package]
 >
 > Shouldn't it invoke clobber_package BEFORE package? I ran it with --
 > trace and it's clearly doing the opposite. What's up?

It does.  But the package task body is entirely empty.  So after it has 
satisfied all of its other dependencies (which do the work of building 
the packaages), then it invokes its last dependency (your 
clobber_package).  Finally it runs the body (which is empty).

The trick is to get clobber_package to be first in the list of package's 
dependencies.  Just move the task :package line to be before the 
creation of the Rake::GemPackageTask.

-- 
-- Jim Weirich      jim at weirichhouse.org     http://onestepback.org
-- In theory, practice and theory are the same.
-- In practice, they are different.


From transfire at gmail.com  Sat Feb 10 18:08:50 2007
From: transfire at gmail.com (TRANS)
Date: Sat, 10 Feb 2007 18:08:50 -0500
Subject: [Rake-devel] backwards dependency?
In-Reply-To: <45CE1FAA.5000704@weirichhouse.org>
References: <4b6f054f0702051209s113e64bfi6ad8c688385c8814@mail.gmail.com>
	<45CE1FAA.5000704@weirichhouse.org>
Message-ID: <4b6f054f0702101508h515d1fbdid7efe046637d6af4@mail.gmail.com>

On 2/10/07, Jim Weirich <jim at weirichhouse.org> wrote:

> It does.  But the package task body is entirely empty.  So after it has
> satisfied all of its other dependencies (which do the work of building
> the packaages), then it invokes its last dependency (your
> clobber_package).  Finally it runs the body (which is empty).
>
> The trick is to get clobber_package to be first in the list of package's
> dependencies.  Just move the task :package line to be before the
> creation of the Rake::GemPackageTask.

ah, I see. thanks!

t.

From zimbatm at oree.ch  Sat Feb 10 20:24:36 2007
From: zimbatm at oree.ch (Jonas Pfenniger)
Date: Sun, 11 Feb 2007 02:24:36 +0100
Subject: [Rake-devel] about the tasklib DSL
Message-ID: <e3caf1460702101724x7110d385ic2c2a598d5031cc8@mail.gmail.com>

Hello,

I really enjoy using Rake but somehow, I'm not able to "get" the tasklib
extension's principles. IMHO, the DSL is not as expressive as the simple
task definition. Why can't I, for example define a test task like that :

test_task, :test_all do |t|
  ...
end

or

task TestTask, :test_all

instead of

TestTask.new(:test_all)...

?

-- 
Cheers,
  zimbatm
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://rubyforge.org/pipermail/rake-devel/attachments/20070211/1aa73c8d/attachment-0001.html 

From murphy at rubychan.de  Sat Feb 17 16:34:58 2007
From: murphy at rubychan.de (murphy)
Date: Sat, 17 Feb 2007 22:34:58 +0100
Subject: [Rake-devel] Invoke tasks more than once
In-Reply-To: <45C0F961.4050009@rubychan.de>
References: <45C0F961.4050009@rubychan.de>
Message-ID: <45D77502.1070703@rubychan.de>

I wrote:
> ... As you can see, I had
> to tweak Task#invoke to allow them to be called repeatedly.
I just noticed that this is handy for other things, like:

  namespace :db do
    desc 'Delete all tables, and migrate DB to latest version.'
    task :reload do
      ENV['VERSION'] = '0'
      Rake::Task['db:migrate'].invoke
      ENV.delete 'VERSION'
      Rake::Task['db:migrate'].invoke false
    end
  end

...which is not possible with the current invoke method.

So, independent from the Observer Mode patch, it would be nice to have
this in Rake 0.8 :)

Actually, it might be preferable to change the semantics of #invoke to
skip the only-once check by default, and check only if the option "true"
is given. If rake is started from the shell, it would act as usual, but
if you use it manually (like in the above example), you would get the
expected result (two invocations), without a parameter.
Alternatively, an invoke_once method might be even better.

[murphy]

From Daniel.Berger at qwest.com  Mon Feb 26 10:28:00 2007
From: Daniel.Berger at qwest.com (Daniel Berger)
Date: Mon, 26 Feb 2007 08:28:00 -0700
Subject: [Rake-devel] Rake 0.7.2 and Ruby 1.8.6
Message-ID: <45E2FC80.3040300@qwest.com>

Hi Jim,

As was brought up on ruby-core, I think it would be wise to put out the 
next release of rake before 1.8.6 goes final. Otherwise, there are going 
to be a LOT of unhappy Rails users.

Any ideas for a release date?

Thanks,

Dan