Ironruby core - Feb 2008 - Support For Referencing Ruby Libs As Embedded Resources?

A recent inquiry on the Ruby.NET list brought up the question as to  
whether or not the ability to reference ruby libs as embedded resources is  
supported or will be supported via IronRuby. e.g.,

require ''res://assembly.dll/foo/bar''

Yes/No/Maybe?  Reasons For/Against/Agnostic?

-- 
/M:D

M. David Peterson
Co-Founder & Chief Architect, 3rd&Urban, LLC
Email: m.david at 3rdandUrban.com | m.david at amp.fm
Mobile: (206) 418-9027
http://3rdandUrban.com | http://amp.fm |  
http://www.oreillynet.com/pub/au/2354

M. David Peterson:
> A recent inquiry on the Ruby.NET list brought up the question as to
> whether or not the ability to reference ruby libs as embedded resources
> is supported or will be supported via IronRuby. e.g.,
>
> require ''res://assembly.dll/foo/bar''
>
> Yes/No/Maybe?  Reasons For/Against/Agnostic?
Is the embedded resource a Ruby file or something else?
What is the feature used for?

- John

On Feb 8, 2008 5:48 PM, John Messerly <jomes at microsoft.com>
wrote:
> M. David Peterson:
>
> > A recent inquiry on the Ruby.NET list brought up the question as to
> > whether or not the ability to reference ruby libs as embedded
resources
> > is supported or will be supported via IronRuby. e.g.,
> >
> > require ''res://assembly.dll/foo/bar''
>
> Is the embedded resource a Ruby file or something else?
> What is the feature used for?
I imagine the idea is to make deployment more robust by putting the
Ruby source somewhere that a user can''t screw up. ;)

At least, that''s what *I''d* want it for...

--
Curt Hagenlocher
curt at hagenlocher.org

On Fri, 08 Feb 2008 18:48:39 -0700, John Messerly <jomes at microsoft.com>
wrote:
> Is the embedded resource a Ruby file or something else?
Yes, a Ruby file.
> What is the feature used for?
Here''s a copy of the pertinent pieces from the related thread on the  
Ruby.NET list.  In short, for ease of redistribution of stdlib source  
files.

On Fri, 08 Feb 2008 18:26:57 -0700, M. David Peterson  
<m.david at xmlhacker.com> wrote:
> On Fri, 08 Feb 2008 13:33:47 -0700, chriso <chrisortman at gmail.com>
wrote:
>> Is there a way to bundle any required libraries into the compiled
>> dll?
>
> You could embed a source file into the assembly and then resolve that  
> file internally to your code base.  However, I''m not sure if there
is
> any officially sanctioned way as of yet to reference embedded resources  
> (e.g. res://assemblyname.dll/foo), though this certainly seems like  
> something that needs to be both considered and discussed.
>
> A temporary hack[1,2] _could_ be used  (though I wouldn''t
recommend this
> for anything beyond experimentation) via creating an assembly
who''s only
> purpose was to return requests for embedded resources to then evaluate  
> the returned Ruby.String, but this would only work for a small fraction  
> of use-cases in which the referenced stdlib didn''t reference any
other
> stdlibs.  In this regard there would need to be some way of indicating  
> to the Ruby.NET Runtime whether it should look for an embedded resource,  
> and if yes, where it should look.
>
> I would think that using the standard res://assemblyname.dll/foo/bar  
> would make the most sense.  Anybody happen to have a sense as to how to  
> properly handle this?  I''ll make a post to the IronRuby list to
see if
> they plan to provide support for embedded resources, and if so, the  
> syntax they plan to use.
>
>
> [1] C# ResourceManager which returns the Ruby.String for the referenced  
> embedded resource.
>
> using System;
> using System.Resources;
> using System.Text;
>
> namespace ResourceLib
> {
>      public class RubyStdLib
>      {
>          static Encoding m_encoding = Encoding.UTF8;
>          static ResourceManager manager =  
> Properties.Resources.ResourceManager;
>
>          public Ruby.String getStdLib(string stdLibName)
>          {
>              return new  
> Ruby.String(m_encoding.GetString((byte[])manager.GetObject(stdLibName)));
>          }
>      }
> }
>
> [2] Modified Ruby
>
> require ''mscorlib.dll''
> a_date = System::DateTime.new(2007, 1, 1)
> puts a_date.AddYears(1).ToString()
> puts ''Hello .net''
>
> input = %{\
> <% high.downto(low) do |n| # set high, low externally %>
> <%= n %> green bottles, hanging on the wall
> <%= n %> green bottles, hanging on the wall
> And if one green bottle should accidentally fall
> There''d be <%= n - 1
> %> green bottles, hanging on the wall
> <% end %>
> }
> high,low = 10, 8
>
> # Beware of the horrific hack just begging to be exploited that follows.
> # Beyond loading the source file directly or using the taint method is
> # there a better way to handle this that''s not immediately
obvious?
>
> require ''ResourceLib.dll''
> rm = ResourceLib::RubyStdLib.new
> erb = eval(rm.getStdLib(''erb''))::ERB.new(input)
> puts erb
> bind = Kernel.binding
> puts bind
> puts erb.run(bind)
>
> puts ''Bye .net''
>


-- 
/M:D

M. David Peterson
Co-Founder & Chief Architect, 3rd&Urban, LLC
Email: m.david at 3rdandUrban.com | m.david at amp.fm
Mobile: (206) 418-9027
http://3rdandUrban.com | http://amp.fm |  
http://www.oreillynet.com/pub/au/2354

On Fri, 08 Feb 2008 19:00:26 -0700, Curt Hagenlocher  
<curt at hagenlocher.org> wrote:
> I imagine the idea is to make deployment more robust by putting the
> Ruby source somewhere that a user can''t screw up.
Yup!
> At least, that''s what *I''d* want it for...
You and I both, Curt! :D

-- 
/M:D

M. David Peterson
Co-Founder & Chief Architect, 3rd&Urban, LLC
Email: m.david at 3rdandUrban.com | m.david at amp.fm
Mobile: (206) 418-9027
http://3rdandUrban.com | http://amp.fm |  
http://www.oreillynet.com/pub/au/2354

Curt Hagenlocher:
> On Feb 8, 2008 5:48 PM, John Messerly <jomes at microsoft.com> wrote:
> > M. David Peterson:
> >
> > > A recent inquiry on the Ruby.NET list brought up the question as
to
> > > whether or not the ability to reference ruby libs as embedded
> > > resources is supported or will be supported via IronRuby. e.g.,
> > >
> > > require ''res://assembly.dll/foo/bar''
> >
> > Is the embedded resource a Ruby file or something else?
> > What is the feature used for?
>
> I imagine the idea is to make deployment more robust by putting the
> Ruby source somewhere that a user can''t screw up. ;)
>
> At least, that''s what *I''d* want it for...
Right, but if you have a DLL already, you can just host IronRuby from C# and
feed it Ruby code however you want to.

I guess I''d be hesitant to change the meaning of "require".
The way we have it now (require ''mscorlib'') is nice because
it''s not changing the behavior. If you really wanted to you could write
mscorlib.rb and run your code on another Ruby implementation.

- John

On Fri, 08 Feb 2008 19:09:25 -0700, John Messerly <jomes at microsoft.com>
wrote:
> Right, but if you have a DLL already, you can just host IronRuby from C#  
> and feed it Ruby code however you want to.
True.  So would that then be the prefered way to go about providing  
support?
> I guess I''d be hesitant to change the meaning of
"require". The way we
> have it now (require ''mscorlib'') is nice because
it''s not changing the
> behavior. If you really wanted to you could write mscorlib.rb and run  
> your code on another Ruby implementation.
Oh, I''m not suggesting changing the meaning, just adding a handler for
the
''res'' protocol that would locate the source file via an
assembly rather
than the file system.  To me, anyway, I don''t see any difference
between
what,

require
''C:\WINDOWS\Microsoft.NET\Framework\v2.0.50727\System.dll''

... means and what,

require ''res://assemblyname.dll/foo/bar''

... means, at least in regards to the fact that they both represent  
references to external libraries.

-- 
/M:D

M. David Peterson
Co-Founder & Chief Architect, 3rd&Urban, LLC
Email: m.david at 3rdandUrban.com | m.david at amp.fm
Mobile: (206) 418-9027
http://3rdandUrban.com | http://amp.fm |  
http://www.oreillynet.com/pub/au/2354

On Feb 8, 2008 6:16 PM, M. David Peterson <m.david at xmlhacker.com>
wrote:
> On Fri, 08 Feb 2008 19:09:25 -0700, John Messerly <jomes at
microsoft.com>
> wrote:
>
> > I guess I''d be hesitant to change the meaning of
"require". The way we
> > have it now (require ''mscorlib'') is nice because
it''s not changing the
> > behavior. If you really wanted to you could write mscorlib.rb and run
> > your code on another Ruby implementation.
>
> Oh, I''m not suggesting changing the meaning, just adding a handler
for the
> ''res'' protocol that would locate the source file via an
assembly rather
> than the file system.
>
After all, if urirequire can do this for "http:"...

--
Curt Hagenlocher
curt at hagenlocher.org
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
http://rubyforge.org/pipermail/ironruby-core/attachments/20080208/71d92fec/attachment.html

I''d personally much rather see it done via something like urirequire
(heck if it''s that needed, someone can write resrequire ;) )

Though there is a debate in there about if require should handle URIs
in general or not, I think it''s one for Ruby in general to have...

On Feb 8, 2008 9:54 PM, Curt Hagenlocher <curt at hagenlocher.org>
wrote:
>
> On Feb 8, 2008 6:16 PM, M. David Peterson <m.david at xmlhacker.com>
wrote:
>
> >
> > On Fri, 08 Feb 2008 19:09:25 -0700, John Messerly <jomes at
microsoft.com>
> > wrote:
> >
> >
> >
> > > I guess I''d be hesitant to change the meaning of
"require". The way we
> > > have it now (require ''mscorlib'') is nice
because it''s not changing the
> > > behavior. If you really wanted to you could write mscorlib.rb and
run
> > > your code on another Ruby implementation.
> >
> > Oh, I''m not suggesting changing the meaning, just adding a
handler for the
> > ''res'' protocol that would locate the source file via
an assembly rather
> > than the file system.
>
>
> After all, if urirequire can do this for "http:"...
>
>
> --
> Curt Hagenlocher
> curt at hagenlocher.org
> _______________________________________________
> Ironruby-core mailing list
> Ironruby-core at rubyforge.org
> http://rubyforge.org/mailman/listinfo/ironruby-core
>
>


-- 
Michael Letterle
[Polymath Programmer]
http://michaeldotnet.blogspot.com

On Fri, 08 Feb 2008 20:11:27 -0700, Michael Letterle  
<michael.letterle at gmail.com> wrote:
> heck if it''s that needed, someone can write resrequire
Assuming this would be seen as standard practice for implementing core  
language extensions, this would probably make the most sense, avoiding any  
compromise and ensuring that there''s a simple escape hatch to fallback
to
require if the resrequire method isn''t supported or if the requested  
embedded resource is unavailable.

-- 
/M:D

M. David Peterson
Co-Founder & Chief Architect, 3rd&Urban, LLC
Email: m.david at 3rdandUrban.com | m.david at amp.fm
Mobile: (206) 418-9027
http://3rdandUrban.com | http://amp.fm |  
http://www.oreillynet.com/pub/au/2354

My /personal/ view is that IronRuby should be Ruby first and .NET
second.  Anything that''s done outside of "normal" behaviour
should be
implemented outside of the core implementation.  I''d almost go as far
as to say that the current method of loading .NET assemblies is
pushing it, but that''s an extreme view.

So yes, if we''re doing something specific to this platform it should
be implemented ontop of the core language not in it.  It would be up
to the person writing the code that uses resrequire (or urirequire) to
perform the graceful fallback.


On Feb 8, 2008 10:35 PM, M. David Peterson <m.david at xmlhacker.com>
wrote:
> On Fri, 08 Feb 2008 20:11:27 -0700, Michael Letterle
> <michael.letterle at gmail.com> wrote:
>
> > heck if it''s that needed, someone can write resrequire
>
> Assuming this would be seen as standard practice for implementing core
> language extensions, this would probably make the most sense, avoiding any
> compromise and ensuring that there''s a simple escape hatch to
fallback to
> require if the resrequire method isn''t supported or if the
requested
> embedded resource is unavailable.
>
> --
> /M:D
>
> M. David Peterson
> Co-Founder & Chief Architect, 3rd&Urban, LLC
> Email: m.david at 3rdandUrban.com | m.david at amp.fm
> Mobile: (206) 418-9027
> http://3rdandUrban.com | http://amp.fm |
> http://www.oreillynet.com/pub/au/2354
> _______________________________________________
>
> Ironruby-core mailing list
> Ironruby-core at rubyforge.org
> http://rubyforge.org/mailman/listinfo/ironruby-core
>


-- 
Michael Letterle
[Polymath Programmer]
http://michaeldotnet.blogspot.com

On Fri, 08 Feb 2008 20:57:58 -0700, Michael Letterle  
<michael.letterle at gmail.com> wrote:
> So yes, if we''re doing something specific to this platform it
should
> be implemented ontop of the core language not in it.  It would be up
> to the person writing the code that uses resrequire (or urirequire) to
> perform the graceful fallback.
That makes complete sense and I can''t help but agree. :)

-- 
/M:D

M. David Peterson
Co-Founder & Chief Architect, 3rd&Urban, LLC
Email: m.david at 3rdandUrban.com | m.david at amp.fm
Mobile: (206) 418-9027
http://3rdandUrban.com | http://amp.fm |  
http://www.oreillynet.com/pub/au/2354

For what it''s worth, this /is/ possible today:

require ''mscorlib''

def resrequire(assembly, resource)



  assem = System::Reflection::Assembly.LoadFile assembly

  stream = assem.GetManifestResourceStream(resource)

  streamreader = System::IO::StreamReader.new stream

  filetext = streamreader.ReadToEnd

  file = File.new resource, ''w''

  file.write filetext

  file.close

  require resource

  File.delete resource



end

I think I''d rather use eval then create a file, but we don''t
have that yet ;)

On Feb 9, 2008 12:45 AM, M. David Peterson <m.david at xmlhacker.com>
wrote:
> On Fri, 08 Feb 2008 20:57:58 -0700, Michael Letterle
> <michael.letterle at gmail.com> wrote:
>
> > So yes, if we''re doing something specific to this platform it
should
> > be implemented ontop of the core language not in it.  It would be up
> > to the person writing the code that uses resrequire (or urirequire) to
> > perform the graceful fallback.
>
> That makes complete sense and I can''t help but agree. :)
>
> --
>
> /M:D
>
> M. David Peterson
> Co-Founder & Chief Architect, 3rd&Urban, LLC
> Email: m.david at 3rdandUrban.com | m.david at amp.fm
> Mobile: (206) 418-9027
> http://3rdandUrban.com | http://amp.fm |
> http://www.oreillynet.com/pub/au/2354
> _______________________________________________
> Ironruby-core mailing list
> Ironruby-core at rubyforge.org
> http://rubyforge.org/mailman/listinfo/ironruby-core
>


-- 
Michael Letterle
[Polymath Programmer]
http://michaeldotnet.blogspot.com