[Ironruby-core] SVN r114 is out

Curt Hagenlocher curth at microsoft.com
Thu Jun 12 23:29:54 EDT 2008

What’s changed since the last release is that we’re now exposing protected members on CLR types as protected members on the Ruby class that’s generated to represent the type.  And in doing so, we’re apparently hiding the (public) no-arg method on the base class.  And what’s more, it looks like the protected one-arg overload is also being exposed as though it were public, and it shouldn’t be.

Improved CLR interop is very high on our priority list.

Nice function, btw, and an excellent example of the power of Ruby.  If only the language didn’t make life so hard on implementers…

From: ironruby-core-bounces at rubyforge.org [mailto:ironruby-core-bounces at rubyforge.org] On Behalf Of Dino Viehland
Sent: Thursday, June 12, 2008 8:21 PM
To: ironruby-core at rubyforge.org
Subject: Re: [Ironruby-core] SVN r114 is out

StringReader defines dispose(bool disposing) and inherits dispose() from TextReader.  So it’s either that dispose(bool) is completely hiding the inherited dispose or the overload resolution is completely broken – probably the former.  It’s probably related to exposing protected members to everyone or just some bug in filtering protected members out.

From: ironruby-core-bounces at rubyforge.org [mailto:ironruby-core-bounces at rubyforge.org] On Behalf Of Ivan Porto Carrero
Sent: Thursday, June 12, 2008 8:08 PM
To: ironruby-core at rubyforge.org
Subject: Re: [Ironruby-core] SVN r114 is out

What could be wrong here?

  def using(o)
      yield if block_given?

This code worked until rev.113 but now it tells me :

>>> $sr = StringReader.new('Hello, world')
=> #<System::IO::StringReader:0x0000060>
>>> $sr.dispose
:0:in `Initialize': wrong number or type of arguments for `dispose' (ArgumentError)

On Fri, Jun 13, 2008 at 2:49 PM, John Lam (IRONRUBY) <jflam at microsoft.com<mailto:jflam at microsoft.com>> wrote:

This update contains the code we demo'd at RailsConf, and a few improvements that we've made since then.

If you rake compile this build, it will generate an ir.cmd file for you in the build\debug directory. The rakefile now looks for your MRI install (by searching your PATH for ruby.exe). If you have your build\debug directory on your path, this file will run correctly.

You will see a bunch of warnings when compiling. We know about these, and will fix them soon (it's the DLR folks deprecating a bunch of old APIs).

I've also added a RubyTestKey.snk file which contains a public key. Our assemblies are delay-signed using this key, and you must enable skip verification to run the assemblies on your dev machine. You'll need to run the runfirst.cmd file in svn\trunk to turn on skip verification for assemblies signed using our public key.

This should simplify some of the SIGN flag problems, although it introduces a new set of problems around signing if you want to distribute the signed assemblies. Unfortunately, we had to enable signing on external builds because there's some bug that only occurs when running an unsigned build. We're investigating this now.



Ironruby-core mailing list
Ironruby-core at rubyforge.org<mailto:Ironruby-core at rubyforge.org>

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://rubyforge.org/pipermail/ironruby-core/attachments/20080612/f574c5d3/attachment.html>

More information about the Ironruby-core mailing list