rubycocoa-talk — talking about RubyCocoa

Re: [Rubycocoa-talk] CocoaRuby demo app

[Eloy D. <elo...@gm...>]

Hey Stephyn,
It's in svn:
http://rubycocoa.svn.sourceforge.net/viewvc/rubycocoa/branches/users/hisa/junk/CocoaRepl/


Cheers,
Eloy

On 6/20/07, Stephyn Butcher <ste...@co...> wrote:
>
> Dear group,
>
> I noticed that HOJIMOTO Hisa uses an interesting app in his
> demonstrations. I think the name is "CocoaRuby" and looks like it has
> 2 panels, one for Ruby code and one for the method results. He
> changes the title of the window, spawns other windows, etc.
>
> I was wondering if it was generally available.
>
> I know that Hisa contributes to this list so he might see this note.
> I sent him an email but you never know how aggressive some SPAM
> filters are.
>
> Cheers,
> Steve
>
>
> -------------------------------------------------------------------------
> This SF.net email is sponsored by DB2 Express
> Download DB2 Express C - the FREE version of DB2 express and take
> control of your XML. No limits. Just data. Click to get it now.
> http://sourceforge.net/powerbar/db2/
> _______________________________________________
> Rubycocoa-talk mailing list
> Rub...@li...
> https://lists.sourceforge.net/lists/listinfo/rubycocoa-talk
>

Re: [Rubycocoa-talk] copyWithZone (NSCopying)

[Jon B. <jo...@jo...>]

I may have answered my own question ...

return self.retain

On Jun 20, 2007, at 10:56 AM, Jon Baer wrote:

> Hi,
>
> I need to implement something which requires the NSCopying protocol  
> + want to know if attr_accessor will handle this automatically or  
> if returning self is how this would/should work:
>
> def copyWithZone
>   OSX::NSLog "Copy with zone called"
>   return self
> end
>
> My object was a normal Ruby object but since it requires this  
> method Ive extending OSX::NSObject and wondering if this will  
> automagically work for doing this, any ideas or pointers?
>
> Thanks.
>
> - Jon
>

[Rubycocoa-talk] CocoaRuby demo app

[Stephyn B. <ste...@co...>]

Dear group,

I noticed that HOJIMOTO Hisa uses an interesting app in his  
demonstrations. I think the name is "CocoaRuby" and looks like it has  
2 panels, one for Ruby code and one for the method results. He  
changes the title of the window, spawns other windows, etc.

I was wondering if it was generally available.

I know that Hisa contributes to this list so he might see this note.  
I sent him an email but you never know how aggressive some SPAM  
filters are.

Cheers,
Steve

[Rubycocoa-talk] copyWithZone (NSCopying)

[Jon B. <jo...@jo...>]

Hi,

I need to implement something which requires the NSCopying protocol +  
want to know if attr_accessor will handle this automatically or if  
returning self is how this would/should work:

def copyWithZone
   OSX::NSLog "Copy with zone called"
   return self
end

My object was a normal Ruby object but since it requires this method  
Ive extending OSX::NSObject and wondering if this will automagically  
work for doing this, any ideas or pointers?

Thanks.

- Jon

[Rubycocoa-talk] TableView (cell question)

[Jon B. <jo...@gm...>]

Need clarification.  Can someone tell me if this is correct form/way  
(it does work):

def tableView_willDisplayCell_forTableColumn_row(tableView, cell,  
tableColumn, row)
   OSX::NSLog("Table view will display cell for table column called")
   case tableColumn.identifier
   when "pic_square"
     OSX::NSLog("Drawing pic_square for row " + row.to_s)
	  OSX::NSLog(@friends[row].pic_square.to_s)
        if (@friends[row].pic_square.length > 0) then
	  url = OSX::NSURL.URLWithString(@friends[row].pic_square.to_s)
	  image = OSX::NSImage.alloc.initWithContentsOfURL(url)
	  cell.setObjectValue(image)
           return cell
	else
	  return cell
	end
   end
end

Im trying to figure out if I do need to return the cell, the docs say  
void, can I just modify the cells @ will + they are retained?

Thanks.

- Jon

Re: [Rubycocoa-talk] Outline view always showing disclosure triangles

[Jon B. <jo...@gm...>]

Growl .. rawr!!!

On Jun 20, 2007, at 7:22 AM, Tim Perrett wrote:

> Ahh interesting, what framework was causing the conflict out of
> interest?
>
> Thanks
>
> Tim
>
> On 20 Jun 2007, at 12:05, Jon Baer wrote:
>
>> Turns out my problem was related to a newer SDK of a Framework I was
>> using so I was making bad delegate calls.  The error is now gone.
>
>
> ---------------------------------------------------------------------- 
> ---
> This SF.net email is sponsored by DB2 Express
> Download DB2 Express C - the FREE version of DB2 express and take
> control of your XML. No limits. Just data. Click to get it now.
> http://sourceforge.net/powerbar/db2/
> _______________________________________________
> Rubycocoa-talk mailing list
> Rub...@li...
> https://lists.sourceforge.net/lists/listinfo/rubycocoa-talk

Re: [Rubycocoa-talk] Outline view always showing disclosure triangles

[Tim P. <he...@ti...>]

Ahh interesting, what framework was causing the conflict out of  
interest?

Thanks

Tim

On 20 Jun 2007, at 12:05, Jon Baer wrote:

> Turns out my problem was related to a newer SDK of a Framework I was
> using so I was making bad delegate calls.  The error is now gone.

Re: [Rubycocoa-talk] Outline view always showing disclosure triangles

[Jon B. <jo...@gm...>]

Im on 10.4.9 w/ macport ruby + svn copy of rc latest.

Turns out my problem was related to a newer SDK of a Framework I was  
using so I was making bad delegate calls.  The error is now gone.

- Jon

On Jun 20, 2007, at 3:51 AM, Tim Perrett wrote:

> Jon you also said you are seeing that error, what config are you
> using? If there is indeed a problem with the bridge how is it you are
> getting that error whilst I don't seem to be?
>

Re: [Rubycocoa-talk] Outline view always showing disclosure triangles

[Tim P. <he...@ti...>]

hmmm thats interesting. I dont see it do that... I am running 10.4.9,  
on the new MacBookPro. When I change it to only return false, I get  
what you would expect... you only see the root, and it has no triangle.

Laurent: the xcode project was posted onto RubyCocoa.com, here is the  
link... http://rubycocoa.com/files/outline-views/OutlineMeRuby.zip

Jon you also said you are seeing that error, what config are you  
using? If there is indeed a problem with the bridge how is it you are  
getting that error whilst I don't seem to be?

Cheers

Tim


On 20 Jun 2007, at 08:42, Dave Baldwin wrote:

> Yes and when I make the change
>
>    def outlineView_isItemExpandable(outlineView, item)
> 	return false
>    end
>
> in OLDataSource.rb there is no difference to the behaviour.
>
> My suspicion that this isn't going across the bridge correctly is
> just that it is the only thing left in the path.
> I am running Ruby 1.8.2 (default apple install) on a G4.  RubyCocoa
> was prebuilt.
>
> Dave.

Re: [Rubycocoa-talk] Core Data or not Core Data

[Tim P. <he...@ti...>]

On 19 Jun 2007, at 20:49, Pierce T. Wetter III wrote:
>
> If you have a list of objects you need to display. You can drag an
> "ArrayContoller" into your nib file, and if you tell it its an
> "Entity", it automatically fetches the data via CoreData. If you do
> the core data tutorial, they show dragging an entity into Interface
> builder which builds a lot of that stuff automatically for you.

Just a quick note about this... I understand about creating an  
ArrayController and so on in the nib, using that method instantiates  
said array controller i understand. Lets say I had an ArrayController  
called "myArray", and I wanted to do something with the objects  
inside that array from within my programming how would I do that? If  
I was working within the class that was designated as the "FileOwner"  
of the nib, would i then just be able to use @myArray to get at it?

Thanks

Tim

Re: [Rubycocoa-talk] Outline view always showing disclosure triangles

[Dave B. <dav...@3d...>]

On 19 Jun 2007, at 22:53, Tim Perrett wrote:

> Hey Dave
>
> I actually wrote that example and it appears to run ok for me? I am
> running the latest version of ruby cocoa also. The method you detail
> which is part of the delegate actually calls this on the node:
>
> return @children.count > 0 ? true : false
>
> which, should give the correct behavior - i.e. if its false that
> makes sure there is no triangle which is the default. Even when i
> added a print statement to check the output i get the correct true or
> false value. What makes you think they are not going over the bridge
> correctly?
>
> When you download the demo fresh, and just build it, do you still see
> the odd behavior you explain?


Yes and when I make the change

   def outlineView_isItemExpandable(outlineView, item)
	return false
   end

in OLDataSource.rb there is no difference to the behaviour.

My suspicion that this isn't going across the bridge correctly is  
just that it is the only thing left in the path.
I am running Ruby 1.8.2 (default apple install) on a G4.  RubyCocoa  
was prebuilt.

Dave.
>
> Thanks
>
> Tim
>
> On 19 Jun 2007, at 22:45, Dave Baldwin wrote:
>
>> RubyCocoa.com has a nice demo of outline views: http://rubycocoa.com/
>> outline-views
>>
>> The screen shot shows an expanded hierarchy with nodes with no
>> children not having the disclosure triangle, as expected.  Running
>> the demo always shows disclosure triangles for every node, even those
>> with no children.
>>
>> The return value of outlineView_isItemExpandable is always being
>> treated as if it were YES.  It looks like the Ruby true and false
>> values are not being converted to YES and NO going across the bridge
>> and trying to use OSX::NO instead just caused errors.
>>
>> I have the same error in my own app that uses outline views and I
>> have confirmed Cocoa AppKit OutlineView example that the
>> outlineView_isItemExpandable method controls the presence or absence
>> of the disclosure triangle.
>>
>> I am using RubyCocoa-0.11.1
>>
>> Dave.
>>
>>
>>
>>
>>
>> --------------------------------------------------------------------- 
>> -
>> ---
>> This SF.net email is sponsored by DB2 Express
>> Download DB2 Express C - the FREE version of DB2 express and take
>> control of your XML. No limits. Just data. Click to get it now.
>> http://sourceforge.net/powerbar/db2/
>> _______________________________________________
>> Rubycocoa-talk mailing list
>> Rub...@li...
>> https://lists.sourceforge.net/lists/listinfo/rubycocoa-talk
>>
>
>
> ---------------------------------------------------------------------- 
> ---
> This SF.net email is sponsored by DB2 Express
> Download DB2 Express C - the FREE version of DB2 express and take
> control of your XML. No limits. Just data. Click to get it now.
> http://sourceforge.net/powerbar/db2/
> _______________________________________________
> Rubycocoa-talk mailing list
> Rub...@li...
> https://lists.sourceforge.net/lists/listinfo/rubycocoa-talk

Re: [Rubycocoa-talk] Crashes with svn version

[Andrei D. <adr...@gm...>]

Yeah, maybe "using" it is a big strong. I just realised that I *don't*
really use it, and I have another question :).

Supposing that I didn't have libffi installed, how do install the one in
misc/libffi ? the build process only creates the .a, but when compiling
rubycocoa.bundle -dynamic is specified etc. I uninstalled libffi, and then,
when running I get a complaint that libffi.dylib does not exist.

I tried running the usual steps of configure etc, but i think there should
be some options, because I get errors in configure

"install-sh can't be found in ./../../"

Thanks,
  Andrei

On 6/20/07, Laurent Sansonetti <lsa...@ap...> wrote:
>
> Hi Andrei,
>
> On Jun 19, 2007, at 2:37 PM, Andrei DRAGOMIR wrote:
>
> > Hello everyone.
> >
> >
> > I have a really strange issue with rubycocoa (svn version) and
> > describe it here, in hope that someone can help me
> >
> > So, I downloaded rubycocoa from svn, config / setup / install, and
> > any sample I try crashes. When running with -d, I get something like
> > this in output:
> >
> > -----
> > > ruby -d HelloWorld.rb
> > ...
> > load of stuff that works
> > ...
> > 2007-06-19 22:00:23.359 ruby[7850] OBJWRP : ocm_closure_handler ...
> > 2007-06-19 22:00:23.359 ruby[7850] DATACNV : decoding method
> > encoding '@ 12@0:4@8' manually
> > 2007-06-19 22:00:23.359 ruby[7850] DATACNV : retval -> @
> > 2007-06-19 22:00:23.359 ruby[7850] DATACNV : skipping first two args
> > 2007-06-19 22:00:23.359 ruby[7850] DATACNV : argc -> 1
> > 2007-06-19 22:00:23.360 ruby[7850] DATACNV : arg[0] -> @
> > 2007-06-19 22:00:23.360 ruby[7850] OBJWRP : ocm_send
> > (NSBundle.bundleWithPath:): args_count=1 ret_type=@
> > 2007-06-19 22:00:23.360 ruby[7850] LIBFFI : argc expected 1 given 1
> > delta 2
> > 2007-06-19 22:00:23.360 ruby[7850] LIBFFI : arg[0] (0x8304) : @
> > 2007-06-19 22:00:23.360 ruby[7850] LIBFFI : retval (0x9fcf0037) : @
> > 2007-06-19 22:00:23.360 ruby[7850] LIBFFI : allocated 4 bytes for
> > the result
> > 2007-06-19 22:00:23.360 ruby[7850] LIBFFI : ffi_call 0x9fcc66ca with
> > 3 args
> > 2007-06-19 22:00:23.360 ruby[7850] LIBFFI : ffi_call done
> > 2007-06-19 22:00:23.360 ruby[7850] LIBFFI : getting return value
> > (0xbfffc570 of type '@')
> > 2007-06-19 22:00:23.360 ruby[7850] LIBFFI : got return value
> > 2007-06-19 22:00:23.360 ruby[7850] LIBFFI : ffi dispatch done
> > 2007-06-19 22:00:23.360 ruby[7850] OBJWRP : ocm_send
> > (bundleWithPath:) done
> > 2007-06-19 22:00: 23.360 ruby[7850] OBJWRP : ocm_closure_handler ok
> > 2007-06-19 22:00:23.360 ruby[7850] OBJWRP : ocm_closure_handler ...
> > 2007-06-19 22:00:23.360 ruby[7850] DATACNV : decoding method
> > encoding 'c8@0:4' manually
> > 2007-06-19 22:00:23.360 ruby[7850] DATACNV : retval -> c
> > 2007-06-19 22:00:23.360 ruby[7850] DATACNV : skipping first two args
> > 2007-06-19 22:00:23.360 ruby[7850] DATACNV : argc -> 0
> > 2007-06-19 22:00:23.360 ruby[7850] OBJWRP : ocm_send
> > (NSBundle#load): args_count=0 ret_type=c
> > 2007-06-19 22:00:23.360 ruby[7850] OBJWRP : found metadata description
> > 2007-06-19 22:00:23.360 ruby[7850] LIBFFI : argc expected 0 given 0
> > delta 2
> > 2007-06-19 22:00:23.360 ruby[7850] LIBFFI : detected 0 array length
> > argument(s)
> > 2007-06-19 22:00:23.361 ruby[7850] LIBFFI : coercing result from
> > octype 'c' to octype 'B'
> > 2007-06-19 22:00:23.361 ruby[7850] LIBFFI : retval (0x9fcf0037) : B
> > 2007-06-19 22:00:23.361 ruby[7850] LIBFFI : allocated 1 bytes for
> > the result
> > 2007-06-19 22:00:23.361 ruby[7850] LIBFFI : ffi_call 0x9fcf5752 with
> > 2 args
> > Illegal instruction
> > ...
> >
> > I have succeeded in tracking down the issue up to the call to
> > <#NSBundle ... >.oc_load, where the NSBundle instance is AppKit.
> >
> > This is called from Framework/Resources/ruby/osx/objc.cocoa.rb, in
> > the line : OSX.require_framework('AppKit')
> >
> > More to the point, the offending code is in oc_import.rb, in the
> > require_framework function:
> >
> > def require_framework...
> >    bundle, path = _bundle_path_for_framework(framework)
> >     bundle.oc_load
> >     if not bundle.isLoaded? then
> >       raise ArgumentError, "Can't load framework '#{framework}'"
> >     end
> >
> >
> > My configuration: MBP Core duo 2, Tiger 10.4.9, ruby 1.8.6 from
> > macports.
> >
> > I also have libffi 2.1 from macports, but I modified the pre-
> > config.rb script to use the one in /opt/local. I suspect that that
> > is where the problem is from, but the problem reproduces with both
> > (the one in misc, and the one in macports).
> >
> > Any ideas / help would be highly appreciated.
>
> Are you sure you tried using the version in misc/libffi? This is the
> first report of this kind that has been submitted yet and I suspect a
> problem in your build environment.
>
> On the other side, the libffi version in macports is different from
> the version we have, on many points, and you should always use the
> version we provide, on systems below or equal to Tiger.
>
> Thanks,
> Laurent
>
> -------------------------------------------------------------------------
> This SF.net email is sponsored by DB2 Express
> Download DB2 Express C - the FREE version of DB2 express and take
> control of your XML. No limits. Just data. Click to get it now.
> http://sourceforge.net/powerbar/db2/
> _______________________________________________
> Rubycocoa-talk mailing list
> Rub...@li...
> https://lists.sourceforge.net/lists/listinfo/rubycocoa-talk
>

Re: [Rubycocoa-talk] How to effectively debug RC errors?

[<jea...@gm...>]

On 6/19/07, Jon Baer <jo...@gm...> wrote:
>
> Anyone have tips on how to figure out what causes this?
>
> Exception raised during posting of notification.  Ignored.
> exception: *** -[NSCFDictionary setObject:forKey:]: attempt to insert
> nil value


the cocoa documentation is indispensable and entirely invaluable when
developing cocoa applications regardless of the implementation language.

this is a cocoa-level exception, the documentation for NSMutableDictionary
says:

http://developer.apple.com/documentation/Cocoa/Reference/Foundation/Classes/NSMutableDictionary_Class/Reference/Reference.html#//apple_ref/occ/instm/NSMutableDictionary/setObject:forKey
:

Raises an NSInvalidArgumentException if aKey or anObject is nil. If you need
to represent a nil value in the dictionary, use NSNull.

cheers,
jean-pierre

Re: [Rubycocoa-talk] Drag + Drop sample

[Jon B. <jo...@gm...>]

Nope, feel free to include.  Will upload more when I get a chance.

Thanks.

- Jon

On Jun 19, 2007, at 10:18 PM, Laurent Sansonetti wrote:

> Hi Jon,
>
> This is a cool example :-)
>
> Would you mind if I commit it to the official samples repository?
>
> # This would mean that it would be covered by the RubyCocoa license.
>
> Thanks,
> Laurent
>
> On Jun 19, 2007, at 10:30 AM, Jon Baer wrote:
>
>> Hi,
>>
>> I threw together a small Drag + Drop sample app, you can grab it  
>> here:
>>
>> http://jonbaer.textdriven.com/rubycocoa/SimpleDrag.zip
>>
>> Let me know if it was what you were looking for.
>>
>> (Forgive mistakes, still learning)
>>
>> - Jon
>>
>> --------------------------------------------------------------------- 
>> ----
>> This SF.net email is sponsored by DB2 Express
>> Download DB2 Express C - the FREE version of DB2 express and take
>> control of your XML. No limits. Just data. Click to get it now.
>> http://sourceforge.net/powerbar/db2/
>> _______________________________________________
>> Rubycocoa-talk mailing list
>> Rub...@li...
>> https://lists.sourceforge.net/lists/listinfo/rubycocoa-talk
>
>
> ---------------------------------------------------------------------- 
> ---
> This SF.net email is sponsored by DB2 Express
> Download DB2 Express C - the FREE version of DB2 express and take
> control of your XML. No limits. Just data. Click to get it now.
> http://sourceforge.net/powerbar/db2/
> _______________________________________________
> Rubycocoa-talk mailing list
> Rub...@li...
> https://lists.sourceforge.net/lists/listinfo/rubycocoa-talk

Re: [Rubycocoa-talk] Outline view always showing disclosure triangles

[Laurent S. <lsa...@ap...>]

Hi Jon,

Would it be possible to post here an Xcode sample project that  
reproduces the problem, so we can investigate further?

Thanks,
Laurent

On Jun 19, 2007, at 3:39 PM, Jon Baer wrote:

> Id like to second that (might be related), I have been noticing this
> error message lately:
>
> 2007-06-19 18:30:28.131 FBMenuBar[2281] Exception raised during
> posting of notification.  Ignored.  exception: *** -[NSCFDictionary
> setObject:forKey:]: attempt to insert nil value
>
> If I wrap setObject_forKey + trap it I get this:
>
> /Users/jonbaer/Desktop/FBMenuBar/build/Debug/FBMenuBar.app/Contents/
> Resources/rb_main.rb:28:in `NSApplicationMain': undefined method
> `new_with_ocid' for false:FalseClass (NoMethodError)
> 	from /Users/jonbaer/Desktop/FBMenuBar/build/Debug/FBMenuBar.app/
> Contents/Resources/rb_main.rb:28
>
> I don't know if that helps or is really part of something I did but I
> also upgraded to latest svn.  I am overriding an NSPanel and
> returning false for a few methods, is this right?
>
> - Jon
>
> On Jun 19, 2007, at 5:45 PM, Dave Baldwin wrote:
>
>> RubyCocoa.com has a nice demo of outline views: http://rubycocoa.com/
>> outline-views
>>
>> The screen shot shows an expanded hierarchy with nodes with no
>> children not having the disclosure triangle, as expected.  Running
>> the demo always shows disclosure triangles for every node, even those
>> with no children.
>>
>> The return value of outlineView_isItemExpandable is always being
>> treated as if it were YES.  It looks like the Ruby true and false
>> values are not being converted to YES and NO going across the bridge
>> and trying to use OSX::NO instead just caused errors.
>>
>> I have the same error in my own app that uses outline views and I
>> have confirmed Cocoa AppKit OutlineView example that the
>> outlineView_isItemExpandable method controls the presence or absence
>> of the disclosure triangle.
>>
>> I am using RubyCocoa-0.11.1
>>
>> Dave.
>>
>>
>>
>>
>>
>> ----------------------------------------------------------------------
>> ---
>> This SF.net email is sponsored by DB2 Express
>> Download DB2 Express C - the FREE version of DB2 express and take
>> control of your XML. No limits. Just data. Click to get it now.
>> http://sourceforge.net/powerbar/db2/
>> _______________________________________________
>> Rubycocoa-talk mailing list
>> Rub...@li...
>> https://lists.sourceforge.net/lists/listinfo/rubycocoa-talk
>
>
> -------------------------------------------------------------------------
> This SF.net email is sponsored by DB2 Express
> Download DB2 Express C - the FREE version of DB2 express and take
> control of your XML. No limits. Just data. Click to get it now.
> http://sourceforge.net/powerbar/db2/
> _______________________________________________
> Rubycocoa-talk mailing list
> Rub...@li...
> https://lists.sourceforge.net/lists/listinfo/rubycocoa-talk

Re: [Rubycocoa-talk] Drag + Drop sample

[Laurent S. <lsa...@ap...>]

Hi Jon,

This is a cool example :-)

Would you mind if I commit it to the official samples repository?

# This would mean that it would be covered by the RubyCocoa license.

Thanks,
Laurent

On Jun 19, 2007, at 10:30 AM, Jon Baer wrote:

> Hi,
>
> I threw together a small Drag + Drop sample app, you can grab it here:
>
> http://jonbaer.textdriven.com/rubycocoa/SimpleDrag.zip
>
> Let me know if it was what you were looking for.
>
> (Forgive mistakes, still learning)
>
> - Jon
>
> -------------------------------------------------------------------------
> This SF.net email is sponsored by DB2 Express
> Download DB2 Express C - the FREE version of DB2 express and take
> control of your XML. No limits. Just data. Click to get it now.
> http://sourceforge.net/powerbar/db2/
> _______________________________________________
> Rubycocoa-talk mailing list
> Rub...@li...
> https://lists.sourceforge.net/lists/listinfo/rubycocoa-talk

Re: [Rubycocoa-talk] Crashes with svn version

[Laurent S. <lsa...@ap...>]

Hi Andrei,

On Jun 19, 2007, at 2:37 PM, Andrei DRAGOMIR wrote:

> Hello everyone.
>
>
> I have a really strange issue with rubycocoa (svn version) and  
> describe it here, in hope that someone can help me
>
> So, I downloaded rubycocoa from svn, config / setup / install, and  
> any sample I try crashes. When running with -d, I get something like  
> this in output:
>
> -----
> > ruby -d HelloWorld.rb
> ...
> load of stuff that works
> ...
> 2007-06-19 22:00:23.359 ruby[7850] OBJWRP : ocm_closure_handler ...
> 2007-06-19 22:00:23.359 ruby[7850] DATACNV : decoding method  
> encoding '@ 12@0:4@8' manually
> 2007-06-19 22:00:23.359 ruby[7850] DATACNV : retval -> @
> 2007-06-19 22:00:23.359 ruby[7850] DATACNV : skipping first two args
> 2007-06-19 22:00:23.359 ruby[7850] DATACNV : argc -> 1
> 2007-06-19 22:00:23.360 ruby[7850] DATACNV : arg[0] -> @
> 2007-06-19 22:00:23.360 ruby[7850] OBJWRP : ocm_send  
> (NSBundle.bundleWithPath:): args_count=1 ret_type=@
> 2007-06-19 22:00:23.360 ruby[7850] LIBFFI : argc expected 1 given 1  
> delta 2
> 2007-06-19 22:00:23.360 ruby[7850] LIBFFI : arg[0] (0x8304) : @
> 2007-06-19 22:00:23.360 ruby[7850] LIBFFI : retval (0x9fcf0037) : @
> 2007-06-19 22:00:23.360 ruby[7850] LIBFFI : allocated 4 bytes for  
> the result
> 2007-06-19 22:00:23.360 ruby[7850] LIBFFI : ffi_call 0x9fcc66ca with  
> 3 args
> 2007-06-19 22:00:23.360 ruby[7850] LIBFFI : ffi_call done
> 2007-06-19 22:00:23.360 ruby[7850] LIBFFI : getting return value  
> (0xbfffc570 of type '@')
> 2007-06-19 22:00:23.360 ruby[7850] LIBFFI : got return value
> 2007-06-19 22:00:23.360 ruby[7850] LIBFFI : ffi dispatch done
> 2007-06-19 22:00:23.360 ruby[7850] OBJWRP : ocm_send  
> (bundleWithPath:) done
> 2007-06-19 22:00: 23.360 ruby[7850] OBJWRP : ocm_closure_handler ok
> 2007-06-19 22:00:23.360 ruby[7850] OBJWRP : ocm_closure_handler ...
> 2007-06-19 22:00:23.360 ruby[7850] DATACNV : decoding method  
> encoding 'c8@0:4' manually
> 2007-06-19 22:00:23.360 ruby[7850] DATACNV : retval -> c
> 2007-06-19 22:00:23.360 ruby[7850] DATACNV : skipping first two args
> 2007-06-19 22:00:23.360 ruby[7850] DATACNV : argc -> 0
> 2007-06-19 22:00:23.360 ruby[7850] OBJWRP : ocm_send  
> (NSBundle#load): args_count=0 ret_type=c
> 2007-06-19 22:00:23.360 ruby[7850] OBJWRP : found metadata description
> 2007-06-19 22:00:23.360 ruby[7850] LIBFFI : argc expected 0 given 0  
> delta 2
> 2007-06-19 22:00:23.360 ruby[7850] LIBFFI : detected 0 array length  
> argument(s)
> 2007-06-19 22:00:23.361 ruby[7850] LIBFFI : coercing result from  
> octype 'c' to octype 'B'
> 2007-06-19 22:00:23.361 ruby[7850] LIBFFI : retval (0x9fcf0037) : B
> 2007-06-19 22:00:23.361 ruby[7850] LIBFFI : allocated 1 bytes for  
> the result
> 2007-06-19 22:00:23.361 ruby[7850] LIBFFI : ffi_call 0x9fcf5752 with  
> 2 args
> Illegal instruction
> ...
>
> I have succeeded in tracking down the issue up to the call to  
> <#NSBundle ... >.oc_load, where the NSBundle instance is AppKit.
>
> This is called from Framework/Resources/ruby/osx/objc.cocoa.rb, in  
> the line : OSX.require_framework('AppKit')
>
> More to the point, the offending code is in oc_import.rb, in the  
> require_framework function:
>
> def require_framework...
>    bundle, path = _bundle_path_for_framework(framework)
>     bundle.oc_load
>     if not bundle.isLoaded? then
>       raise ArgumentError, "Can't load framework '#{framework}'"
>     end
>
>
> My configuration: MBP Core duo 2, Tiger 10.4.9, ruby 1.8.6 from  
> macports.
>
> I also have libffi 2.1 from macports, but I modified the pre- 
> config.rb script to use the one in /opt/local. I suspect that that  
> is where the problem is from, but the problem reproduces with both  
> (the one in misc, and the one in macports).
>
> Any ideas / help would be highly appreciated.

Are you sure you tried using the version in misc/libffi? This is the  
first report of this kind that has been submitted yet and I suspect a  
problem in your build environment.

On the other side, the libffi version in macports is different from  
the version we have, on many points, and you should always use the  
version we provide, on systems below or equal to Tiger.

Thanks,
Laurent

Re: [Rubycocoa-talk] Outline view always showing disclosure triangles

[Jon B. <jo...@gm...>]

Id like to second that (might be related), I have been noticing this  
error message lately:

2007-06-19 18:30:28.131 FBMenuBar[2281] Exception raised during  
posting of notification.  Ignored.  exception: *** -[NSCFDictionary  
setObject:forKey:]: attempt to insert nil value

If I wrap setObject_forKey + trap it I get this:

/Users/jonbaer/Desktop/FBMenuBar/build/Debug/FBMenuBar.app/Contents/ 
Resources/rb_main.rb:28:in `NSApplicationMain': undefined method  
`new_with_ocid' for false:FalseClass (NoMethodError)
	from /Users/jonbaer/Desktop/FBMenuBar/build/Debug/FBMenuBar.app/ 
Contents/Resources/rb_main.rb:28

I don't know if that helps or is really part of something I did but I  
also upgraded to latest svn.  I am overriding an NSPanel and  
returning false for a few methods, is this right?

- Jon

On Jun 19, 2007, at 5:45 PM, Dave Baldwin wrote:

> RubyCocoa.com has a nice demo of outline views: http://rubycocoa.com/
> outline-views
>
> The screen shot shows an expanded hierarchy with nodes with no
> children not having the disclosure triangle, as expected.  Running
> the demo always shows disclosure triangles for every node, even those
> with no children.
>
> The return value of outlineView_isItemExpandable is always being
> treated as if it were YES.  It looks like the Ruby true and false
> values are not being converted to YES and NO going across the bridge
> and trying to use OSX::NO instead just caused errors.
>
> I have the same error in my own app that uses outline views and I
> have confirmed Cocoa AppKit OutlineView example that the
> outlineView_isItemExpandable method controls the presence or absence
> of the disclosure triangle.
>
> I am using RubyCocoa-0.11.1
>
> Dave.
>
>
>
>
>
> ---------------------------------------------------------------------- 
> ---
> This SF.net email is sponsored by DB2 Express
> Download DB2 Express C - the FREE version of DB2 express and take
> control of your XML. No limits. Just data. Click to get it now.
> http://sourceforge.net/powerbar/db2/
> _______________________________________________
> Rubycocoa-talk mailing list
> Rub...@li...
> https://lists.sourceforge.net/lists/listinfo/rubycocoa-talk

Re: [Rubycocoa-talk] Drag + Drop sample

[Sal S. <sal...@gm...>]

thanks that was exactly what I was looking for

-Sal

Re: [Rubycocoa-talk] Outline view always showing disclosure triangles

[Tim P. <he...@ti...>]

Hey Dave

I actually wrote that example and it appears to run ok for me? I am  
running the latest version of ruby cocoa also. The method you detail  
which is part of the delegate actually calls this on the node:

return @children.count > 0 ? true : false

which, should give the correct behavior - i.e. if its false that  
makes sure there is no triangle which is the default. Even when i  
added a print statement to check the output i get the correct true or  
false value. What makes you think they are not going over the bridge  
correctly?

When you download the demo fresh, and just build it, do you still see  
the odd behavior you explain?

Thanks

Tim

On 19 Jun 2007, at 22:45, Dave Baldwin wrote:

> RubyCocoa.com has a nice demo of outline views: http://rubycocoa.com/
> outline-views
>
> The screen shot shows an expanded hierarchy with nodes with no
> children not having the disclosure triangle, as expected.  Running
> the demo always shows disclosure triangles for every node, even those
> with no children.
>
> The return value of outlineView_isItemExpandable is always being
> treated as if it were YES.  It looks like the Ruby true and false
> values are not being converted to YES and NO going across the bridge
> and trying to use OSX::NO instead just caused errors.
>
> I have the same error in my own app that uses outline views and I
> have confirmed Cocoa AppKit OutlineView example that the
> outlineView_isItemExpandable method controls the presence or absence
> of the disclosure triangle.
>
> I am using RubyCocoa-0.11.1
>
> Dave.
>
>
>
>
>
> ---------------------------------------------------------------------- 
> ---
> This SF.net email is sponsored by DB2 Express
> Download DB2 Express C - the FREE version of DB2 express and take
> control of your XML. No limits. Just data. Click to get it now.
> http://sourceforge.net/powerbar/db2/
> _______________________________________________
> Rubycocoa-talk mailing list
> Rub...@li...
> https://lists.sourceforge.net/lists/listinfo/rubycocoa-talk
>

Re: [Rubycocoa-talk] Core Data or not Core Data

[Tim P. <he...@ti...>]

For sure I totally agree - it can be very difficult to keep it all  
black and white, especially when there are a lot of IB outlets  
kicking about I tend to find :)

Thanks for your advice, i'll make the changes and let the list know  
how it goes!!!

All the best

Tim



On 19 Jun 2007, at 20:49, Pierce T. Wetter III wrote:

> But its rarely so cut and dried as to where to put things because
> MVC rarely maps well to precise divisions.

[Rubycocoa-talk] Outline view always showing disclosure triangles

[Dave B. <dav...@3d...>]

RubyCocoa.com has a nice demo of outline views: http://rubycocoa.com/ 
outline-views

The screen shot shows an expanded hierarchy with nodes with no  
children not having the disclosure triangle, as expected.  Running  
the demo always shows disclosure triangles for every node, even those  
with no children.

The return value of outlineView_isItemExpandable is always being  
treated as if it were YES.  It looks like the Ruby true and false  
values are not being converted to YES and NO going across the bridge  
and trying to use OSX::NO instead just caused errors.

I have the same error in my own app that uses outline views and I  
have confirmed Cocoa AppKit OutlineView example that the  
outlineView_isItemExpandable method controls the presence or absence  
of the disclosure triangle.

I am using RubyCocoa-0.11.1

Dave.

[Rubycocoa-talk] Crashes with svn version

[Andrei D. <adr...@gm...>]

Hello everyone.


I have a really strange issue with rubycocoa (svn version) and describe it
here, in hope that someone can help me

So, I downloaded rubycocoa from svn, config / setup / install, and any
sample I try crashes. When running with -d, I get something like this in
output:

-----
> ruby -d HelloWorld.rb
...
load of stuff that works
...
2007-06-19 22:00:23.359 ruby[7850] OBJWRP : ocm_closure_handler ...
2007-06-19 22:00:23.359 ruby[7850] DATACNV : decoding method encoding '@
12@0:4@8' manually
2007-06-19 22:00:23.359 ruby[7850] DATACNV : retval -> @
2007-06-19 22:00:23.359 ruby[7850] DATACNV : skipping first two args
2007-06-19 22:00:23.359 ruby[7850] DATACNV : argc -> 1
2007-06-19 22:00:23.360 ruby[7850] DATACNV : arg[0] -> @
2007-06-19 22:00:23.360 ruby[7850] OBJWRP : ocm_send (
NSBundle.bundleWithPath:): args_count=1 ret_type=@
2007-06-19 22:00:23.360 ruby[7850] LIBFFI : argc expected 1 given 1 delta 2
2007-06-19 22:00:23.360 ruby[7850] LIBFFI : arg[0] (0x8304) : @
2007-06-19 22:00:23.360 ruby[7850] LIBFFI : retval (0x9fcf0037) : @
2007-06-19 22:00:23.360 ruby[7850] LIBFFI : allocated 4 bytes for the result
2007-06-19 22:00:23.360 ruby[7850] LIBFFI : ffi_call 0x9fcc66ca with 3 args
2007-06-19 22:00:23.360 ruby[7850] LIBFFI : ffi_call done
2007-06-19 22:00:23.360 ruby[7850] LIBFFI : getting return value (0xbfffc570
of type '@')
2007-06-19 22:00:23.360 ruby[7850] LIBFFI : got return value
2007-06-19 22:00:23.360 ruby[7850] LIBFFI : ffi dispatch done
2007-06-19 22:00:23.360 ruby[7850] OBJWRP : ocm_send (bundleWithPath:) done
2007-06-19 22:00: 23.360 ruby[7850] OBJWRP : ocm_closure_handler ok
2007-06-19 22:00:23.360 ruby[7850] OBJWRP : ocm_closure_handler ...
2007-06-19 22:00:23.360 ruby[7850] DATACNV : decoding method encoding 'c8@0:4'
manually
2007-06-19 22:00:23.360 ruby[7850] DATACNV : retval -> c
2007-06-19 22:00:23.360 ruby[7850] DATACNV : skipping first two args
2007-06-19 22:00:23.360 ruby[7850] DATACNV : argc -> 0
2007-06-19 22:00:23.360 ruby[7850] OBJWRP : ocm_send (NSBundle#load):
args_count=0 ret_type=c
2007-06-19 22:00:23.360 ruby[7850] OBJWRP : found metadata description
2007-06-19 22:00:23.360 ruby[7850] LIBFFI : argc expected 0 given 0 delta 2
2007-06-19 22:00:23.360 ruby[7850] LIBFFI : detected 0 array length
argument(s)
2007-06-19 22:00:23.361 ruby[7850] LIBFFI : coercing result from octype 'c'
to octype 'B'
2007-06-19 22:00:23.361 ruby[7850] LIBFFI : retval (0x9fcf0037) : B
2007-06-19 22:00:23.361 ruby[7850] LIBFFI : allocated 1 bytes for the result

2007-06-19 22:00:23.361 ruby[7850] LIBFFI : ffi_call 0x9fcf5752 with 2 args
Illegal instruction
...

I have succeeded in tracking down the issue up to the call to <#NSBundle ...
>.oc_load, where the NSBundle instance is AppKit.

This is called from Framework/Resources/ruby/osx/objc.cocoa.rb, in the line
: OSX.require_framework('AppKit')

More to the point, the offending code is in oc_import.rb, in the
require_framework function:

def require_framework...
   bundle, path = _bundle_path_for_framework(framework)
    bundle.oc_load
    if not bundle.isLoaded? then
      raise ArgumentError, "Can't load framework '#{framework}'"
    end


My configuration: MBP Core duo 2, Tiger 10.4.9, ruby 1.8.6 from macports.

I also have libffi 2.1 from macports, but I modified the
pre-config.rbscript to use the one in /opt/local. I suspect that that
is where the
problem is from, but the problem reproduces with both (the one in misc, and
the one in macports).

Any ideas / help would be highly appreciated.

Thanks in advance
  Andrei

Re: [Rubycocoa-talk] Core Data or not Core Data

[Pierce T. W. I. <pi...@tw...>]

On Jun 19, 2007, at 11:37 AM, Tim Perrett wrote:

> Thanks Pierce, thats really helpfull
>
> I think i will use coredata for this project I am doing as it seems
> to fit the bill!
>
> Otherwise, when your choosing how to set up the coredata application,
> I am getting a little conufussed with is how to best arrange my
> application, and also how to reference coredata arrays that we
> instansiate in the nib files in the ruby code? is it by name?

If you have a list of objects you need to display. You can drag an  
"ArrayContoller" into your nib file, and if you tell it its an  
"Entity", it automatically fetches the data via CoreData. If you do  
the core data tutorial, they show dragging an entity into Interface  
builder which builds a lot of that stuff automatically for you.

  Basic idea behind CoreData & bindings is that you bind things by a  
name.

  Say you have a really simple interface: a checkbox.

  Let's say the value of the checkbox is whether or not to explode.

  Now normally to implement the UI, you would have to:

    1. Setup the checkbox with the value of the "shouldExplode"  
variable.

    2. Monitor when the user clicked on the checkbox. Usually you  
would map this to an action, then in the action, you would have to  
pull the new value of the checkbox, then update the variable.

    3. Re-setup the checkbox everytime you change shouldExplode.

    What bindings do is let you bind the checkbox to the value of  
"shouldExplode" on some object. Then the UI will call shouldExplode()  
to get the value, and call setShouldExplode to set the value.

    The binding stuff can be chained, so if the shouldExplode variable  
is 5 objects deep, and you bind to:

    object1.object2.object3.object4.object5.shouldExplode, that's just  
the same thing as:

     object1().object2().object3().object4().object5().shouldExplode()

   That is, Key-Value-Coding is really just a convention on how to  
find stuff, plus some optimizations behind the scenes to be really fast.

   Key Value Observing is built on top of KVC to let you see when  
certain things change, which is how bindings work. So if you bind a  
checkbox to shouldExplode, and you change that value in code, the  
checkbox knows to update. Conversely when the checkbox changes, it  
knows to update the value.

    It gets more complicated when you have an array of things and a  
table view showing a scrolling list of checkboxes, but in that case,  
you're effectively putting a for loop in there. The binding is a bit  
weird to me because you bind at the column level down. So the binding  
will look something like:

   arrayController.arrangedObjects.shouldExplode

  Which tells a table view: arrangedObjects has a list of objects,  
bind the checkbox value to shouldExplode for all of them.

>
>
> The model holds the data mapping stuff, so thats fine I understand
> how that works with the gui. Then I am going to have a controller to
> actually execute the actions (probably with a managed object alias
> like in Eloys screen cast for then accessing the other array) and
> feed back to the gui with the relevant outlets. Is it ok to have a
> WindowController subclass do this? So it would be the nibs owner as
> well.

    In Frictionless ActionBrowser is just a regular NSObject. I  
suppose it could have been a WindowController subclass (and I would  
have gotten some default behavior I probably need) but its not  
required. I have it as the nibs owner and its also the delegate for a  
number of views. I'll probably have to separate some of the delegate  
functionality out into other objects.

> I know that all sounds so vauge, I am new to writing cocoa
> applications so I guess this is more of a general question on how do
> you arrange things when you start having a various windows and
> objects kicking around. For the most part I presume that the bulk of
> the code for actually making 'things work' would go in that windows
> WindowControler?

   That would keep in with the MVC model, yes.

   But its rarely so cut and dried as to where to put things because  
MVC rarely maps well to precise divisions.

  Pierce

Re: [Rubycocoa-talk] Core Data or not Core Data

[Tim P. <he...@ti...>]

Thanks Pierce, thats really helpfull

I think i will use coredata for this project I am doing as it seems  
to fit the bill!

Otherwise, when your choosing how to set up the coredata application,  
I am getting a little conufussed with is how to best arrange my  
application, and also how to reference coredata arrays that we  
instansiate in the nib files in the ruby code? is it by name?

The model holds the data mapping stuff, so thats fine I understand  
how that works with the gui. Then I am going to have a controller to  
actually execute the actions (probably with a managed object alias  
like in Eloys screen cast for then accessing the other array) and  
feed back to the gui with the relevant outlets. Is it ok to have a  
WindowController subclass do this? So it would be the nibs owner as  
well.

I know that all sounds so vauge, I am new to writing cocoa  
applications so I guess this is more of a general question on how do  
you arrange things when you start having a various windows and  
objects kicking around. For the most part I presume that the bulk of  
the code for actually making 'things work' would go in that windows  
WindowControler?

Any advice is very much welcomed!!!

Tim