rubycocoa-talk — talking about RubyCocoa

Re: [Rubycocoa-talk] New ways to not install Ruby things.

[Dave H. <gr...@gr...>]

On Dec 8, 2005, at 4:00, Jonathan Paisley wrote:

>
> Try removing /Library/Frameworks/RubyCocoa.framework altogether and 
> rerunning the source install.

sudo rm  -R /Library/Frameworks/RubyCocoa.framework
sudo ruby install.rb install

Attempt to run "test"
"can't open library: /usr/lib/libruby.1.dylib  (No such file or 
directory, errno = 2)"

[Rubycocoa-talk] assertion failure when using nstableview

[Steven D.A. <st...@ne...>]

I am using rubycocoa version 0.4.2 and ruby 1.8.4 in a dictionary  
application I'm writing.  As part of the dictionary, each word has a  
number of synonyms with which it may be associated; these are  
presented in an NSTableView.  I have a button that allows a user to  
add a row to the table, where they can then type in the word.

Since I have several tables like this, instead of implementing each  
one separately, since the logic is very similar, I have defined a  
generic function for adding a row to a table; supply the attribute  
name as a string and we get the attribute and then apply the logic to  
add the new row.  The code looks like this:

01    def addGeneric(sender, attribute)
02        # Edit a single-column tableview.
03        attribute = self.instance_variable_get("@#{attribute}")
04        num_rows = attribute.numberOfRows - 1
05        idx_set = NSIndexSet.indexSetWithIndex(num_rows)
06        attribute.selectRowIndexes_byExtendingSelection(idx_set, true)
07        attribute.editColumn_row_withEvent_select(0, num_rows, nil,  
true)
08    end

This code selects the last row in the tableview.  For example, if  
there are three rows containing values "moe", "curly" and "larry",  
respectively, then "larry" would be selected.  If you typed  
something, "larry" would be overwritten.  This isn't quite what I  
want.  I want the _next_ row to become editable, allowing the user to  
then type in a new value that wasn't previously in the list.

Given this, I assume that it will work if I just remove the "-1" on  
line 4.  But when I test this, when I click the add button, I get  
this error:

2005-12-09 01:44:38.390 rubydict[7485] *** Assertion failure in - 
[NSTableView editColumn:row:withEvent:select:], TableView.subproj/ 
NSTableView.m:3855
/Library/Frameworks/RubyCocoa.framework/Versions/A/Resources/ruby/osx/ 
objc/oc_wrapper.rb:17:in `NSApplicationMain': NSApplicationMain -  
NSInternalInconsistencyException - Invalid parameter not satisfying:  
_lastSelectedRow == row (OSX::OCException)
	from /Volumes/secure_fw/Documents/Personal/Development/ 
new_dictionary/rubydict/build/Development/rubydict.app/Contents/ 
Resources/rb_main.rb:23

Any idea what might be going on here?  BTW, very similar logic works  
in Python using PyObjC.

Thanks in advance,
steve

Re: [Rubycocoa-talk] New ways to not install Ruby things.

[kimura wataru<ki...@us...>]

Thanks for your reporting. 

I wrote a patch to fix this problem.

----------
Index: install.rb
===================================================================
RCS file: /cvsroot/rubycocoa/src/install.rb,v
retrieving revision 1.17
diff -u -r1.17 install.rb
--- install.rb	2 Nov 2005 14:10:15 -0000	1.17
+++ install.rb	8 Dec 2005 13:58:57 -0000
@@ -563,7 +563,7 @@
 
   def extconf
     opt = @options['config-opt'].join(' ')
-    command "#{config('ruby-prog')} #{curr_srcdir}/extconf.rb #{opt}"
+    command "#{config('ruby-prog')} \"#{curr_srcdir}/extconf.rb\" #{opt}"
   end
 
   def config_dir_data( rel )
----------

Wed, Dec 7, 2005 5:58:37 PM, Dave Howell wrote:
>
>On Dec 7, 2005, at 17:50, Dave Howell wrote:
>
>> I guess I'll try compiling from source . . . .
>
>"ruby install.rb config" failed. It can't handles spaces in the path.
>
>I guess I'll have to relocate it to some directory that doesn't have 
>any spaces in its path . . .
>
>


-- 
kimura wataru

Re: [Rubycocoa-talk] New ways to not install Ruby things.

[Jonathan P. <jp...@dc...>]

> As my earlier messages mentioned, I'm using Panther (as it happens,  
> it's the current release, so it's 10.3.9), the error message is  
> "can't open library:

My apologies. I hadn't realised that there were now separate dmg  
distributions for Panther and Tiger.

> /usr/lib/libruby.1.dylib  (No such file or directory, errno = 2)",  
> I'm using 1.8.2, which means I can't be using the "standard Ruby  
> installation," and indeed, if I hadn't completely obliterated it  
> from my OS (so, yes, I've removed bits. All the bits. Every last  
> bit I could find), it would be happily compiling the *wrong*  
> version of Ruby into my code and making my life incredibly miserable.

Can you confirm that the ruby installation picked up from the command  
line is your self-installed 1.8.2?

    which ruby
    ruby -v

> What steps did I take? I hit "Build and Run." For all of my current  
> projects, even ones that have not changed at all since they last  
> compiled successfully, I now get the same error message.

Can you check that your self-compiled RubyCocoa has definitely been  
installed in /Library/Frameworks? It sounds like perhaps the one from  
the dmg is still there.

Try removing /Library/Frameworks/RubyCocoa.framework altogether and  
rerunning the source install.

> The question is: What did the pre-built Panther Install package  
> screw up that the "ruby install.rb" script wasn't smart enough to  
> fix? "libruby.1.dylib" is installed into /usr/local/lib. How do I  
> force XCode to look where the library really is, and why didn't  
> RubyCocoa's "install.rb" file set it up correctly in the first  
> place? This was all working with RubyCocoa 0.4.1. Now it's not.

The reference to /usr/lib/libruby.1.dylib is probably coming from the  
RubyCocoa framework. You can check with the 'otool -L' command, which  
shows library dependencies (the example below is for the 0.4.2  
Panther dmg):

   otool -L /Library/Frameworks/RubyCocoa.framework/RubyCocoa
../RubyCocoa:
	@executable_path/../Frameworks/RubyCocoa.framework/Versions/A/ 
RubyCocoa (compatibility version 1.0.0, current version 1.0.0)
	/System/Library/Frameworks/Cocoa.framework/Versions/A/Cocoa  
(compatibility version 1.0.0, current version 9.0.0)
	/usr/lib/libruby.1.dylib (compatibility version 1.6.0, current  
version 1.6.0)
	/usr/lib/libSystem.B.dylib (compatibility version 1.0.0, current  
version 71.1.4)

Re: [Rubycocoa-talk] New ways to not install Ruby things.

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

On 8 Dec 2005, at 02:18, Dave Howell wrote:

>
> On Dec 7, 2005, at 17:58, Dave Howell wrote:
>
>> "ruby install.rb config" failed. It can't handles spaces in the path.
>>
>> I guess I'll have to relocate it to some directory that doesn't  
>> have any spaces in its path . . .
>
> Got it installed.
>
> Still can't run a Ruby app from XCode, though:
>
> can't open library: /usr/lib/libruby.1.dylib  (No such file or  
> directory, errno = 2)
>
> Well, no kidding. That's because it's in /usr/local/lib.
>
> Quit and restart XCode. Nope.
>
> Log out and in. Nope.
>
> Reboot. Nope.


Have you tried running from terminal?

try:
which ruby
ruby -v

have you setup your shell path environment to include /usr/local/bin?

Dave.

>
>
>
> OK, so what did the Panther Install package screw up that the "ruby  
> install.rb" script wasn't smart enough to fix?
>
>
>
>
>
>
> -------------------------------------------------------
> This SF.net email is sponsored by: Splunk Inc. Do you grep through  
> log files
> for problems?  Stop!  Download the new AJAX search engine that makes
> searching your log files as easy as surfing the  web.  DOWNLOAD  
> SPLUNK!
> http://ads.osdn.com/?ad_id=7637&alloc_id=16865&op=click
> _______________________________________________
> Rubycocoa-talk mailing list
> Rub...@li...
> https://lists.sourceforge.net/lists/listinfo/rubycocoa-talk

Re: [Rubycocoa-talk] New ways to not install Ruby things.

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

On 8 Dec 2005, at 01:50, Dave Howell wrote:

> With the announcement of 0.4.2 I thought I might be able to move  
> forward on some Ruby projects.
>
> Go to web site. Go to download page. Download Panther installer.  
> Install. Watch Ruby fail to compile.
>

I thought 0.4.2 included access to tiger only functionality (i.e.  
core data).  Don't know what it will do when it cannot find this.
May have nothing to do with your problem...

Dave.

> The download page says "Before downloading, you may want to read  
> the Release Notes and ChangeLog." There do not appear to be any  
> "Release Notes" on that page, nor any obvious links to same.
>
>
> If I instead do
>
> Go to web site. Follow link to "main page." Follow link to "Getting  
> RubyCocoa", then I see a notice that the binary distributions are  
> for ruby 1.6.8. I'm using ruby 1.8.2. I'll just guess that's why it  
> can't compile.
>
> Yes? No?
>
> I guess I'll try compiling from source . . . .
>
>
>
> -------------------------------------------------------
> This SF.net email is sponsored by: Splunk Inc. Do you grep through  
> log files
> for problems?  Stop!  Download the new AJAX search engine that makes
> searching your log files as easy as surfing the  web.  DOWNLOAD  
> SPLUNK!
> http://ads.osdn.com/?ad_id=7637&alloc_id=16865&op=click
> _______________________________________________
> Rubycocoa-talk mailing list
> Rub...@li...
> https://lists.sourceforge.net/lists/listinfo/rubycocoa-talk

Re: [Rubycocoa-talk] New ways to not install Ruby things.

[Dave H. <gr...@gr...>]

On Dec 8, 2005, at 0:36, Jonathan Paisley wrote:

> On 8 Dec 2005, at 1:50, Dave Howell wrote:
>
>> Go to web site. Go to download page. Download Panther installer. 
>> Install. Watch Ruby fail to compile.
>
> Please give precise details of the failure you experienced. What steps 
> did you take, and what results did you get? "fail to compile" isn't 
> enough information to help diagnose the problem.
>
> What version of OS X are you using? Have you done anything to the 
> standard Ruby installation (e.g., removed bits of it)?

As my earlier messages mentioned, I'm using Panther (as it happens, 
it's the current release, so it's 10.3.9), the error message is "can't 
open library: /usr/lib/libruby.1.dylib  (No such file or directory, 
errno = 2)", I'm using 1.8.2, which means I can't be using the 
"standard Ruby installation," and indeed, if I hadn't completely 
obliterated it from my OS (so, yes, I've removed bits. All the bits. 
Every last bit I could find), it would be happily compiling the *wrong* 
version of Ruby into my code and making my life incredibly miserable.

What steps did I take? I hit "Build and Run." For all of my current 
projects, even ones that have not changed at all since they last 
compiled successfully, I now get the same error message.

The question is: What did the pre-built Panther Install package screw 
up that the "ruby install.rb" script wasn't smart enough to fix? 
"libruby.1.dylib" is installed into /usr/local/lib. How do I force 
XCode to look where the library really is, and why didn't RubyCocoa's 
"install.rb" file set it up correctly in the first place? This was all 
working with RubyCocoa 0.4.1. Now it's not.

I could reset the files by restoring from my backups, but I rather 
doubt that the file containing the corrupted path is in the /usr tree, 
and I can't roll back my entire hard drive two days.

Re: [Rubycocoa-talk] New ways to not install Ruby things.

[Jonathan P. <jp...@dc...>]

On 8 Dec 2005, at 1:50, Dave Howell wrote:

> Go to web site. Go to download page. Download Panther installer.  
> Install. Watch Ruby fail to compile.

Please give precise details of the failure you experienced. What  
steps did you take, and what results did you get? "fail to compile"  
isn't enough information to help diagnose the problem.

What version of OS X are you using? Have you done anything to the  
standard Ruby installation (e.g., removed bits of it)?

Re: [Rubycocoa-talk] New ways to not install Ruby things.

[Dave H. <gr...@gr...>]

On Dec 7, 2005, at 17:58, Dave Howell wrote:

> "ruby install.rb config" failed. It can't handles spaces in the path.
>
> I guess I'll have to relocate it to some directory that doesn't have 
> any spaces in its path . . .

Got it installed.

Still can't run a Ruby app from XCode, though:

can't open library: /usr/lib/libruby.1.dylib  (No such file or 
directory, errno = 2)

Well, no kidding. That's because it's in /usr/local/lib.

Quit and restart XCode. Nope.

Log out and in. Nope.

Reboot. Nope.



OK, so what did the Panther Install package screw up that the "ruby 
install.rb" script wasn't smart enough to fix?

Re: [Rubycocoa-talk] New ways to not install Ruby things.

[Dave H. <gr...@gr...>]

On Dec 7, 2005, at 17:50, Dave Howell wrote:

> I guess I'll try compiling from source . . . .

"ruby install.rb config" failed. It can't handles spaces in the path.

I guess I'll have to relocate it to some directory that doesn't have 
any spaces in its path . . .





This is why I didn't install RubyCocoa 0.4.2 as soon as it was 
released; I don't recall ever having installed *any* Ruby-related 
software without having to to deal with multiple problems during the 
install. Sigh.

Re: [Rubycocoa-talk] New ways to not install Ruby things.

[Dave H. <gr...@gr...>]

With the announcement of 0.4.2 I thought I might be able to move 
forward on some Ruby projects.

Go to web site. Go to download page. Download Panther installer. 
Install. Watch Ruby fail to compile.

The download page says "Before downloading, you may want to read the 
Release Notes and ChangeLog." There do not appear to be any "Release 
Notes" on that page, nor any obvious links to same.


If I instead do

Go to web site. Follow link to "main page." Follow link to "Getting 
RubyCocoa", then I see a notice that the binary distributions are for 
ruby 1.6.8. I'm using ruby 1.8.2. I'll just guess that's why it can't 
compile.

Yes? No?

I guess I'll try compiling from source . . . .

Re: [Rubycocoa-talk] Re: SyncServices and RubyCocoa

[Jonathan P. <jp...@dc...>]

On 5 Dec 2005, at 22:53, Rupert BARROW wrote:

> Funnily enough, I am fiddling with Apple's SyncServices tutorial :  
> when their Objective-C example sets @"YES", the == test works well,  
> whereas when I set "YES" from Ruby, the == test fails; however the  
> isEqualString succeeds !

The string literal @"YES" in the Objective C source will get interned  
by the linker (and so is the very same object).

The conversion from a ruby string "YES" to Objective C involves the  
creation of a new NSString object, so it'll never match up.

Re: [Rubycocoa-talk] Re: SyncServices and RubyCocoa

[Rupert B. <rup...@fr...>]

Jonathan,

Le 11 nov. 05 =C3=A0 12:51, Jonathan Paisley a =C3=A9crit :

> On 11 Nov 2005, at 10:19, Rupert BARROW wrote:
>
>> 		if ([record valueForKey:@"changed"] =3D=3D @"YES")		=
	// this crashes
>
> This is unrelated to the main issue, but note that comparison =20
> doesn't make sense.
>
> You need to compare NSStrings with isEqualToString:

You're right again !
Funnily enough, I am fiddling with Apple's SyncServices tutorial : =20
when their Objective-C example sets @"YES", the =3D=3D test works well, =20=

whereas when I set "YES" from Ruby, the =3D=3D test fails; however the =20=

isEqualString succeeds !
I hope to post Ruby-fied samples soon.

Ciao,
Rup

PS: pardon all these "stupid" questions : patching pieces of ObjC and =20=

Ruby together seems to be one of the causes of my difficulties, the =20
other being my inexperience in these matters. However I am making =20
progress, and the result is looking good.
One side effect of all this experimentation is that I have totally =20
thrashed my Address Book and phone contacts, as well as =20
SyncroSpector : names have disappeared, only company names appear (I =20
think !) Rather funny, actually, although I am still stuck with =20
usuless contact in my phone :-)((   Restoring backups of Address Book =20=

is of no avail, and deleting ~/Library...SyncServices stuff does =20
absolutely nothing to tickle my PB. Still trying ...

Re: [Rubycocoa-talk] Tool/headless GUI with RubyCocoa

[Rupert B. <rup...@fr...>]

Le 5 d=C3=A9c. 05 =C3=A0 22:12, Jonathan Paisley a =C3=A9crit :

> Are you talking about a class you have created natively in =20
> Objective C that you want to extend in Ruby-land?

I don't want to extend it, I just want to instantiate it.

> If so, you need to do
>
>    OSX.ns_import :YourClassName
>
> After that, a new class under the OSX module will be created with =20
> the appropriate name.

Thanks, this also works for simple instantiation.
I can now go ahead 'til the next hurdle : thanks for you repetitive =20
help, Jonathan.

Cheers,
Rup=

Re: [Rubycocoa-talk] Tool/headless GUI with RubyCocoa

[Jonathan P. <jp...@dc...>]

On 5 Dec 2005, at 21:00, Rupert BARROW wrote:

> Of course, this is what we do for OSX (Cocoa) objects, but I have  
> never created an instance in Ruby from a custom Objective-class.  
> How would you go about doing this, especially in a headless  
> application, for a Model or Controller Objective-C class ?

Are you talking about a class you have created natively in Objective  
C that you want to extend in Ruby-land?

If so, you need to do

    OSX.ns_import :YourClassName

After that, a new class under the OSX module will be created with the  
appropriate name. You can then subclass it as you do with other built- 
in classes:

  class MySubclass < OSX::YourClassName
    ...
  end

Please let me know if I've misunderstood.

Cheers,
Jonathan

Re: [Rubycocoa-talk] Tool/headless GUI with RubyCocoa

[Rupert B. <rup...@fr...>]

Thanks Jonathan,
Of course, this is what we do for OSX (Cocoa) objects, but I have =20
never created an instance in Ruby from a custom Objective-class. How =20
would you go about doing this, especially in a headless application, =20
for a Model or Controller Objective-C class ?

Rup


Le 5 d=C3=A9c. 05 =C3=A0 18:47, Jonathan Paisley a =C3=A9crit :

>
> On 4 Dec 2005, at 18:50, Rupert BARROW wrote:
>
>> I would like to develop a headless tool (app without a GUI) with =20
>> RubyCocoa : how can I create (in Ruby) Objective-C objects ?  It's =20=

>> difficult to #import header.h files  :-)  , or to connect with =20
>> ib_outlets !
>
> You can probably do it already. You don't need to import any header =20=

> files - most of the classes you need will have been imported into =20
> ruby (see the appkit.rb, foundation.rb files etc in RubyCocoa).
>
> Does that make sense?
>
>
> require 'osx/cocoa'
> include OSX
>
> object =3D NSWhatever.alloc.initWithWhatever:...
>
>
>

Re: [Rubycocoa-talk] Tool/headless GUI with RubyCocoa[MESSAGE NOT SCANNED]

[Jonathan P. <jp...@dc...>]

On 4 Dec 2005, at 18:50, Rupert BARROW wrote:

> I would like to develop a headless tool (app without a GUI) with  
> RubyCocoa : how can I create (in Ruby) Objective-C objects ?  It's  
> difficult to #import header.h files  :-)  , or to connect with  
> ib_outlets !

You can probably do it already. You don't need to import any header  
files - most of the classes you need will have been imported into  
ruby (see the appkit.rb, foundation.rb files etc in RubyCocoa).

Does that make sense?


require 'osx/cocoa'
include OSX

object = NSWhatever.alloc.initWithWhatever:...

Re: [Rubycocoa-talk] valueForKey:@"changed" == @"YES" ??? (ObjC vs. RubyCocoa strings ?)

[Jonathan P. <jp...@dc...>]

On 5 Dec 2005, at 16:36, kimura wataru wrote:

> In Objective-C, "==" operator for Cocoa objects tests by address. We
> can test values of strings with "isEqualToString:" method.
>
> I guess the objc compiler makes same address to literal string @"YES".
>
> Rupert BARROW wrote:
>> What is going on ?  I have made every effort for the two ObjC and
>> RubyCocoa strings be equal : same value ("YES"), same class
>> (NSCFString); what else ?  Why does the test
>> 		if ([record valueForKey:@"changed"] == @"YES")
>> never verify ?

I think I mentioned something like this being broken a short while  
ago (even though it comes from Apple's sample code - grr!).

As far as I know, it works sometimes because some strings are  
'interned', and so the literal @"YES" NSString/CFString object is  
actually the same one being returned from the valueForKey: method.  
Relying on this behaviour is not good.

Re: [Rubycocoa-talk] valueForKey:@"changed" == @"YES" ??? (ObjC vs. RubyCocoa strings ?)

[kimura wataru<ki...@us...>]

Hi,

In Objective-C, "==" operator for Cocoa objects tests by address. We
can test values of strings with "isEqualToString:" method.

I guess the objc compiler makes same address to literal string @"YES".


Rupert BARROW wrote:
>Hi,
>
>I've been spending most of the night trying to understand the  
>following problem :
>

>When I run the program, the traces from the Objective-C code show :
>
>	record changed=YES, class=NSCFString, res=(null)
>	record changed=YES, class=NSCFString, res=(null)
>	record changed=YES, class=NSCFString, res=(null)
>	record changed=YES, class=NSCFString, res=(null)
>	record changed=YES, class=NSCFString, res=(null)
>	changedRecords=()
>
>What is going on ?  I have made every effort for the two ObjC and  
>RubyCocoa strings be equal : same value ("YES"), same class  
>(NSCFString); what else ?  Why does the test
>		if ([record valueForKey:@"changed"] == @"YES")
>never verify ?
>
>
>Thanks for any help, I'm going to bed    %:(
>Rup
>

[Rubycocoa-talk] Tool/headless GUI with RubyCocoa

[Rupert B. <rup...@fr...>]

Hi,

I would like to develop a headless tool (app without a GUI) with  
RubyCocoa : how can I create (in Ruby) Objective-C objects ?  It's  
difficult to #import header.h files  :-)  , or to connect with  
ib_outlets !

Thanks for any help,
Rupert

[Rubycocoa-talk] valueForKey:@"changed" == @"YES" ??? (ObjC vs. RubyCocoa strings ?)

[Rupert B. <rup...@fr...>]

Hi,

I've been spending most of the night trying to understand the  
following problem :

I have Objective-C checking for changed objects like this :

	while (record = [recordEnumerator nextObject]) {
		NSLog(@"record changed=%@, class=%@, res=%@",
					[record valueForKey:@"changed"],
					[[record valueForKey:@"changed"] class],
					[record valueForKey:@"changed"] == @"YES");
		if ([record valueForKey:@"changed"] == @"YES")
			[changedRecords addObject:record];
	}
	NSLog(@"changedRecords=%@", [changedRecords description]);

I have RubyCocoa code updating these objects like this :

	def changed
		@_changed
	end
	self.addRubyMethod_withType("changed".to_sym, "v@:@")
	def setChanged(v)
		@_changed = v
	end
	self.addRubyMethod_withType("setChanged:".to_sym, "v@:@")
	def setChangedYes
		setChanged(OSX::NSString.alloc.initWithString("YES"))
	end
	def setChangedNo
		setChanged(OSX::NSString.alloc.initWithString("NO"))
	end

When I run the program, the traces from the Objective-C code show :

	record changed=YES, class=NSCFString, res=(null)
	record changed=YES, class=NSCFString, res=(null)
	record changed=YES, class=NSCFString, res=(null)
	record changed=YES, class=NSCFString, res=(null)
	record changed=YES, class=NSCFString, res=(null)
	changedRecords=()

What is going on ?  I have made every effort for the two ObjC and  
RubyCocoa strings be equal : same value ("YES"), same class  
(NSCFString); what else ?  Why does the test
		if ([record valueForKey:@"changed"] == @"YES")
never verify ?


Thanks for any help, I'm going to bed    %:(
Rup

Re: [Rubycocoa-talk] Core Image

[Jonathan P. <jp...@dc...>]

On 2 Dec 2005, at 9:13, Dave Baldwin wrote:

> I have just tried to use Core Image from rubycocoa and started  
> getting errors like:
> RBException_NameError - uninitialized constant CIFilter  
> (OSX::OCException)
>
> looking through the rubycocoa sources it looks like Core Image  
> isn't included.  Is this correct or am I missing something?

I think you are correct - CoreImage isn't wrapped yet.

You could possibly have a go yourself by looking at how the WebKit  
interface is handled - purely from ruby. See osx/webkit.rb and osx/ 
objc/webkit.rb. To summarise (you need to look at osx/objc/webkit.rb  
to see the details):

  (1) load the CoreImage framework
  (2) ns_import the classes needed
  (3) define integer constants/enums
  (4) create NSString (or other object) constants by using  
objc_symbol_to_obj

Hope that helps,
Jonathan

[Rubycocoa-talk] Core Image

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

I have just tried to use Core Image from rubycocoa and started  
getting errors like:
RBException_NameError - uninitialized constant CIFilter  
(OSX::OCException)

looking through the rubycocoa sources it looks like Core Image isn't  
included.  Is this correct or am I missing something?

Thanks,

Dave.

Re: [Rubycocoa-talk] Re: Newbie: stack level too deep??[MESSAGE NOT SCANNED]

[Jonathan P. <jp...@dc...>]

On 28 Nov 2005, at 20:24, Frank Illenberger wrote:

> thanks for the information. The Sydney project sounds interesting.  
> How would I proceed to get to a Cocoa app that may call methods  
> ruby classes (via RubyCocoa) from different threads?
> Is it straight forward to start with Syndey?

I'm afraid I don't know a lot about Sydney other than what I've read  
on that page. I have had a quick look at it, but I think it's not  
really ready for use with RubyCocoa, unless you're an experienced  
developer with a lot of time to figure out what's going on.

Maybe somebody on the Sydney mailing list would be able to give a  
better answer.

Jonathan

Re: [Rubycocoa-talk] Re: Newbie: stack level too deep??

[Frank I. <ill...@ma...>]

Jonathan,

thanks for the information. The Sydney project sounds interesting.  
How would I proceed to get to a Cocoa app that may call methods ruby  
classes (via RubyCocoa) from different threads?
Is it straight forward to start with Syndey?

Cheers

  Frank


On 28.11.2005, at 10:53, Jonathan Paisley wrote:

> On 28 Nov 2005, at 7:21, Frank Illenberger wrote:
>
>> I could dig a little deeper and found the following stack trace is  
>> repsonsible for a "Stack level too deep" exception:
>
> Ah, the problem is that the ruby code is being executed on a  
> secondary NSThread. Unfortunately, Ruby is not thread-safe with  
> respect to OS-level threads (which an NSThread is).
>
> Ruby checks from time to time that the stack has not overflowed,  
> and since the non-main thread will have a different area of memory  
> for its stack than the main thread, Ruby thinks a stack overflow  
> has occurred.
>
> I'm afraid that there's probably no good way to work around this at  
> the moment. It may be possible to get RubyCocoa running with Sydney  
> [1], which apparently gives OS-level thread support. Do let us know  
> if you give that a shot.
>
> Sorry I don't have better news for you
> Cheers
> Jonathan
>
>
> [1] http://blog.fallingsnow.net/articles?page=2
>
>
>
> -------------------------------------------------------
> This SF.net email is sponsored by: Splunk Inc. Do you grep through  
> log files
> for problems?  Stop!  Download the new AJAX search engine that makes
> searching your log files as easy as surfing the  web.  DOWNLOAD  
> SPLUNK!
> http://ads.osdn.com/?ad_id=7637&alloc_id=16865&op=click
> _______________________________________________
> Rubycocoa-talk mailing list
> Rub...@li...
> https://lists.sourceforge.net/lists/listinfo/rubycocoa-talk