headers
Nicola Pero | 1 Jan 2002 22:19
Picon

Re: NSScrollView

Ok - I've applied to CVS your changes to NSScrollView's -tile to set up
rulers.  

Incidentally, I've rewritten the rest of NSScrollView's -tile as well so
that - hopefully - it's easier to understand.

Your contribution was much appreciated, let me know if it works, I hope so
:-)
Permalink | Reply | 0 comments |
headers
Nicola Pero | 2 Jan 2002 08:48
Picon

Re: Small NSTextView patch

> Hi,
> 
> I've modified NSTextView: -mouseDown so that if our text storage contains 
> attributed strings with the NSLinkAttributeName and we 'click' on such 
> strings, NSTextView: -clickedOnLink: atIndex: is called so NSTextView's 
> delegate textView: clickedOnLink: atIndex: method will be called too.
> 
> I included a patch with this mail.

Hi Ludovic, 

thanks - interesting patch - had a look into it and ... well the doc is
unclear but I felt like changing it - what about the following code for
managing links ?

 /* Manage the case that the click is on a link (a link is some chars
     with the NSLinkAttributeName set to something which is not-null).  */
  {
    id link;

    /* What exactly is this link object, it's up to the programmer who
       is using the NSTextView and who originally created the link
       object and saved it under the NSLinkAttributeName in the text.
       If I were to use it, I would normally use a NSURL as the link
       object.  */
    link = [_textStorage attribute: NSLinkAttributeName
			 atIndex: startIndex
			 effectiveRange: NULL];

    if (link != null)
(Continue reading)

Permalink | Reply | 0 comments |
headers
Adam Fedor | 2 Jan 2002 19:08

class method forward invocations

In this example, the class forwardInvocation method doesn't get called, 
instead I get:

Jan 02 11:04:33 longstring[8059] Got object forwardInvocation
Jan 02 11:04:33 longstring[8059] GSFFCallInvocation.m:754  Assertion 
failed in GSInvocationCallback.  GSFFCallInvocation: Class 'MyClass' 
does not respond to forwardInvocation: for 'intValue'
Uncaught exception NSInternalInconsistencyException, reason: 
GSFFCallInvocation.m:754  Assertion failed in GSInvocationCallback. 
GSFFCallInvocation: Class 'MyClass' does not respond to 
forwardInvocation: for 'intValue'
Aborted (core dumped)

-------------------------------------------------------------------------------
I'm wondering if this fix will create some new problem:

diff -c -r1.15 GSFFCallInvocation.m
*** GSFFCallInvocation.m        17 Dec 2001 14:31:42 -0000      1.15
--- GSFFCallInvocation.m        2 Jan 2002 18:07:28 -0000
***************
*** 215,221 ****
             return class_get_instance_method (object_get_class
                                                 (receiver), sel);
           }
!       else if (object_is_meta_class (receiver))
           {
             return class_get_class_method (object_get_meta_class
                                              (receiver), sel);
--- 215,221 ----
             return class_get_instance_method (object_get_class
(Continue reading)

Permalink | Reply | 1 comment |
headers
Adam Fedor | 2 Jan 2002 21:06

Re: ScriptKit patch

Anthony W. Juckel wrote:

> I've recently been looking into the ScriptKit for an application that 
> I've started.  Unfortunately, I was unable to get run-tests.sh to run 
> properly without the following patch.  It simply re-orders some of the 
> data arguments to let* within generateRealScript in 
> ScriptKit/Library/GuileInterpreter.m.  I'm not sure if this was 
> necessary because of some peculiarities in my guile set-up, or if others 
> would have the same problem with a recent version of guile (I'm using a 
> 1.4 debian package), but I thought I would post the patch to the list to 
> see what others thought.

I added this patch. I had the same problem with my setup (guile 1.4) 
which this fixed. Thanks.

--

-- 
Adam Fedor, Digital Optics            | I'm glad I hate spinach, because
http://www.doc.com                    | if I didn't, I'd eat it, and you
http://www.gnustep.org                | know how I hate the stuff.
Permalink | Reply | 0 comments |
headers
Willem Rein Oudshoorn | 3 Jan 2002 22:26
Picon
Picon
Favicon

Re: class method forward invocations

I think this patch is correct.  This was my first contact
with the ObjC runtime and I remember that I found the code
counter intuitive.  But I probably got confused by
the the conversion from 
        receiver -> metaclass.
and got distracted by other problems.

Wim Oudshoorn.

Adam Fedor <fedor <at> doc.com> writes:
> I'm wondering if this fix will create some new problem:
> 
> diff -c -r1.15 GSFFCallInvocation.m
> *** GSFFCallInvocation.m        17 Dec 2001 14:31:42 -0000      1.15
> --- GSFFCallInvocation.m        2 Jan 2002 18:07:28 -0000
> ***************
> *** 215,221 ****
>              return class_get_instance_method (object_get_class
>                                                  (receiver), sel);
>            }
> !       else if (object_is_meta_class (receiver))
>            {
>              return class_get_class_method (object_get_meta_class
>                                               (receiver), sel);
> --- 215,221 ----
>              return class_get_instance_method (object_get_class
>                                                  (receiver), sel);
>            }
> !       else if (object_is_class (receiver))
>            {
(Continue reading)

Permalink | Reply | 0 comments |
headers
Dominique Dumont | 4 Jan 2002 10:39
Picon

patch for gnustep-make on HPUX IA64


Hello

On HPUX IA64, the suffix of the shared library is .so instead of .sl.

Here's a small patch so that gnustep-make can produce the correct
shared lib.

Hope this helps (and besh wishes for 2002).

--

-- 
Dominique_Dumont <at> grenoble.hp.com
Attachment (gnustep-make-1.0.1.diff): text/x-patch, 451 bytes
Permalink | Reply | 1 comment |
headers
Nicola Pero | 4 Jan 2002 09:12
Picon

Re: patch for gnustep-make on HPUX IA64

Thanks - I applied this patch to the CVS repository

On 4 Jan 2002, Dominique Dumont wrote:

> 
> Hello
> 
> On HPUX IA64, the suffix of the shared library is .so instead of .sl.
> 
> Here's a small patch so that gnustep-make can produce the correct
> shared lib.
> 
> Hope this helps (and besh wishes for 2002).
> 
>
Permalink | Reply | 0 comments |
headers
Stephen Brandon | 4 Jan 2002 19:08
Picon

objc classes don't load at runtime - gui/win32

Hi,

I reported this problem some time ago but made no progress
at that time. Today though I have discovered what might be a
clue - so I'd appreciate anyone who has knowledge of the
objc runtime having a look at this and seeing if it's
relevant to the problem.

Setup:
latest CVS gnustep on Win2000/Cygwin tools/mingw32.

Background:
The basic problem is that gnustep-gui classes are not found
at runtime. gnustep-base classes are fine, and other
frameworks that I use (MusicKit, SndKit etc) also load ok.

Diagnostic tools:
I compile libobjc-1.1.0 with the debug option on, so all
loading and caching of imps, modules and classes is
reported.

New clue:
For the first time, I see some diagnostic output that seems
to relate to a gnustep-gui class, thus hopefully pointing to
the fact that it is being at least partially loaded. In this
case, I get some of the following in the output:

...
Record selector NSEraseRect:[(null)] as: 1459
Record selector NSHighlightRect:[(null)] as: 1460
(Continue reading)

Permalink | Reply | 1 comment |
headers
Adam Fedor | 5 Jan 2002 06:35

Re: objc classes don't load at runtime - gui/win32

Stephen Brandon wrote:

> 
> ...
> received module: GSCompatibility.m
> gathering selectors from module: GSCompatibility.m
> received module: PerformSound.m
> gathering selectors from module: PerformSound.m
> received module: SndConditionLock.m
> Record selector NSEraseRect:[(null)] as: 1459
> Record selector NSHighlightRect:[(null)] as: 1460
> Record selector NSRectClip:[(null)] as: 1461
> Record selector NSRectClipList::[(null)] as: 1462
> Record selector NSRectFill:[(null)] as: 1463
> (etc)
> 
> Needless to say, my SndConditionLock class makes absolutely
> no reference to any of the graphics functions/selectors.
> There are not even any headers imported that would refer to
> appkit things. Looking at the order of output from
> gnustep-objc-1.1.0/init.c, these selectors should definitely
> be related to the preceding module name (as far as I can
> make out).
> 

Well these lines would show up if you used these selectors in that 
module. If you haven't, and there is little reason to do so, that's very 
odd.

I have to say I'm baffled by the whole thing. I've tried some things
(Continue reading)

Permalink | Reply | 0 comments |
headers
Adam Fedor | 7 Jan 2002 00:59

Re: Euro support in GNUstep

Martin Brecher wrote:

> Hi Adam, -
> 
> since you're planning to do a release of GNUstep core very soon I'd like to 
> propose a switch of the Language files to be Euro-compliant as since Jan 1st 
> the Euro is the only official (as defined by law) currency in the involved 
> countries.
> 
> The following changes should do the trick (since we don't have a Euro symbol 
> yet):
> 
> NSCurrencySymbol = "EUR";
> NSInternationalCurrencyString = EUR;
> NSPositiveCurrencyFormatString = "EUR9,999.00";
> NSNegativeCurrencyFormatString = "EUR-9,999.00";
> 
> 
> Taking a look at the /System/Libraries/Resources/Languages dir, the following 
> files need to be changed:
> Dutch French German Italian
> 

If I make this change is it going to mess up any programs that expect 
the local currency (i.e. conversion programs)?  I suppose it is not used 
that much but it seems like it might be nice to either have a new set of 
strings for the Euro or at least have a set of strings for the old 
currency for compatibility.

--

--
(Continue reading)

Permalink | Reply | 5 comments |

Gmane