Both GNUstep and Cocotron have already achieved this “OS independence and
portability” better than you think, on a wider variety of platforms - Linux as
well as Windows and FreeBSD. So if you open up the Objective-C compatibility
layer and leave the Foundation to them you may have even better OS
ner wrote:
>
>
>> On Dec 3, 2015, at 12:01 PM, Maxthon Chan wrote:
>>
>> Dear Swift developers:
>>
>> Maybe you have never heard of it, but there have been several ongoing
>> efforts, like GNUstep and Cocotron, at maintaining an open source Fo
Dear Swift developers:
Maybe you have never heard of it, but there have been several ongoing efforts,
like GNUstep and Cocotron, at maintaining an open source Foundation
reimplementation for alternative operating systems like Linux. It seemed to me
that the current release of Swift did not put
I am thinking rewriting buildtool using CoreFoundation API to break dependency
on gnustep-make and gnustep-base.
I am thinking creating forks of base, libobjc2 and corebase projects and
rearrange some code, to resolve some compatibility issues and modernise code to
take as many advantage as pos
clang is surprisingly fast if built with LLVM link-time optimisation - even
faster than GCC, from my experience. So I am setting up a new server with this
LTO’d clang. But then I get a SIGSEGV when testing libobjc2 built with this
augment clang. Can anyone test and tell me the reason why?
Build
clang is surprisingly fast if built with LLVM link-time optimisation - even
faster than GCC, from my experience. So I am setting up a new server with this
LTO’d clang. But then I get a SIGSEGV when testing libobjc2 built with this
augment clang. Can anyone test and tell me the reason why?
Build
eversing Apple's Cocoa to make OS X apps run under Linux...
Sent from my iPhone
> On 2013年6月19日, at 21:28, Luboš Doležel wrote:
>
>> On Wed, 19 Jun 2013 21:18:18 +0800, Maxthon Chan wrote:
>> Given that:
>>
>> 1) Apple did not patent Cocoa and/or their impleme
n took
> effect on the mail server.
>
> Greg
>
>
>> On Wed, Jun 19, 2013 at 10:54 AM, Maxthon Chan wrote:
>> Forgot to mention, OS X 10.9 is based on iOS 6.
>>
>> Sent from my iPhone
>>
>>> On 2013年6月19日, at 22:33, Gregory Casamento wro
les, ignoring the values that many of the
>> people on this list share with the organisation that "owns", in its way, the
>> GNUstep project. It is not a surprise that these people are, as a result,
>> asking you to avoid repeating your mistakes or even calling for you t
oring the values that many of the
>> people on this list share with the organisation that "owns", in its way, the
>> GNUstep project. It is not a surprise that these people are, as a result,
>> asking you to avoid repeating your mistakes or even calling for you to b
Given that:
1) Apple did not patent Cocoa and/or their implementation of CoreFoundation, and
2) GNUstep contains no Apple code
We will be good, as Apple have no reason to drag us into trouble. We are not
infringing any IP at all if that two conditions is met.
Just look at Mono which is a binary
ote:
>
>> On 19 Jun 2013, at 13:26, Maxthon Chan wrote:
>>
>> This analyse is based on OS X 10.8 SDK from Xcode 4.6, and OS X 10.9 DP1
>> binaries.
>
> You have been asked repeatedly before not to post information gleaned in this
> manner to this list.
>
$for fmwk in Foundation CoreFoundation; do class-dump
/System/Library/Framework/$fmwk.framework -H -o ~/Developer/$fmwk; done
And I got a set of class dump of Foundation and CoreFoundation of OS X. And
examining the files I think I found something:
* A significant portion of what OS X called Fo
https://www.dropbox.com/s/gb370rbxti65gim/GNUstep.tar.gz
Here it is. Tested on Xcode 4.6 and Xcode 5. Drop them to where Xcode carried
their OS X SDK and toolchain and you are good to go.
Compiling does not work now, as the target OS is Linux which Xcode does not
support, and it does not show .
If someone really does not need ObjC stuff here, they will go to CFLite instead.
On Jun 17, 2013, at 20:57, Stefan Bidi wrote:
> I was specifically trying not to link against the ObjC runtime and
> gnustep-base. The reason for that is because I would like to support
> building corebase withou
One thing failed: I cannot cross-compile libc++. This can be problematic on the
final toolchain as Xcode defaults to it.
On Jun 17, 2013, at 11:32, Maxthon Chan wrote:
> Partial success: now I can get editing work on Xcode with all code completion
> things. So the tasks coming now is t
Partial success: now I can get editing work on Xcode with all code completion
things. So the tasks coming now is to make sure it edits with Objective-C.
Cross-building will be a total different story as clearly OS X does not run ELF
and Xcode have guards against it. (However it does make it easi
th the Linux and OS X
> portions of the toolchain.
>
> On Sun, Jun 16, 2013 at 2:20 PM, Maxthon Chan wrote:
> I will build a toolchain based on my Ubuntu installation for Xcode 5:
>
> * Whatever version of basic system libraries copied from Ubuntu 13.04 as of
> now
> *
You mean sort of like using an environment variable GS_LOAD_DEBUG=BFD (or GS_LOAD_DEBUG=LLDB for LLDB, and GS_LOAD_DEBUG=NO to explicitly prohibit it) to control it? By default this BFD (or LLDB) bundle is not loaded (changeable at base ./configure) and can be loaded on need?2013年6月16日 下午11:31,Rich
targeting 64-bit Linux with gold,
* Templates based on OS X templates
* SenTestingKit
* and maybe network-backed LLDB remote debugging
Sent from my iPhone
> On 2013年6月16日, at 19:53, David Chisnall wrote:
>
>> On 16 Jun 2013, at 12:49, Maxthon Chan wrote:
>>
>> Also on t
> On 2013年6月16日, at 19:39, David Chisnall wrote:
>
>> On 16 Jun 2013, at 12:28, Maxthon Chan wrote:
>>
>> Then it is going to be something awkward: one cannot legally link 2-clause
>> BSDL binary against GPL one, and CGIKit is 2-clause BSDL
>
> Please talk
against in LGPL without promoting to GPL, which in turn can be linked
against in BSDL like CGIKit)
On Jun 16, 2013, at 19:19, David Chisnall wrote:
> On 16 Jun 2013, at 12:14, Maxthon Chan wrote:
>
>> I know my CGIKit, a simple CGI-based Objective-C web service library, may
>> r
I know my CGIKit, a simple CGI-based Objective-C web service library, may rely
on this - when an exception is caught it will create a HTTP 500 page with good
error backtrace, if the programmer so desires.
On Jun 16, 2013, at 19:04, David Chisnall wrote:
> On 16 Jun 2013, at 11:52, Gregory Cas
t is not really the place to ask for legal advice such as this as
> none of us are lawyers.
>
> Gregory Casamento
> greg.casame...@gmail.com
> Open Logic Corporation, President/CEO
> GNUstep Lead Developer
>
>
> On Jun 16, 2013, at 6:40 AM, Maxthon Chan wrote:
>
&
Is there any way to prevent GPL from spreading into GNUstep Base (and
everything that is built on top of it) when using BFD support? Is it possible
to:
1) Use some alternative technologies, like LLVM and its subproject LLDB, or
2) Split the BFD support into a loadable bundle which is loaded inst
I think the latest version of CFLite is in APSL v2 which is FSF approved.
On Jun 16, 2013, at 18:30, Gregory Casamento wrote:
> I stand corrected. My apologies. As I recall they are incompatible because
> there is a clause in the apsl which allows apple to revoke your right to use
> the code.
NDA, anything
>>> someone else created without improper access to the trade secret
>>> information or anything that you can prove was commonly known to the
>>> general public.
>>>
>>> That's why I was warning you. Typically it's dangerous to a
the information covered in the NDA, anything someone
>> else created without improper access to the trade secret information or
>> anything that you can prove was commonly known to the general public.
>>
>> That's why I was warning you. Typically it's dangerous t
> I don't want a posting here to become any grounds for Apple to cause any
> issues for GNUstep.
>
> Gregory
>
>
>> On Sat, Jun 15, 2013 at 5:57 AM, Maxthon Chan wrote:
>> I am watching a video session from WWDC 2013 and Apple revealed their way of
>> i
I am watching a video session from WWDC 2013 and Apple revealed their way of
implementing tagged pointers: use the sole LSB, 1 for tagged pointer and 0 for
normal ones. This is applied across all platforms, i386, amd64, armv7 and
armv7s (compatible to armv7a in Coretx-A15 and partly compatible w
I will give you my settings once I got home (use libsupc++.a under 64-bit
Ubuntu 13.04). However my settings is calling for LLVMgold.so and
clang+binutils LTO support.
Sent from my iPhone
> On 2013年6月14日, at 17:48, David Chisnall wrote:
>
>
>> On 14 Jun 2013, at 02:54, Frank Rehwinkel wrote
unctions is included by a C++ module.
>
> So no, I don't typically have things like what you wrote down there :-)
>
> Regards,
>
> Ivan Vučica
> via phone
>
>> On 14. 6. 2013., at 08:58, Maxthon Chan wrote:
>>
>> I mean exposing a function or two fr
time (while
> giving us the bonus of modular, reusable, easily maintainable code). Why
> would you want/need to unhide details?
>
>
>> On Fri, Jun 14, 2013 at 2:59 AM, Maxthon Chan wrote:
>> Copied from Xcode and Mail saved all formatting from code highlighting as
>>
t;> On Wed, Jun 12, 2013 at 12:25 PM, Maxthon Chan wrote:
>> Do you guys have a boilerplate that you are very used to build code on top
>> of? I have one like this (sans include guard which depend on header file
>> name):
>>
>> #include
>> #include
>
> So I just don't know how to get this library built. Not as simple as getting
> a port compiled I guess.
>
>
>
> On Wed, Jun 12, 2013 at 7:07 PM, Frank Rehwinkel
> wrote:
> libsupc++.a does have symbols in it.
> nm -g reports no symbols on libsupc++.so.1.
&
Oh it just hit me - maybe you want libsupc++.a instead of .so, which (may) ask
for libc++ to build.
I used to simply force linking libobjc2 against libc++ but that tend to make
app load a little slower and consume some more memory (and then defeat this
performance penalty by enforcing link-time
Yup. This is the library responsible of that.
It is C++ ABI library, essentially a runtime library like libobjc2. (libsupc++
<-> libobjc2, libc++/libstdc++ <-> libgnustep-base, you can think of it that
way.)
在 2013-6-13，上午6:39，David Chisnall 写道：
>
> On 12 Jun 2013, at 23:35, Frank Rehwinkel
_cxxabiv1::__class_type_inf
> o const*, void**) const'
> libobjc.so.4.6: undefined reference to `std::type_info::__is_pointer_p()
> const'
> clang-3.4: error: linker command failed with exit code 1 (use -v to see
> invocation)
>
>
> On Wed, Jun 12, 2013 at 4:43 PM
Apple have already fused (most of) Foundation and CF into
CoreFoundation.framework. We can totally combine them too - provide a
libgnustep-corebase that contains (almost) all Foundation and all
CoreFoundation, and a libgnustep-base that contains the rest of Foundation,
linking against CoreFound
time I used
> cmake. So that explains why this problem didn't surface the first time.
>
> But I'm left with the first problem reported. That -base won't link with the
> libobjc.so.4.6. Even when -lsup++ is added to the link options.
>
>
>> On Wed, J
;t surface the first time.
>
> But I'm left with the first problem reported. That -base won't link with the
> libobjc.so.4.6. Even when -lsup++ is added to the link options.
>
>
>> On Wed, Jun 12, 2013 at 2:31 PM, Maxthon Chan wrote:
>> Hmm? Maybe the -s
>
> libobjc.so.4.6: undefined reference to `std::type_info::__is_pointer_p()
> const'
> clang-3.4: error: linker command failed with exit code 1 (use -v to see
> invocation)
> ninja: build stopped: sub
pc++. libsupc++ is in both the FreeBSD 9.1 host and the jail.
>
> Is using libsupc++ an option for building libobjc2 and gnustep-base?
>
>
> On Wed, Jun 12, 2013 at 1:57 PM, Maxthon Chan wrote:
> Download and build it separately using the same SVN head version.
>
> with libsupc++. libsupc++ is in both the FreeBSD 9.1 host and the jail.
>
> Is using libsupc++ an option for building libobjc2 and gnustep-base?
>
>
> On Wed, Jun 12, 2013 at 1:57 PM, Maxthon Chan wrote:
> Download and build it separately using the same SVN head version.
Download and build it separately using the same SVN head version.
If you built that, you should also build libc++ (LLVM's C++ runtime) and
(maybe) link it against libc++rt.
Also, when building gnustep-base, specify -I/usr/local/include and
-L/usr/local/lib appropriately.
在 2013-6-13，上午1:54，Fra
Do you guys have a boilerplate that you are very used to build code on top of?
I have one like this (sans include guard which depend on header file name):
#include
#include
#include
#include
#include
// Feature testers
#ifndef __has_feature
#define __has_feature(x) 0
#endif
#ifndef __has_
I am sort of looking at a microkernel design - inside the kernel there are only
2 drivers: tmpfs (implemented on top of NSDictionary and NSData) and tar (on
top of NSArray and NSData) to make initramfs (which is a simple tar archive of
all pre-filesystem kexts and services) work, and everything
So I can safely do this? (Macro rewritten as inline function)
static inline __attribute__((always_inline)) BOOL objc_isSmallObject(id obj)
{
return (((intptr_t)obj) & 1) ? YES : NO;
}
在 2013-6-8，上午4:10，David Chisnall 写道：
> On 7 Jun 2013, at 16:29, Luboš Doležel wrote:
>
>> for toll-fr
in an
Objective-C header, as seen)
#include
// Prototype:
int main(int argc, char **argv)
{
return CGIApplicationMain(argc, argv, NULL,
"CGITestApplicationDelegate");
}
在 2013-6-7，下午4:47，David Chisnall 写道：
> On 7 Jun 2013, at 09:42, Maxthon Chan wrote:
>
>>
Well can I (just like NSApplication):
1) In supercalss, define the shared instance as id
2) In superclass, return the shared instance as id or instancetype
3) In superclass, DO NOT set up yet.
4) In superclass, set up in the method asking for the shared instance, which
always use [[self alloc] in
etype)sharedSingleton
{
@synchronized (__singleton)
{
if (!__singleton)
__singleton = [[self alloc] init];
}
return __singleton;
}
// …
@end
在 2013-6-7，下午4:26，David Chisnall 写道：
> On 7 Jun 2013, at 07:29, Maxthon Cha
Just asking, with ARC, is this a good choice on implementing singleton?
//Singleton.h
#import
@interface Singleton : NSObject
+ (instancetype)defaultSingleton;
// …
@end
extern Singleton *DefaultSingleton // Of course this is optional
// Singleton.m
Singleton *DefaultSingleton
@implementation Si
ot sure, but it might also be enough to draw
> maps.
>
> I'd also suggest looking into implementing CATiledLayer in QuartzCore :-)
>
> Regards,,
>
> Ivan Vučica
> via phone
>
> On 4. 6. 2013., at 08:27, Maxthon Chan wrote:
>
>> Here is a list of ta
Here is a small patch I made so that libobjc2 can be built on environments
where -std=c++11 presents in CXXFLAGS.
The last time when I built libobjc2 linking against libsupc++ with clang SVN
(itself is built with C++11 enabled, together with lldb and linked against
libc++) enabling C++11, I hit
. bridging). In this
> particular case, things are done the way they are done because it is the only
> sensible way without having tons special cases.
>
> Stef
>
>
> On Tue, Jun 4, 2013 at 1:19 PM, Maxthon Chan wrote:
> I never intend to do what Apple did in their exact
were made. The most important of
> all being how CF types interact with ObjC classes (i.e. bridging). In this
> particular case, things are done the way they are done because it is the only
> sensible way without having tons special cases.
>
> Stef
>
>
> On Tue, Jun 4,
time quite a few important design decisions were made. The most important of
> all being how CF types interact with ObjC classes (i.e. bridging). In this
> particular case, things are done the way they are done because it is the only
> sensible way without having tons special cases.
unLoop. It will have an isa
pointer of this class.
@interface GSCFRunLoop : NSProxy
// …
@end
在 2013-6-5，上午1:52，Ivan Vučica 写道：
> Maxton,
>
> On 4. 6. 2013., at 19:27, Maxthon Chan wrote:
>
>>> My reason on saying that you may not be clear what Apple is doing:
>&g
If you are worried about legal issues, just see how happily ReactOS project is
going. Their aim is a *complete Windows clone* that is *binary compatible* and
Microsoft have no issue with them. If Apple have an issue about this, they
should have already sued FSF and shut this project down.
在 201
下面是被转发的邮件：
> 发件人: Maxthon Chan
> 主题: 回复： __NSCF** classes and libobjc2 dependency on C++ runtime
> 日期: 2013年6月5日 GMT+0800上午1时20分19秒
> 收件人: David Chisnall
>
> My reason on saying that you may not be clear what Apple is doing:
>
> It seemed to me that you have totally
下面是被转发的邮件：
> 发件人: Maxthon Chan
> 主题: 回复： CoreLocation clone
> 日期: 2013年6月4日 GMT+0800下午2时10分09秒
> 收件人: Sebastian Reitenbach
>
> Given that, I would like to back port. Also, I would like to eliminate the
> need for a tile generating service too, as GNUstep have
and 3.
在 2013-6-4，下午2:06，Sebastian Reitenbach 写道：
>
> On Tuesday, June 4, 2013 02:45 CEST, Ivan Vučica wrote:
>
>>
>> On 4. 6. 2013., at 01:02, Maxthon Chan wrote:
>>
>>> From what I know, you can use Android location services without Google Maps
>
>>
>>> The MapKit for OS X in question seemed used only public Google Maps API. It
>>> is freely licensed as well, no key required and the author himself is not
>>> in trouble as well.
>>
>> Did you read Google Maps API terms of service and, more impo
code over to
a Linux box and remotely debug. It should not be that difficult to build
cross-compiling binutils and LLVM/clang targeting Linux under OS X, and consist
them into a SDK/toolchain for Xcode.
在 2013-6-4，上午8:58，Ivan Vučica 写道：
> On 4. 6. 2013., at 02:14, Maxthon Chan wr
下面是被转发的邮件：
> 发件人: Maxthon Chan
> 主题: 回复： CoreLocation clone
> 日期: 2013年6月4日 GMT+0800上午9时07分39秒
> 收件人: Ivan Vučica
>
> Agree on most of that, but one problem about tiles: can we do that ourselves,
> distributed, use Display PostScript (probably backed up by Cairo or
Can you compile it with -fobjc-arc and check again? Probably with ARC you can
have some ideas.
在 2013-6-4，上午7:33，Germán Arias 写道：
> For autocomplete I have a class GSAutocompleteWIndow. The
> method -complete: (in NSTextVIew) do something like:
>
> GSAutocompleteWindow *window = [GSAutocomplet
And by what I say as crowd-sourcing, I mean ask users that are willing to share
for information. What we are asking for here is several numbers: AP BSSID,
transmission power, and its geological location (latitude and longitude, as
well as height). The request is done by the device uploading a li
required and the author himself is not in
trouble as well.
在 2013-6-4，上午6:19，Ivan Vučica 写道：
> On 3. 6. 2013., at 23:56, Maxthon Chan wrote:
>
>> Well given Android is (at least partially) open-sourced, is it possible to
>> search their source code for their Wi-Fi geolocation API?
I did some research on how Apple implemented Objective-C exceptions and
CoreFoundation on iOS (iOS 6.1.3 on iPhone 4S) and I found something that we
can use.
If you want me to test anything, please reply to this Email and I will test it
for you. I have three devices to test: iOS 6.1.3 on iPhone
Don't panic, as Apple will switch back in no more than 2 years - when they
reached OS X 10.10 (OS X 10.9 is pretty much confirmed in this year's WWDC - I
am a paid iOS developer so I have access to these "Apple Prerelease"
information. Or maybe at that time there will be no more OS X, and the ne
Well given Android is (at least partially) open-sourced, is it possible to
search their source code for their Wi-Fi geolocation API? Implement that as a
loadable bundle to prevent licensing issue. (this CoreLocation clone can be
LGPL so that this bundle, probably under Apache license as derived
下面是被转发的邮件：
> 发件人: Maxthon Chan
> 主题: 回复： CoreLocation clone
> 日期: 2013年6月4日 GMT+0800上午4时06分54秒
> 收件人: David Wetzel
>
> If so, how about hack into it and clone CoreLocation? As of now, we need to
> hack out 1) using Google API (lots of NSURLConnection things, maybe
下面是被转发的邮件：
> 发件人: Maxthon Chan
> 主题: 回复： OS X/objc2: is clang required to build GNUstep-base?
> 日期: 2013年6月3日 GMT+0800上午3时29分01秒
> 收件人: Frank Rehwinkel
>
> 1) Do you have Xcode installed? I mean Xcode 4.6.2.
> 2) If you have Xcode installed, do you have command line tool
Sorry I messed up with the mail list.
下面是被转发的邮件：
> 发件人: Maxthon Chan
> 主题: 回复： GNUStep on MacOS
> 日期: 2013年5月28日 GMT+0800上午4时22分25秒
> 收件人: Frank Rehwinkel
>
> I cannot even build the Additions on OS X.
>
> As I used to introduce, I am working on a Objective-C HTTP ser
Well somehow I managed to compiled it with CXXFLAGS="-std=c++11 -stdlib=libc++"
and linked libobjc (with Objective-C++ support) against libc++ (LLVM's C++11
standard library), that is I made a huge hump of libraries, libobjc2,
libdispatch and LLVM's libc++ all linked together.
I always use libo
That doesn't matter if we keep internal consistency.
That is, the memory management of CoreFoundation objects are all done via
bridging to Objective-C code, no matter if it is toll-free or not (For
non-toll-free objects, the private Objective-C class corresponding to the
CoreFoundation object i
cuments.
在 2013-5-17，上午3:30，Luboš Doležel 写道：
> On 05/16/2013 08:26 PM, Maxthon Chan wrote:
>> Well Apple documented CFRetain and CFRelease as replacements of
>> retain/release methods on NSObject which is prohibited with ARC, and
>> from an incident I discovered that:
>&g
I have some portability issues when I mix ARC with some occasional manual
retains (especially when hand writing encoders and decoders).
The offending function is CFRetain and CFRelease, which is, as documented by
Apple, recommended as replacements of (forbidden by ARC) -[NSObject retain] and
-[
xcodeproj bundles? As well as
provide those Xcode command line tools, including xcrun, xcodebuild and like?
在 2013-4-14，下午6:42，David Chisnall 写道：
> How is that in any way even remotely relevant to a discussion of gmodel
> support in -gui?
>
> David
>
> On 13 Apr 2013, a
JSON-based protocol over HTTPS to the server, hence I wrote
AutoProtocol.
It is open sourced, but license is not yet decided. If you guys are interested,
I will LGPL it (after removing my project-specific code); or else, i will BSDL
it.
在 2013-4-13，下午3:57，Riccardo Mottola 写道：
> Maxthon Ch
Um, I think my project, a partly open-source social aggregation software,
WebFusion, will have a Linux version that is a direct port (as in almost 100%
code reused) from the Mac version. Hence it will not be the last known
application.
You can check out my GitHub for this project:
https://gith
Case seemed to be dismissed. This issue seemed is more about reflective KVC.
在 2013-4-11，下午8:51，Maxthon Chan 写道：
> #import
>
> int main(void)
> {
> @autoreleasepool
> {
> NSURL *URL = [NSURL URLWithString:@"https://www.google.com/";
against Apple's
Foundation)
By the way, my extension will export a load of extra API which is technically a
wrapper around libcurl functions.
在 2013-4-11，下午5:14，Richard Frith-Macdonald 写道：
>
> On 11 Apr 2013, at 09:59, Maxthon Chan wrote:
>
>> About this issue, I am rewritin
iling list, as I wont be looking
>> into this any further.
>>
>> On 10.04.2013 09:49, Maxthon Chan wrote:
>>> Well you can just try grab data from https://www.google.com/ using
>>> NSURLConnection. It should return an HTML page with status code 200 (the
>>
which prints some output that is
> obviously right or wrong, or which asset()s the correct output.
>
> David
>
> On 9 Apr 2013, at 17:19, Maxthon Chan wrote:
>
>> I will just copy/paste some code from my actual project. Compile without
>> -DGNUSTEP:
>>
&g
Sorry, messed up again.
下面是被转发的邮件：
> 发件人: Maxthon Chan
> 主题: 回复： Bug: libobjc2 property name introspection.
> 日期: 2013年4月10日 GMT+0800上午12时33分52秒
> 收件人: David Chisnall
>
> Here is one, exposes this bug well:
>
> #import
> #import
>
> @interface TFoo
Ignore the missing methods. They are handled by the base class using method
"forwarding". (I abused it to send JSON-based data to a server, calling methods
there.)
在 2013-4-10，上午12:13，David Chisnall 写道：
> Please send me a (simple) test case.
>
> David
>
> On 9 Ap
When I enumerate a class for all its properties and querying its name, the name
of last property is always cobbled. I have to use some "quick & dirty" way (i.e
add a dummy property at the end) to prevent this from affecting. OS X does not
have this issue, though.
Can anyone figure out what is w
Hi Everyone:
I am writing a project for OS X, iOS and Linux, and a cross-platform lore
library is called for. This following code works fine under both OS X and iOS,
but are very problematic under Linux using GNUstep:
- (NSData *)dataWithData:(NSData *)data fromMethod:(NSString *)method
error:
Based on your research, I am seriously doubting what is actually inside the
CoreFoundation binary: is it possible that actually all Objective-C code that
involves toll-free bridging or like, or maybe even more, that is documented as
located in Foundation, is actually located right in CoreFoundat
However this is the Apple approach. Can we just think out of the box and
maintain just API interoperability?
As I said, Apple used Core Foundation to serve both Cocoa and Carbon
simultaneously, which is asking for a C API (Cocoa was in Objective-C and
Carbon was C++, and there wasn't a good imp
800, Maxthon Chan wrote:
>> I think I scored one on suggesting implementing CF in ObjC here. :)
>
> Well, I have my doubts about this claim. Especially because of
>
>>> "If you look at the Darwin sources, you'll see that the iPhone's
>>> CoreFoundation cla
I think I scored one on suggesting implementing CF in ObjC here. :)
There are very few iOS code available out there, so it is very difficult to
hack into that.
By the way, I do have an paid Apple iOS Developer account so maybe I can verify
something for you? Maybe check iPhone's CF framework's
You should remember as of now, OS X code is still "fat binaries" that runs on
both 32-bit and 64-bit environments and Carbon libraries are still there, so OS
X still needs that C-based CF to support Carbon and Cocoa in the same time.
However on iOS there is no Carbon in the first place thus they
Can we make a GSPointerWrapper class or like to wrap non-objects into objects
so that they can be safely placed into the existing NSArray/NSDictionary
objects in order to implement CF objects in Objective-C? And as I said,
CoreFoundation is designed to support both Foundation and Carbon on OS X,
An alternative idea: stick to original NeXTSTEP method?
As what I know, NeXTSTEP use Display PostScript to manifest GUI, and is
"essentially PDF". So is it possible to use that old method to output PDF, that
is, get the portion out from the DPS frame?
在 2013-2-27，下午7:35，Robert Slover 写道：
> If
About reimplementing UIKit, you can actually map UIKit to AppKit. Those code
are largely similar.
p.s. Is there any tool that converts Xcode project into Makefile?
在 2013-2-26，下午4:21，Luboš Doležel 写道：
> On 25.2.2013 11:17, Ivan Vučica wrote:
>> I think it might also be a better year for me to
Dear Johan:
Hello.
1) Objective-C 2.0 is creation of Apple Inc. and is incorporated into OS X 10.5
(partially) / 10.7 (completely) and iOS 4.3 (partially) / 5.0 (completely). For
GNUstep, the latest svn build of GNUstep can support most of the Objective-C
2.0 features if using clang as the com
Well if you want to use ARC, you need libobjc2 instead of original libobjc, and
compile using clang with -fobjc-arc flag, assuming you are using a 64-bit
machine. You can read Apple's documentation on ARC.
在 2013-2-5，下午5:42，Riccardo Mottola 写道：
> Maxthon Chan wrote:
>> Just an
Just an idea, will ARC work?
在 2013-2-5，下午5:20，Riccardo Mottola 写道：
> Hi,
>
> when writing the server side with DO I do (since I copied it from the
> examples I found around the net, including our tutorial):
>
>// we set up distributed objects
>port1 = [NSPort port];
>port2 = [NSP
100 matches
Mail list logo
