The kit is built and packaged. Only Xcode code highlighting and autocomplete
work though. I will try tap into Xcode 5 bots after it is gold and maybe use
that as a remote debug/build framework.
发自我的 iPad
在 2013-6-17，12:37，Chan Maxthon 写道：
> Xcode can cross compile, but only if there is a
Xcode can cross compile, but only if there is a way for it to run the resulting
code. However by properly implementing a fake SDK folder, you can make its code
highlighting and autocomplete work. I have already ported a set of basic Linux
headers over and it worked fine. Now I am going to copy l
How about use SenTestingKit and CXXTest? Since I am building cross tools to
build GNUstep code targeting Linux on OS X for Xcode 4.6 and Xcode 5, those
kits from 4.6 sounds reasonable to use for that purpose.
发自我的 iPad
在 2013-6-17，9:50，Stefan Bidi 写道：
> I'd like to extend gnustep-tests to be
Mark 1 solved by someone else - there is an existing distribution of
Linux-targeting tool chain for OS X. I am using it out of the box. Sadly it
uses GCC so I need a cross-compiling clang. I do have clang 3.3 source code and
building one targeting Linux amd64 should be trivial. I think the bigge
Do you guys use LTO? I find its performance boost impressive. When I am
bootstrapping clang, I compared performance of clang built with gcc (-O3, no
LTO, but gcc have a better optimizer) and LLVM LTO (-O3 -flto, but LLVM
optimizer is less aggressive than gcc's) and it gave my clang a 50%~100%
p
You may want to focusing on translating LLVM bitcode to that target - clang
comes with full C, C++ and Objective-C support and by implementing that your
project will be able to translate a lot more programs in a more sophisticated
level to your target platform.
发自我的 iPhone
在 2013-6-10，21:27，bu
g's methods.
>
> Lubos
>
>
>
> Dne 9. června 2013 0:08:40 Chan Maxthon napsal:
>> Can you just dump it from -[NSString dataUsingEncoding:] which is way
>> simpler?
>>
>> 发自我的 iPhone
>>
>> 在 2013-6-9，6:06，Luboš Doležel 写道：
>>
"figuring out if a pointer points to a small or regular
> object"...
>
> On Sat, Jun 8, 2013 at 11:51 PM, Chan Maxthon wrote:
>> I am sort of looking into the concept of running Objective-C in a kernel of
>> an operating system, which would require implementing a
Can you just dump it from -[NSString dataUsingEncoding:] which is way simpler?
发自我的 iPhone
在 2013-6-9，6:06，Luboš Doležel 写道：
> Hi,
>
> GNUstep's NSString currently lacks
> getBytes:maxLength:usedLength:encoding:options:range:remainingRange:
>
> (added in OS X 10.5).
>
> I wanted to implemen
only the linking would
ask for a Linux box. This is why I am asking this question.
发自我的 iPhone
在 2013-6-9，5:15，Ivan Vučica 写道：
> On Sat, Jun 8, 2013 at 9:55 PM, Chan Maxthon wrote:
>> I am expecting a solution that works on both OS X and GNUstep.
>
> Why?
>
> This is an
I am expecting a solution that works on both OS X and GNUstep.
发自我的 iPhone
在 2013-6-9，1:28，David Chisnall 写道：
> No. See the email you replied to for why not.
>
> David
>
> On 8 Jun 2013, at 18:20, Maxthon Chan wrote:
>
>> So I can safely do this? (Macro rewritten as inline function)
>>
>>
Indeed I am not that sure about how GNUstep currently handles exceptions but I
am giving you ideas from some crash reports with symbols from an Apple system.
And the conclusion on toll-free bridging is about the latest Apple
implementation. You should really get a Mac and poke around. Just for y
Just wondering, do you guys need/want a CoreLocation clone? That framework is
more or less useless on computers, but a extremely useful for mobile devices,
and I know there is an effort in porting GNUstep onto Android.
发自我的 iPad
___
Gnustep-dev mailing
That is, if I have an existing OS X app for 10.8 (I #ifndef GNUSTEP out'ed all
Apple-specific code) I can port and build it as-is, with all xib files in place?
发自我的 iPad
在 2013-6-4，3:30，Gregory Casamento 写道：
> Gorm doesn't understand nibs beyond 10.5's nib format. Beyond this nibs are,
> pri
This is what I am describing. Apple is still using the NeXT bundle layout that
is currently used by GNUstep, but only on iOS. On OS X the new Cocoa layout
(that one with the Contents directory) is used. My suggestion was, once again,
while keeping compatibility, default to the Cocoa layout, mean
SMB/CIFS supported
that. That is, despite NTFS does both symbolic and hard links, I cannot easily
transfer it from a UNIX system (I generally develop on OS X) to a Windows
system.
发自我的 iPhone
在 2013-5-23，1:59，Niels Grewe 写道：
>
> On 22.05.2013 18:47BST Chan Maxthon wrote:
>
>&g
architecture.
发自我的 iPhone
在 2013-5-22，20:55，Niels Grewe 写道：
>
> On 22.05.2013 14:44CEST Chan Maxthon wrote:
>
>> Including OS X? I mean I intend to compile a single bundle, with a single
>> copy of all resources, containing both OS X binaries linked against Apple's
>&g
3 14:38CEST Chan Maxthon wrote:
>
>> I mean not only cross compiling, but also constructing a single bundle with
>> one copy of all resources and numerous binaries for different platforms. It
>> is intended to work right out of the box on all systems that it is compiled
&g
Well I meant build a single bundle with one copy of resources and various
architectures of binaries. That is, a bundle that will simultaneously valid for
OS X, Linux i386, Linux amd64, FreeBSD amd64 and more.
发自我的 iPad
在 2013-5-22，20:32，Niels Grewe 写道：
>
> On 22.05.2013 14:22CES
Since you mentioned this, how about adopt current OS X bundle structure as
default GNUstep structure? The pro of this is that OS X bundle actually can
allow multiple platforms of binaries coexist in a single wrapper. OS X used
Contents/MacOS (and a fat Mach-O binary may reside in it) as the fold
When I tried to compile libobjc2 trunk or 1.7rc1 on Ubuntu server using clang
3.4 svn, clearly the objc++ part is calling for c++11 from its asking for
c++abi library and c++11 headers. However some CMakeFile issues are preventing
it from being built as C++11. A big patch will come up, fixing th
I still have an old server that does not run 64-bit OS. However unit tests tend
to be very problematic on the older 32-bit platform, even though the OS itself
is modern enough (Ubuntu 13.04, Linux 3.8) Can anyone figure out why?
发自我的 iPad
___
Gnustep-d
dding into existing pages. But if you need a beautiful and complex
> web app, this is something you should look at.
>
> And if you'll be writing WebUIKit, why not base it on Objective-J and
> Cappuccino's Foundation library? :-)
>
> Sent from my iPad
>
> On 17. 5
destructor dealloc
is used for CoreFoundation objects as well. Modern CFRetain and CFRelease
functions are mere wrappers around objc_retain and objc_release functions from
libobjc2.
发自我的 iPhone
在 2013-5-18，5:35，Chan Maxthon 写道：
> I discovered this from stack traces of toll-free bridged ca
I discovered this from stack traces of toll-free bridged calls.
All CoreFoundation objects have an Objective-C isa pointer. For toll-free
bridged objects, this pointer points to an Objective-C class which is a
subclass of the corresponding class cluster public class, like for CFArray the
subcla
This have a clearly different aim. Despite starting this server is as simple as
a call -[CGIServer start] but it is intended to be used as independent server.
发自我的 iPad
在 2013-5-17，13:52，Richard Frith-Macdonald 写道：
>
> On 16 May 2013, at 20:50, Maxthon Chan wrote:
>
>> Well it turned out th
Well that thing never compiled for me, using trunk libobjc2 and trunk
llvm/clang on my server, let alone I have portability in mind (Written under OS
X, it is required to build on Linux as well, using trunk libobjc2, trunk
llvm/clang and full Objective-C ARC.). And the reason I spawned this proj
Just thinking, is it possible to instead of implementing everything again, just
maintain the differences and implement GNUstep extensions on top of OS X
frameworks? (That is, use Foundation and AppKit by Apple for Base and GUI, and
only provide API diffs, and also a set of tools to convert Gorm
Okay?
发自我的 iPad
在 2013-4-11，1:18，Fred Kiefer 写道：
> You should rather reply back to the mailing 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 r
Hmm it seemed that my runtime is way outdated. I will try rebuild it and retest
it again later. I need sleep now.
发自我的 iPad
在 2013-4-10，0:45，David Chisnall 写道：
> On 9 Apr 2013, at 17:43, Chan Maxthon wrote:
>
>> Well you may have to stick to that line, as my issue is about
Well you may have to stick to that line, as my issue is about code ported over
from OS X and use gnustep-make to build.
发自我的 iPad
在 2013-4-10，0:41，David Chisnall 写道：
> Yes, that's the output I see. Your proposed compile line is obviously not
> reproduceable, because the output from gnustep-c
You should check the one used in iOS 6.1 as shipped with Xcode 4.6. It is
actually free on Mac App Store.
发自我的 iPhone
在 2013-3-15，17:04，Luboš Doležel 写道：
> On Fri, 15 Mar 2013 16:46:36 +0800, Maxthon Chan wrote:
>> I do have the binary, but I have legal worries about this, unless it
>> is a ki
Being creative, can we implement it in alternative ways, like allowing a
directory to carry a bundle that will not only provide localized folder names
but also much more?
发自我的 iPhone
在 2013-3-14，7:17，Riccardo Mottola 写道：
> Hi,
>
> On 03/13/13 22:41, Fred Kiefer wrote:
>>
>> More information
Thus I am once again suggesting implementing the entire CoreBase in Objective-C
and layer it above Base - that will save us all that hassle and can survive
future Base changes if compiled with non-fragile ABI support.
发自我的 iPhone
在 2013-3-12，1:07，David Chisnall 写道：
> On 11 Mar 2013, at 16:33,
How about reversing the relationship on GNUstep: implement CoreBase in
Objective-C. Everything except CFRetain() and CFRelease() can be built with
ARC, and CF objects are typedef'd to their respective Objective-C counterpart.
Apple invented CF to serve Carbon and Cocoa simultaneously, but we don
Maybe the format string is bad?
发自我的 iPhone
在 2013-2-11，5:01，Steven LeMaire 写道：
> Hi Everyone,
>
> I really don't understand why the code below isn't working for me. I've tried
> setting a locale and timezone on the date formatter, but it's had no affect.
> I'm running this on Slackware 13.3
Really, you should make that connection an instance variable or private
property, and then ARC it.
发自我的 iPhone
在 2013-2-5，18:06，Richard Frith-Macdonald 写道：
>
> On 5 Feb 2013, at 09:42, Riccardo Mottola wrote:
>
>> Maxthon Chan wrote:
>>> Just an idea, will ARC work?
>> I don't know what it A
Actually, it was 4-way: 32-bit x86, amd64, 32-bit PowerPC and to some extent
64-bit PowerPC.
发自我的 iPad
在 2013-1-27，19:35，Ivan Vučica 写道：
> On 26. 1. 2013., at 14:10, Fred Kiefer wrote:
>
>> Apple allows the code to use 32bit versions of the core bundles even on
>> 64bit system. I don't know
Just an idea, can you try link it to the WebKit from Chromium source tree? Or
wrap Chromium WebKit for GNUstep?
发自我的 iPad
在 2013-1-25，16:19，Luboš Doležel 写道：
> Dne 25. ledna 2013 6:11:24 Abhi Beckert napsal:
>> On Thu, Jan 24, 2013 at 10:16 PM, Luboš Doležel wrote:
>>
>> > I tried to make t
My current location and position is just prohibitively expensive and difficult
for me to be there. So I will not be there, but keep me posted please.
发自我的 iPhone
在 2013-1-12，1:42，Quentin Mathé 写道：
> Hi David,
>
> Le 11 janv. 2013 à 10:55, David Chisnall a écrit :
>
>> Hi Everyone,
>>
>> I'm
d based on Apple's CoreFoundation. It
will be less effort to do that if you are using FreeBSD. I have Linux so it is
a lot more work to do.
发自我的 iPad
在 2013-1-9，1:08，Marcus Müller 写道：
>
> On 08.01.2013, at 18:02, Chan Maxthon wrote:
>
>> You can mix ARC and non-ARC code, given
You can mix ARC and non-ARC code, given that those code are placed in separate
files, then link them together. So you can safely implement the missing link in
ARC code using Objective-C categories, and link them together with non-ARC
GNUstep and your project in ARC.
发自我的 iPad
在 2013-1-9，0:55，
Why wouldn't you guys just use libobjc2 and ARC?
发自我的 iPad
在 2013-1-8，23:42，Richard Frith-Macdonald 写道：
>
> On 8 Jan 2013, at 15:35, Pirmin Braun wrote:
>
>> Am Tue, 8 Jan 2013 14:51:07 +
>> schrieb David Chisnall :
>>
>>> On 8 Jan 2013, at 14:41, Pirmin Braun wrote:
>>>
NSSt
Given you are dare enough, will you make a libSystem.dylib replace for Linux,
by packaging Linux libc, libobjc2, libm and libdispatch into a single "new"
libc.so? This may make compiling Objective-C code as straight as on OS X.
Maxthon T. Chan
发自我的 iPad
_
Current chosen parts of the build: svn clang, Apple's Core Foundation, Apple's
libobjc (will have so name libobjc.so.8 since it came off OS X 10.8.2),
Cocotron's Foundation, GNUstep's everything else.
Sent from my iPad
On 2012-10-31, at 4:12, Chan Maxthon wrote:
> Ap
Apple actually opened its Core Foundation and implemented its Foundation kit as
a wrapper around it. Is it possible for GNUstep to incorporate the Apple code,
maybe with a branding exception on the GPL (APSL is basically GPL but with a
branding exception) or forking it? Now I am trying to build
46 matches
Mail list logo
