> On 23 May 2019, at 08:15, David Chisnall <[email protected]> wrote:
>
> Hi,
>
> I'd not looked at the Swift corelibs version of libdispatch. Looking at its
> headers, there are some feature test macros that check if you have
> Objective-C support, which seem to only enable Objective-C interop on Apple
> platforms:
>
> https://github.com/apple/swift-corelibs-libdispatch/blob/631821c2bfcf296995a8424f148a9365470c2210/os/object.h#L58
>
> You could add some checks to there to enable them on non-Mac platforms and
> see what happens when you build them…
Good idea :)
Defining -DOS_OBJECT_HAVE_OBJC_SUPPORT=1 in the ${TOOL_NAME}_CPPFLAGS variable
in GNUmakefile successfully navigated that set of tests, and then it fell over
because dispatch.h #includes <objc/NSObject.h> rather than
<Foundation/NSObject.h>. A symlink fixed that :)
So now Server.m compiles. There’s still quite a long way to go to see if it
works, apparently we don’t have NS{Mutable}OrderedSet in GNUstep just yet, so
I’ll mock one up with an NS{Mutable}Array and an NS{Mutable}Set and see where
it barfs next :)
The board that’s running all this is one of these
<https://www.friendlyarm.com/index.php?route=product/product&path=69&product_id=180>
- I still find it kind of amazing that something this small can run “Cocoa”
apps :)
Cheers!
Simon
>
> David
>
> On 23/05/2019 15:13, Simon Gornall wrote:
>> [I posted this in gnu.gnustep.discuss, but then realised that the last post
>> before mine was back in January. Posting here where there’s more traffic :)]
>> Hi all,
>> So I've just installed Debian (Sid) and got GNUstep installed with runtime
>> 2.0 / clang-8 etc. by following
>> http://wiki.gnustep.org/index.php/Building_GNUstep_under_Debian_FreeBSD#libdispatch
>>
>> <http://wiki.gnustep.org/index.php/Building_GNUstep_under_Debian_FreeBSD#libdispatch>
>> As far as I can tell, it all worked fine. I got a few tests fail where the
>> above link didn't mention, but they seemed to be to do with Objc++, and I'm
>> not planning on using that. I put that down to running sid.
>> I have a project that compiles on my Mac, and I'm trying to port it over -
>> part of it is a web-socket service, which uses libdispatch and blocks to
>> process the data streams. On the Mac, there aren't any warnings. On the
>> linux box, I get:
>> nano:~/src/b-leet.com/bleetd <http://b-leet.com/bleetd>> make messages=yes
>> This is gnustep-make 2.7.0. Type 'make print-gnustep-make-help' for help.
>> Running in gnustep-make version 2 strict mode.
>> Making all for tool bleetd...
>> /usr/bin/clang-8 bleetd/classes/server/Server.m -c \
>> -MMD -MP -Ibleetd/classes/server -Ibleetd/classes/websockets -DGNUSTEP
>> -DGNUSTEP_BASE_LIBRARY=1 -DGNU_GUI_LIBRARY=1 -DGNUSTEP_RUNTIME=1
>> -D_NONFRAGILE_ABI=1 -DGNUSTEP_BASE_LIBRARY=1 -fno-strict-aliasing
>> -fexceptions -fobjc-exceptions -D_NATIVE_OBJC_EXCEPTIONS -pthread -fPIC
>> -Wall -DGSWARN -DGSDIAGNOSE -Wno-import -g -O2 -fblocks
>> -fobjc-runtime=gnustep-2.0 -fblocks -fconstant-string-class=NSConstantString
>> -I. -I/home/simon/GNUstep/Library/Headers -I/usr/local/include \
>> -o obj/bleetd.obj/bleetd/classes/server/Server.m.o
>> In file included from bleetd/classes/server/Server.m:10:
>> In file included from bleetd/classes/websockets/PSWebSocketServer.h:16:
>> bleetd/classes/websockets/PSWebSocket.h:69:1: error: property with 'retain
>> (or strong)' attribute must be of object type
>> @property (nonatomic, strong) dispatch_queue_t delegateQueue;
>> ^
>> In file included from bleetd/classes/server/Server.m:10:
>> bleetd/classes/websockets/PSWebSocketServer.h:47:1: error: property with
>> 'retain (or strong)' attribute must be of object type
>> @property (nonatomic, strong) dispatch_queue_t delegateQueue;
>> ^
>> 2 errors generated.
>> The GNUmakefile I put together looks like:
>> ---8<------8<------8<--- Cut Here ---8<------8<------8<---
>> include $(GNUSTEP_MAKEFILES)/common.make
>> TOOL_NAME = bleetd
>> bleetd_CPPFLAGS = -Ibleetd/classes/server \
>> -Ibleetd/classes/websockets
>> bleetd_OBJC_FILES = $(shell find bleetd -name '*.m')
>> include $(GNUSTEP_MAKEFILES)/tool.make
>> ---8<------8<------8<--- Cut Here ---8<------8<------8<---
>> So the question: Would I normally expect a 'dispatch_queue_t' to still be
>> something that can be put into a property in GNUstep ? I'm really only
>> familiar with the Mac side of things - I’m new to GNUstep, and in Mac-land
>> they're reference-counted objects. I guess I’m asking: is this a bug, did I
>> do something wrong during install, or is it expected behaviour ?
>> Cheers
>> Simon
>> _______________________________________________
>> Discuss-gnustep mailing list
>> [email protected]
>> https://lists.gnu.org/mailman/listinfo/discuss-gnustep
>
> _______________________________________________
> Discuss-gnustep mailing list
> [email protected]
> https://lists.gnu.org/mailman/listinfo/discuss-gnustep
_______________________________________________
Discuss-gnustep mailing list
[email protected]
https://lists.gnu.org/mailman/listinfo/discuss-gnustep
