Hi Brian,

The builds I provide are signed properly to run on other systems. 2.6.0rc1 is 
here:

https://www.dropbox.com/s/dj7dx4swnh1pnas/wsjtx.app.zip?dl=0

Also I posted some instructions for building M1 here:

http://folabs.com/building-wsjt-x-apple-m1.html

best,
alex


> On Jul 3, 2022, at 12:12 PM, Brian Moran via wsjt-devel 
> <wsjt-devel@lists.sourceforge.net> wrote:
> 
> 
> Hi Alex and and Olof; I too have been building M1-specific binaries (for 
> 2.6.0 rc-1).
> 
> There certainly are some caveats in building specifically for M1 chips at 
> this time, starting with specifying some of the CMake variables required: 
> 
> CMAKE_OSX_ARCHITECTURES -> arm64 (this chooses M1-specific binary generation)
> CMAKE_OSX_DEPLOYMENT_TARGET -> 12.0 (a good minimum for M1 code generation -- 
> the 10.2 default doesn't support M1, leads to linker errors)
> 
> These variables must be configured in TWO places, both in the top-level build 
> directory, as well as in the build's wsjtx-build location.
> 
> Then, of course, one must have the dependencies for compiling and linking. 
> One dependency is the M1-specific Qt libraries to link against; one way to 
> install those is via a package manager such as Homebrew 
> (https://formulae.brew.sh/formula/qt@5). 
> 
> Some other current issues in building from the source tarball:
> If one is building the "install" target, the generated App gets put in an 
> inconvenient location 
> If one is building the .DMG ("package" target), then all of the binaries in 
> the DMG must be re-signed with a locally trusted certificate, even to install 
> and test the DMG locally.
> Building hamlib in conjunction with wsjtx (from the source tarball) can 
> require that ./configure be run again in the hamlib src directory to 
> configure platform-specific symbols
> 
> Alex, you might find that if you provide the arm64 binary to someone, you'll 
> also need to get the receiver to re-sign the binary so it runs on their 
> system -- native M1 binaries are a little more locked down than Intel-code 
> binaries.  To avoid that annoyance, there would need to be signed 
> distributions available, or a willingness for the recipient of the binaries 
> to install the Xcode code signing tool to re-sign the binaries.
> 
> Since the Intel-compiled binaries do run on M1 Macs, the need isn't as acute 
> to make M1-specific binaries available... yet. 
> 
> I'd love to hear from others that are building M1-specific arm64 binaries
> 
> -Brian N9ADG
> 
> 
> 
> 
> 
>> On Sun, Jul 3, 2022 at 11:18 AM Olof Johansson via wsjt-devel 
>> <wsjt-devel@lists.sourceforge.net> wrote:
>> Hi Alex,
>> 
>> Do you happen to have your modifications on a public git repo somewhere? I 
>> would like to rebuild from sources locally, if possible.
>> 
>> Also (to the maintainers): what's holding this up from being merged into the 
>> upstream repo? Would be nice to see it making its way in, and I'd love to 
>> see official m1 binaries at some point.
>> 
>> 
>> -Olof
>> 
>> 
>>> On Wed, Mar 30, 2022 at 9:27 AM Alex Lelievre via wsjt-devel 
>>> <wsjt-devel@lists.sourceforge.net> wrote:
>>> Hi there,
>>> 
>>> I’ve been testing my native build of WSJTX 2.5.4 on Apple M1 now for quite 
>>> some time and it works great.  If anyone wants to give it a shot, I can 
>>> post my binary that you can drag and drop into your current install.  Let 
>>> me know if there is any interest.
>>> 
>>> alex 
>>> K6LOT
>>> 
> _______________________________________________
> wsjt-devel mailing list
> wsjt-devel@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/wsjt-devel
_______________________________________________
wsjt-devel mailing list
wsjt-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/wsjt-devel

Reply via email to