[R-pkg-devel] Minimum macOS version ?
One package I stand behind as maintainer does
## Take care of 10.14 requirement for Intel macOS
if test x"${uname}" = x"Darwin" -a x"${machine}" = x"x86_64"; then
AC_MSG_CHECKING([for Darwin x86_64 use minimum version override])
CXX17_MACOS="-mmacosx-version-min=10.14"
AC_MSG_RESULT([${CXX17_MACOS}])
fi
We now contemplate going to 10.15 as minimum macOS version. Is that an issue
for CRAN?
Apologies for the somewhat ignorant question but I was not aware where I
could have looked this up. So thanks in advance for any pointers.
Thanks, Dirk
--
dirk.eddelbuettel.com | @eddelbuettel | e...@debian.org
__
R-package-devel@r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/r-package-devel
Re: [R-pkg-devel] Minimum macOS version ?
Dirk,
the minimum required version for the high-sierra build is 10.13 and for big-sur
build is 11.0 (as the names imply). Although it is not unrealistic to move
Intel to macOS 10.14, it would be more problematic to move to 10.15 since it is
the version that killed 32-bit support so 10.14 is actually more prevalent than
10.15 as a lot of users will never upgrade due to that loss. Given that Intel
binaries are there only to support legacy hardware it's unlikely to move
forward much more (i.e. I can see the immediate argument saying that those
macOS version are unsupported, but so are the machines running them, but that
doesn't mean R users won't use them).
What exactly do you need the higher macOS versions for? Generally, there were
no major changes in the Mach-O format recently, so it shouldn't really matter -
the main difference would be SDK/run-time - are there specific features you
require? Some of it could be addressed by providing more recent run-times
separately - that's why I'm asking.
Cheers,
Simon
> On 9/12/2022, at 9:19 AM, Dirk Eddelbuettel wrote:
>
>
> One package I stand behind as maintainer does
>
> ## Take care of 10.14 requirement for Intel macOS
> if test x"${uname}" = x"Darwin" -a x"${machine}" = x"x86_64"; then
> AC_MSG_CHECKING([for Darwin x86_64 use minimum version override])
> CXX17_MACOS="-mmacosx-version-min=10.14"
> AC_MSG_RESULT([${CXX17_MACOS}])
> fi
>
> We now contemplate going to 10.15 as minimum macOS version. Is that an issue
> for CRAN?
>
> Apologies for the somewhat ignorant question but I was not aware where I
> could have looked this up. So thanks in advance for any pointers.
>
> Thanks, Dirk
>
> --
> dirk.eddelbuettel.com | @eddelbuettel | e...@debian.org
>
> __
> R-package-devel@r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-package-devel
>
__
R-package-devel@r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/r-package-devel
Re: [R-pkg-devel] Minimum macOS version ?
Simon, On 9 December 2022 at 10:00, Simon Urbanek wrote: | the minimum required version for the high-sierra build is 10.13 and for big-sur build is 11.0 (as the names imply). Although it is not unrealistic to move Intel to macOS 10.14, it would be more problematic to move to 10.15 since it is the version that killed 32-bit support so 10.14 is actually more prevalent than 10.15 as a lot of users will never upgrade due to that loss. Given that Intel binaries are there only to support legacy hardware it's unlikely to move forward much more (i.e. I can see the immediate argument saying that those macOS version are unsupported, but so are the machines running them, but that doesn't mean R users won't use them). | | What exactly do you need the higher macOS versions for? Generally, there were no major changes in the Mach-O format recently, so it shouldn't really matter - the main difference would be SDK/run-time - are there specific features you require? Some of it could be addressed by providing more recent run-times separately - that's why I'm asking. In this case: std::filesystem which is from C++20 and errors on macOS. We will #ifdef it out of existence on that platform but use it on other platforms where we can. Thaks, Dirk -- dirk.eddelbuettel.com | @eddelbuettel | e...@debian.org __ R-package-devel@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-package-devel
Re: [R-pkg-devel] Minimum macOS version ?
Dirk, thanks, got it. If you have a work-around avoiding std::filesystem then can use something like // no filesystem support before macOS Catalina #if defined(__MAC_OS_X_VERSION_MIN_REQUIRED) && __MAC_OS_X_VERSION_MIN_REQUIRED < 101500 // no std::filesystem #endif (derived from https://github.com/nlohmann/json/pull/3101 - some of the comments there may be of use - it also mentions MinGW etc.) The "offical" macros for macOS availability tests are AvailabilityMacros.h but they boil down to the above. Cheers, Simon > On 9/12/2022, at 11:10 AM, Dirk Eddelbuettel wrote: > > > Simon, > > On 9 December 2022 at 10:00, Simon Urbanek wrote: > | the minimum required version for the high-sierra build is 10.13 and for > big-sur build is 11.0 (as the names imply). Although it is not unrealistic to > move Intel to macOS 10.14, it would be more problematic to move to 10.15 > since it is the version that killed 32-bit support so 10.14 is actually more > prevalent than 10.15 as a lot of users will never upgrade due to that loss. > Given that Intel binaries are there only to support legacy hardware it's > unlikely to move forward much more (i.e. I can see the immediate argument > saying that those macOS version are unsupported, but so are the machines > running them, but that doesn't mean R users won't use them). > | > | What exactly do you need the higher macOS versions for? Generally, there > were no major changes in the Mach-O format recently, so it shouldn't really > matter - the main difference would be SDK/run-time - are there specific > features you require? Some of it could be addressed by providing more recent > run-times separately - that's why I'm asking. > > In this case: std::filesystem which is from C++20 and errors on macOS. > > We will #ifdef it out of existence on that platform but use it on other > platforms where we can. > > Thaks, Dirk > > -- > dirk.eddelbuettel.com | @eddelbuettel | e...@debian.org > __ R-package-devel@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-package-devel
