Re: [R-SIG-Mac] R-SIG-Mac Digest, Vol 241, Issue 5
Mikael, the OpenMP binaries are entirely unrelated to the R release. Also you cannot put libiomp anywhere else without modification, because it contains the path and is signed. Given that you need additional flags for OpenMP anyway, this is not really an issue, since you have to know what you are doing in the first place so you need several flags to make this work. If you are interested in OpenMP for your package then I'd suggest trying a static iomp build. Cheers, Simon > On Apr 19, 2023, at 1:43 AM, Mikael Jagan wrote: > > The OpenMP binaries provided here > >https://mac.r-project.org/openmp/ > > still have the usr/local/ prefix. Is that intentional? If that prefix were > simply deleted, then users could unpack with (for example) > >tar -xvf openmp-14.0.6-darwin20-Release.tar.gz -C /opt/R/$(uname -m) > > And then not need > >CPPFLAGS += -I/usr/local/include >LDFLAGS += -L/usr/local/lib > > just to include the header or link the shared library. > > Mikael > >> Date: Tue, 18 Apr 2023 15:18:03 +1200 >> From: Simon Urbanek >> To: R list >> Subject: [R-SIG-Mac] Please test R 4.3.0 RC, more information below >> Message-ID: <19b7bfe5-2bb9-491d-99a2-bf572a5b4...@r-project.org> >> Content-Type: text/plain; charset="us-ascii" >> Dear Mac users, >> the next major R release R 4.3.0 is coming very soon so I would like to ask >> you to, please, test the pre-releases from >> https://mac.R-project.org >> such that any possible new issues are detected *before* the release. The >> upcoming release will be for macOS 11 and higher only to leverage new >> compilers (important for modern C++ support) and libraries. As usual, there >> are two releases: one for the older Intel Macs (x86_64) and one for the new >> M1/M2 Macs (arm64) - please install the matching version. >> R users can stop reading here, just test, please! >> Package developers, please check the CRAN result pages for your package to >> make sure it passes checks. Also please note that we are now using a >> universal GNU Fortran compiler that works both on Intel and arm64 Macs - you >> can download an Apple Installer from >> https://mac.r-project.org/tools/ >> There are two major changes for the Intel build relevant to advanced users: >> a) .Platfrom$pkgType is "mac.binary.big-sur-x86_64" since we are targeting a >> new macOS version (this is to match arm64 which uses >> "mac.binary.big-sur-arm64"). It means that binaries for contributed packages >> in repositories live in bin/macosx/big-sur-x86_64/contrib/4.3. >> b) the Intel build now matches the arm64 build such that dependent libraries >> are expected in /opt/R/x86_64 (to match arm64's /opt/R/arm64) to avoid >> conflicts with other tools. You can still add -I/usr/local/include to >> CPPFLAGS or -L/usr/local/lib to LDFLAGS if you want to use 3rd party >> libraries from /usr/local (refer to R and package documentation for details). >> Finally, R will no longer sets the macos-version-min flag by default, which >> means R will build for the currently installed macOS version. This allows >> users to build packages that use the latest SDK and capabilities of the >> latest macOS without any changes. However, if you are distributing package >> binaries for use others, we strongly recommend that you set >> export MACOSX_DEPLOYMENT_TARGET=11.0 >> when compiling binaries which will make sure that they will run on macOS 11 >> and higher, otherwise your binary may not be compatible. >> CRAN also uses the macOS11 SDK via >> export SDKROOT=/Library/Developer/CommandLineTools/SDKs/MacOSX11.sdk >> However, depending on your Xcode version you may not have such SDK, so it is >> optional. >> Cheers, >> Simon > ___ R-SIG-Mac mailing list R-SIG-Mac@r-project.org https://stat.ethz.ch/mailman/listinfo/r-sig-mac
Re: [R-SIG-Mac] Please test R 4.3.0 RC, more information below
On 18/04/2023 12:58 p.m., Duncan Murdoch wrote: Hi Simon. I'm seeing crashes when I try to compile rgl and run it in the new R on my Intel notebook. I only installed R, I didn't update all the libs/tools/etc. The crash report is copied below. One oddity I notice on the 4th line of the report: it says it is running "R 4.2.3 GUI 1.79 Big Sur Intel build (8220)". When I run R.app and `version`, I see > version _ platform x86_64-apple-darwin20 arch x86_64 os darwin20 system x86_64, darwin20 status RC major 4 minor 3.0 year 2023 month 04 day13 svn rev84266 language R version.string R version 4.3.0 RC (2023-04-13 r84266) nickname Already Tomorrow but when I choose "About R" in the GUI, I see 4.2.3 reported. Maybe I've got both installed, and that's the cause of the crash? Duncan Murdoch Sorry, I think that was (mostly) a false alarm. I may have had some old object files which weren't recompiled after the install; at least after a new clean build I'm not seeing the crash any more. The bad version report in R.app is still there. Duncan Murdoch ___ R-SIG-Mac mailing list R-SIG-Mac@r-project.org https://stat.ethz.ch/mailman/listinfo/r-sig-mac
Re: [R-SIG-Mac] Please test R 4.3.0 RC, more information below
Hi Simon. I'm seeing crashes when I try to compile rgl and run it in the new R on my Intel notebook. I only installed R, I didn't update all the libs/tools/etc. The crash report is copied below. One oddity I notice on the 4th line of the report: it says it is running "R 4.2.3 GUI 1.79 Big Sur Intel build (8220)". When I run R.app and `version`, I see > version _ platform x86_64-apple-darwin20 arch x86_64 os darwin20 system x86_64, darwin20 status RC major 4 minor 3.0 year 2023 month 04 day13 svn rev84266 language R version.string R version 4.3.0 RC (2023-04-13 r84266) nickname Already Tomorrow but when I choose "About R" in the GUI, I see 4.2.3 reported. Maybe I've got both installed, and that's the cause of the crash? Duncan Murdoch - Translated Report (Full Report Below) - Process: R [73019] Path: /Applications/R.app/Contents/MacOS/R Identifier:org.R-project.R Version: R 4.2.3 GUI 1.79 Big Sur Intel build (8220) Code Type: X86-64 (Native) Parent Process:launchd [1] User ID: 501 Date/Time: 2023-04-18 12:52:03.3210 -0400 OS Version:macOS 12.6.5 (21G531) Report Version:12 Anonymous UUID:2FE3AC85-3E2D-5190-5592-56ECA0EE7FA1 Sleep/Wake UUID: 26969829-5168-4FF8-8E98-CBCCF23BAFDB Time Awake Since Boot: 23 seconds Time Since Wake: 2451 seconds System Integrity Protection: enabled Crashed Thread:0 Dispatch queue: com.apple.main-thread Exception Type:EXC_BAD_ACCESS (SIGSEGV) Exception Codes: KERN_INVALID_ADDRESS at 0x0004c9f0 Exception Codes: 0x0001, 0x0004c9f0 Exception Note:EXC_CORPSE_NOTIFY Termination Reason:Namespace SIGNAL, Code 11 Segmentation fault: 11 Terminating Process: exc handler [73019] VM Region Info: 0x4c9f0 is not in any region. Bytes before following region: 4414989840 REGION TYPESTART - END [ VSIZE] PRT/MAX SHRMOD REGION DETAIL UNUSED SPACE AT START ---> __TEXT 1072c3000-10734f000[ 560K] r-x/r-x SM=COW ...tents/MacOS/R Thread 0 Crashed:: Dispatch queue: com.apple.main-thread 0 libR.dylib 0x124689ab9 Rf_install + 73 1 rgl.so 0x11ff0f504 rgl::X11GUIFactory::createWindowImpl(rgl::Window*) + 52 (x11gui.cpp:655) 2 rgl.so 0x11ff00806 rgl::Window::Window(rgl::View*, rgl::GUIFactory*) + 118 (gui.cpp:162) 3 rgl.so 0x11fedf294 rgl::Device::Device(int, bool) + 148 (device.cpp:16) 4 rgl.so 0x11fedfec5 rgl::DeviceManager::createTestWindow() + 37 (devicemanager.cpp:167) 5 rgl.so 0x11ff00f39 rgl_init + 313 (init.cpp:97) 6 libR.dylib 0x10799dfde R_doDotCall + 862 (dotcode.c:877) 7 libR.dylib 0x1079ee818 bcEval + 105112 (eval.c:8002) 8 libR.dylib 0x1079d4713 Rf_eval + 531 (eval.c:1013) 9 libR.dylib 0x1079f47ac R_execClosure + 2284 10 libR.dylib 0x1079f34e3 Rf_applyClosure + 483 (eval.c:2113) 11 libR.dylib 0x1079dba87 bcEval + 27911 (eval.c:7414) 12 libR.dylib 0x1079d4713 Rf_eval + 531 (eval.c:1013) 13 libR.dylib 0x1079f47ac R_execClosure + 2284 14 libR.dylib 0x1079f34e3 Rf_applyClosure + 483 (eval.c:2113) 15 libR.dylib 0x1079dba87 bcEval + 27911 (eval.c:7414) 16 libR.dylib 0x1079d4713 Rf_eval + 531 (eval.c:1013) 17 libR.dylib 0x1079f2bac forcePromise + 172 (eval.c:833) 18 libR.dylib 0x107a00434 FORCE_PROMISE + 8 (eval.c:5467) [inlined] 19 libR.dylib 0x107a00434 getvar + 372 (eval.c:5508) 20 libR.dylib 0x1079d8b5f bcEval + 15839 (eval.c:7198) 21 libR.dylib 0x1079d4713 Rf_eval + 531 (eval.c:1013) 22 libR.dylib 0x1079f2bac forcePromise + 172 (eval.c:833) 23 libR.dylib 0x107a00434 FORCE_PROMISE + 8 (eval.c:5467) [inlined] 24 libR.dylib 0x107a00434 getvar + 372 (eval.c:5508) 25 libR.dylib 0x1079d8b5f bcEval + 15839 (eval.c:7198) 26 libR.dylib 0x1079d4713 Rf_eval + 531 (eval.c:1013) 27 libR.dylib 0x1079f2bac forcePromise + 172 (eval.c:833) 28 libR.dylib
Re: [R-SIG-Mac] R-SIG-Mac Digest, Vol 241, Issue 5
The OpenMP binaries provided here https://mac.r-project.org/openmp/ still have the usr/local/ prefix. Is that intentional? If that prefix were simply deleted, then users could unpack with (for example) tar -xvf openmp-14.0.6-darwin20-Release.tar.gz -C /opt/R/$(uname -m) And then not need CPPFLAGS += -I/usr/local/include LDFLAGS += -L/usr/local/lib just to include the header or link the shared library. Mikael Date: Tue, 18 Apr 2023 15:18:03 +1200 From: Simon Urbanek To: R list Subject: [R-SIG-Mac] Please test R 4.3.0 RC, more information below Message-ID: <19b7bfe5-2bb9-491d-99a2-bf572a5b4...@r-project.org> Content-Type: text/plain; charset="us-ascii" Dear Mac users, the next major R release R 4.3.0 is coming very soon so I would like to ask you to, please, test the pre-releases from https://mac.R-project.org such that any possible new issues are detected *before* the release. The upcoming release will be for macOS 11 and higher only to leverage new compilers (important for modern C++ support) and libraries. As usual, there are two releases: one for the older Intel Macs (x86_64) and one for the new M1/M2 Macs (arm64) - please install the matching version. R users can stop reading here, just test, please! Package developers, please check the CRAN result pages for your package to make sure it passes checks. Also please note that we are now using a universal GNU Fortran compiler that works both on Intel and arm64 Macs - you can download an Apple Installer from https://mac.r-project.org/tools/ There are two major changes for the Intel build relevant to advanced users: a) .Platfrom$pkgType is "mac.binary.big-sur-x86_64" since we are targeting a new macOS version (this is to match arm64 which uses "mac.binary.big-sur-arm64"). It means that binaries for contributed packages in repositories live in bin/macosx/big-sur-x86_64/contrib/4.3. b) the Intel build now matches the arm64 build such that dependent libraries are expected in /opt/R/x86_64 (to match arm64's /opt/R/arm64) to avoid conflicts with other tools. You can still add -I/usr/local/include to CPPFLAGS or -L/usr/local/lib to LDFLAGS if you want to use 3rd party libraries from /usr/local (refer to R and package documentation for details). Finally, R will no longer sets the macos-version-min flag by default, which means R will build for the currently installed macOS version. This allows users to build packages that use the latest SDK and capabilities of the latest macOS without any changes. However, if you are distributing package binaries for use others, we strongly recommend that you set export MACOSX_DEPLOYMENT_TARGET=11.0 when compiling binaries which will make sure that they will run on macOS 11 and higher, otherwise your binary may not be compatible. CRAN also uses the macOS11 SDK via export SDKROOT=/Library/Developer/CommandLineTools/SDKs/MacOSX11.sdk However, depending on your Xcode version you may not have such SDK, so it is optional. Cheers, Simon ___ R-SIG-Mac mailing list R-SIG-Mac@r-project.org https://stat.ethz.ch/mailman/listinfo/r-sig-mac
Re: [R-SIG-Mac] Please test R 4.3.0 RC, more information below
On 18/04/2023 11:01, Göran Broström wrote: About the Fortran compiler: Den 2023-04-18 kl. 05:18, skrev Simon Urbanek: Dear Mac users, Package developers, please check the CRAN result pages for your package to make sure it passes checks. Also please note that we are now using a universal GNU Fortran compiler that works both on Intel and arm64 Macs - you can download an Apple Installer from https://mac.r-project.org/tools/ I have GNU Fortran (GCC) 12.0.1 20220312 (experimental) in /opt/R/arm64/gfortran/bin/gfortran Should it be uninstalled/removed in some way before installing gfortran-12.2-universal.pkg, and if so, how? Not necessary, as the newer compiler installs into /opt/gfortran. You may need to change FC and FLIBS in a ~/.R/Makevars if you use one, and you may need to adjust your path if you use Fortran other than for R packages. To remove the older version, delete /opt/R/arm64/gfortran/ . -- Brian D. Ripley, rip...@stats.ox.ac.uk Emeritus Professor of Applied Statistics, University of Oxford ___ R-SIG-Mac mailing list R-SIG-Mac@r-project.org https://stat.ethz.ch/mailman/listinfo/r-sig-mac
Re: [R-SIG-Mac] Please test R 4.3.0 RC, more information below
About the Fortran compiler: Den 2023-04-18 kl. 05:18, skrev Simon Urbanek: Dear Mac users, Package developers, please check the CRAN result pages for your package to make sure it passes checks. Also please note that we are now using a universal GNU Fortran compiler that works both on Intel and arm64 Macs - you can download an Apple Installer from https://mac.r-project.org/tools/ I have GNU Fortran (GCC) 12.0.1 20220312 (experimental) in /opt/R/arm64/gfortran/bin/gfortran Should it be uninstalled/removed in some way before installing gfortran-12.2-universal.pkg, and if so, how? Thanks, Göran Cheers, Simon ___ R-SIG-Mac mailing list R-SIG-Mac@r-project.org https://stat.ethz.ch/mailman/listinfo/r-sig-mac ___ R-SIG-Mac mailing list R-SIG-Mac@r-project.org https://stat.ethz.ch/mailman/listinfo/r-sig-mac