On Friday, July 31, 2020 at 3:51:45 PM UTC-7, Dima Pasechnik wrote: > > On Fri, Jul 31, 2020 at 11:41 PM Samuel Lelièvre > <[email protected] <javascript:>> wrote: > > > > 2020-07-31 22:26 UTC, Dima Pasechnik: > > > >> I decided to configure with > > > >> > > > >> --with-system-gcc=no --with-system-gfortran=no > > > > > > --with-system-gcc must be yes on macOS, I think. > > > You actually cannot build Sage with "real" gcc/g++ on macOS, > > > you need system's gcc/g++, which is actually not gcc/g++, but > clang/clang++ > > > > > > system-gcc will come from Xcode, not from Homebrew. > > > > In fact, you can set `--with-system-gcc=no`, and Sage will build > > its own `gcc` (using the system `clang` to do so), and then use > > that `gcc` and its `gfortran` to build most other packages. > > Does this still work?
Not "still", rather "again", i.e., since Sage 9.1. This configuration is called "homebrew-macos-standard-gcc_spkg". https://github.com/sagemath/sage/runs/910575934 Unfortunately, there are indeed several packages that do not succeed to build with the installed gcc, as you can see in this run. Some of these build failures are already tracked. For example: https://trac.sagemath.org/ticket/29613 for givaro. I have marked it as "minor" because, as you say, there is little point in using the gcc from the SPKG when both the XCode "gcc" (clang) and the homebrew compilers are better options. Samuel -- please do open the ticket - for the broken logic of the --with-system-XYZ arguments. -- You received this message because you are subscribed to the Google Groups "sage-devel" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. To view this discussion on the web visit https://groups.google.com/d/msgid/sage-devel/e70097f3-21a4-49d6-b815-905db447be5fo%40googlegroups.com.
