Hi All,
Is there any reason to keep support for g77? The last release was in 2006 and gfortran has been available since 2005. I admit that there is no reason to drop current support, apart from getting rid of some code, so perhaps the question should be: how much work should we spend maintaining it?
Chuck
On Mon, Dec 1, 2014 at 7:05 AM, Charles R Harris charlesr.harris@gmail.com wrote:
Hi All,
Is there any reason to keep support for g77?
Yes. The Windows binary builds still use it. We seem to be getting there with the Mingw64 toolchain and gfortran, but I'd like to see that proven (i.e. do a release with it and have that out for >6 months) before considering dropping g77.
The last release was in 2006 and gfortran has been available since 2005. I admit that there is no reason to drop current support, apart from getting rid of some code, so perhaps the question should be: how much work should we spend maintaining it?
AFAIK gh-5315 is the first PR in a long time that requires extra work for g77. I'd say that the current extra workload is acceptable.
Ralf
Charles R Harris charlesr.harris@gmail.com wrote:
Is there any reason to keep support for g77? The last release was in 2006 and gfortran has been available since 2005. I admit that there is no reason to drop current support, apart from getting rid of some code, so perhaps the question should be: how much work should we spend maintaining it?
Until we have Carl Kleffner's gfortran toolchain ready the Windows build of SciPy depends on it.
Accelerate Framework on MacOS X also uses g77 ABI.
Sturla
-
Charles R Harris -
Ralf Gommers -
Sturla Molden