Iain, Thanks once again for your quick reply.
This problem, which at first seemed simple, has gotten a bit more complicated, and consumed a huge amount of my head-scratching and time. Your help should get me to the solution quickly now. The only versions that my usual source has that are on your list of having the initialization bug fixed seem to be for Macs with the Apple M1 or M2 chip. So I’ll look elsewhere for this fix. Once again, thank you very much! Leigh > On Jul 22, 2023, at 1:14 PM, Iain Sandoe <i...@sandoe.co.uk> wrote: > > Hi Leigh, > >> On 22 Jul 2023, at 20:05, Leigh House <ha...@newmexico.com> wrote: >> > >> Thanks for your fast reply! >> >> My Mac is an intel iMac from 2019. > > That should be fine with “upstream” sources, should you decide to build from > source (but that should not be necessary, there are several places providing > gfortran for mac). > >> I didn’t keep detailed notes about where I got the compiler package, though >> I’ve often gone to hpc.sourceforge.net in the past, under “Computation >> Tools”. I suspect that is where I got my current compiler from. >> >> And I didn’t keep notes about which .dylib file I had to find and copy into >> /usr/local/lib, though from a colleague’s experience it may have been >> libgfortran.5.dylib. > > I checked the version(s) for which the initialization bug is fixed: > > GCC-10.4 > GCC-11.3 <<- so you are almost certainly seeing it with 11.2. > GCC-12.1 > GCC-13.1 > (and current development ’trunk’). > > This was a very unusual case - we try to be backward compatible as much as > possible, but the change was out of our hands. > > So, I’d recommend that you see if your “usual source” has an update - or, > alternately, go to one of the ‘OSS distributions’ like Homebrew. > > HTH, > Iain > >
