On Thu, 24 Mar 2022 18:58:57 +0000 Brian Callahan <bcal...@posteo.net> wrote:
> > Index: patches/patch-configure_ac > > =================================================================== > > RCS file: /cvs/ports/lang/algol68g/patches/patch-configure_ac,v > > retrieving revision 1.2 > > diff -u -p -r1.2 patch-configure_ac > > --- patches/patch-configure_ac 13 Mar 2022 21:03:42 -0000 1.2 > > +++ patches/patch-configure_ac 24 Mar 2022 09:06:28 -0000 > > @@ -1,4 +1,4 @@ > > -Enable build on arm64 and armv7 > > +Enable build on all archs > > Build the compiler > > Mention that we are IEEE 754 compliant > > Remove hardcoded optimization flags > > @@ -12,7 +12,7 @@ Index: configure.ac > > # OpenBSD. > > # > > -*86-*-openbsd* | *86_64-*-openbsd*) > > -+*86-*-openbsd* | *86_64-*-openbsd* | arm*-*-openbsd* | aarch*-*-openbsd*) > > ++*-*-openbsd*) > > AC_DEFINE(BUILD_BSD, 1, [Define this if OpenBSD was detected]) > > - AC_MSG_WARN([configuring interpreter-only]) > > + AC_DEFINE(HAVE_IEEE_754, 1, [Define this if IEEE_754 compliant]) > > > > > > It was set this way because everyone was under the impression the > compiler only worked in x86 and arm. If it works universally, go for it. jca's diff also fixes the build on powerpc and powerpc64. The compiler might work because src/a68g/compiler.c emits C code. The powerpc is a PowerBook G4 running at 750 MHz, which might be too slow for the time limit. It got 3 test failures, a68g: runtime error: 1: time limit exceeded... FAIL: test-set/13-math.a68 a68g: runtime error: 1: time limit exceeded... FAIL: test-set/14-mersenne.a68 a68g: runtime error: 1: too many parallel units (platform supports 4 parallel units), ... FAIL: test-set/20-quicksort-parallel.a68 The powerpc64 got only 1 failure, 20-quicksort-parallel.a68 --George
