Re: gcc12 not building on macOS 15 aarch64 (gfortran woes)

[Jason Bacon <jtocino%gmx.com@localhost>]

On 2/23/25 07:35, Jason Bacon wrote:
> On 2/23/25 07:02, Jason Bacon wrote:
> > On 2/22/25 19:46, Greg Troxel wrote:
> > > Some data points (pkgsrc-2024Q4 with mk/ and lang/ at -current):
> > >
> > > macOS 10.13 x86_64: OK: 9 10 12 13. FAIL: 14
> > > macOS 13 x86_64: FAIL: 7 8 9. OK: 10 12 13 14
> > > macOS 15 x86_64: [will run tests tomorrow]
> > >
> > > It would be useful for you to update to current and try 13 again; it
> > > seems ok patch-wise in cvs.
> > >
> > > Also, can anyone test on macOS 15 aarch64, or with older macOS versions?
> > > (I realize "stuck on X" has probably not yet arrived with aarch64 macs -
> > > but surely that's coming soon.)
> > >
> > >
> > >
> > > This is making me lean to requiring 14 for the macos/aarch64 case,
> > > leaving macos/x86-64 at 12, and also hoping someone fixes 12 or the
> > > apple toolchain, whichever is buggy.
> > >
> > > (I remember that we have precedents where aarch64 needs newer versions.)
> >
> > I just in the process of testing gcc14 with the same idea, to make it 
> > default on aarch64, when I experienced a failure that looks like the 
> > same one we discussed several weeks ago.  I thought this had been 
> > fixed since then, and in fact gcc14 was installed and working until 
> > pkg_chk removed it last night.
> >
> > macOS:
> > Software:
> >
> >      System Software Overview:
> >
> >        System Version: macOS 15.3 (24D60)
> >        Kernel Version: Darwin 24.3.0
> >        Boot Volume: Macintosh HD
> >        Boot Mode: Normal
> >        Computer Name: Administrator’s Mac mini
> >        User Name: Jason Bacon (bacon)
> >        Secure Virtual Memory: Enabled
> >        System Integrity Protection: Enabled
> >        Time since boot: 10 days, 10 hours, 58 minutes
> >
> > Xcode command-line tools:
> > version: 16.2.0.0.1.1733547573
> >
> >
> > => Applying pkgsrc patch /Users/bacon/Pkgsrc/pkgsrc/lang/gcc14/../../ 
> > lang/gcc14/patches/patch-libstdc++-v3_libsupc++_new__opa.cc
> > Hmm...  Looks like a unified diff to me...
> > The text leading up to this was:
> > --------------------------
> > |$NetBSD: patch-libstdc++-v3_libsupc++_new__opa.cc,v 1.1 2024/06/28 
> > 18:35:16 wiz Exp $
> > |
> > |Ensure we can find SunOS std::aligned_alloc if using it.
> > |
> > |--- libstdc++-v3/libsupc++/new_opa.cc.orig     2024-05-21 
> > 07:47:42.000000000 +0000
> > |+++ libstdc++-v3/libsupc++/new_opa.cc
> > --------------------------
> > Patching file libstdc++-v3/libsupc++/new_opa.cc using Plan A...
> > Hunk #1 succeeded at 62.
> > done
> > ERROR: Patching failed due to modified or broken patch file(s):
> > ERROR: /Users/bacon/Pkgsrc/pkgsrc/lang/gcc14/../../lang/gcc14/patches/ 
> > patch-libgcc_config.host
> > ERROR: 
> > ==========================================================================
> > ERROR:
> > ERROR: Some of the selected build options and/or local patches may be 
> > incompatible.
> > ERROR: Please try building with fewer options or patches.
> > ERROR:
> > ERROR: 
> > ==========================================================================
> > *** Error code 1
> >
> > Stop.
> > bmake[8]: stopped making "reinstall" in /Users/bacon/Pkgsrc/pkgsrc/ 
> > lang/ gcc14
> > *** Error code 1
>
> I just resurrected wip/gcc14, and it works fine.  It also has 76 more 
> patches than lang/gcc14:
>
> Darwin tarpon.acadix  bacon ~/Pkgsrc/pkgsrc/wip/gcc14 1022: (pkgsrc): ls 
> patches/ | wc -l
>        98
> Darwin tarpon.acadix  bacon ~/Pkgsrc/pkgsrc/wip/gcc14 1023: (pkgsrc): 
> ls ../../lang/gcc14/patches/ | wc -l
>        22

Scratch that...

My CVS tree was corrupted somehow, maybe a disk full during an update, 
so some of the patches were missing from lang/gcc14.  Retesting it now 
and will report back IFF there are failures.


--
Life is a game.  Play hard.  Play fair.  Have fun.