[MacPorts] #74191: Argyll 3.5.0 fails on M4max / Xcode CLT 26 --> workaround
MacPorts
noreply at macports.org
Mon Jul 6 20:56:31 UTC 2026
#74191: Argyll 3.5.0 fails on M4max / Xcode CLT 26 --> workaround
--------------------------------+-----------------------------
Reporter: MStraeten | Owner: (none)
Type: defect | Status: new
Priority: Low | Milestone: MacPorts Future
Component: ports | Version: 2.12.5
Keywords: m4max xcode clt 26 | Port: argyll
--------------------------------+-----------------------------
On MacBook M4 Max, Xcode CLT 26 GenRMGam crashes with EXC_BAD_ACCESS
during recursive tree traversal in radial_point_triang() (gamut.c around
line 3989).
It seems to be triggered by compiler optimization (-Os / -O2).
Workaround: Lowering optimization to -O0 completely resolves the crash.
{{{
sudo port install argyll configure.optflags="-O0"
}}}
longterm solution: add
{{{
configure.optflags -O0
}}}
to the Portfile - but not sure, if that's appropriate when just one
failing processor/sdk combination is known
Background (analysis by gemini):
When optimization flags like -Os (optimize for size) or -O2 are active,
the compiler makes aggressive assumptions about pointer arithmetic and
memory layouts to make the machine code as compact and fast as possible.
In gamut.c — which heavily relies on complex tree structures and recursive
quadtrees — the new Clang compiler erroneously optimizes away critical
safety checks (such as if (np != NULL)) or miscalculates pointer offsets.
It likely presumes certain memory states are "undefined behavior" or
impossible. Because the M4 Max uses a massive, differently mapped virtual
address space compared to older Intel or M1 chips, these bad compiler
assumptions cause the pointers to jump to invalid addresses (e.g., jumping
from 0x00000001... to 0x00000008...), triggering an immediate
EXC_BAD_ACCESS crash.
--
Ticket URL: <https://trac.macports.org/ticket/74191>
MacPorts <https://www.macports.org/>
Ports system for macOS
More information about the macports-tickets
mailing list