On 1/8/26 13:10, Segher Boessenkool wrote:
Hi!
On Thu, Jan 08, 2026 at 12:38:24PM -0700, Sandra Loosemore wrote:
diff --git a/gcc/config/rs6000/darwin.opt b/gcc/config/rs6000/darwin.opt
index 8e1f8f05e6c..bc3344fc730 100644
--- a/gcc/config/rs6000/darwin.opt
+++ b/gcc/config/rs6000/darwin.opt
@@ -20,18 +20,18 @@
; <http://www.gnu.org/licenses/>.
Waltivec-long-deprecated
-Driver Alias(mwarn-altivec-long)
+Driver Undocumented Alias(mwarn-altivec-long)
The original patch submission for this (from 2004,
https://gcc.gnu.org/pipermail/gcc-patches/2004-November/156015.html)
says (as does the current code)
"warning ("use of %<long%> in AltiVec types is deprecated; use %<int%>");"
which is exactyly the point (and such code does not work under -m64,
which then gives the
"error ("use of %<long%> in AltiVec types is invalid for 64-bit code");"
faltivec
-Driver
+Driver Undocumented
The documentation for the Darwin -faltivec optuion already says
"The "-faltivec" option should have been called "-maltivec" all along."
So no, let's please no make it worse :-(
I'm all for getting rid of stupid warnings by improving our code, but I
don't like making things even worse :-(
I'm not sure exactly what your complaint is, here. I am not removing
these options or changing their behavior, just marking them
"Undocumented". These two options are already undocumented; neither has
an entry in the manual or even a short documentation string printed by
--help. The purpose of marking them "Undocumented" in the .opt file is
to indicate that the lack of documentation is *intentional*. I've been
doing this elsewhere for options that are deprecated or obsolete but
still accepted as do-nothing flags for backward compatibility, that give
warnings about being obsolete if you do try to use them, that have been
renamed, etc. The GCC manual is already too long without documenting
options that users shouldn't even be trying to use....
-Sandra
