On Fri, Jan 16, 2026 at 1:38 PM Rainer Orth <[email protected]> wrote: > > Hi Richard, > > >> So unless additional concerns come up, I'll commit this patch in a > >> couple of days. > > > > This probably warrants documenting in gcc-16/changes.html > > that's what I meant to do. I originally planned to refer from the > caveats section there to the patch posting, but that's probably too > low-level for regular users. I'll probably go ahead and add a > user-centric description to a gcc-16 porting_to.html, describing the > interoperability impact with GCC < 16, clang, and Studio 12.6 CC (which > also implements the GNU C++ ABI). I'd forgotten about this in the patch > description, but have already some of this in response to Jonathan's > comment on the ABI break in the PR. > > > Does this mean that ABI-wise non-fixincluded and fixincluded > > versions do not inter-operate? I wonder if it makes sense > > to diagnose int8_t versions that do not adhere to the (updated) > > ABI then (not sure if there's a suitable backend hook). > > It meanwhile occured to me that I might be able to do better than what I > currently have. The idea runs as follows: > > * Make INT8_TYPE etc. compile-time configurable via some -mint8_t-compat > (or -msolaris-compat) option, off by default. > > * Have that predefine __INT8_T_COMPAT__ if on. > > * Wrap the fixed definitions in <sys/int_types.h> in #ifndef > __INT8_T_COMPAT__. > > I guess this might work, but (at least initially) only for C-family > languages. At least D (or others that can interlink with C++) would > need their own handling, e.g. gdc definition a matching version tag to > control the definitions in libdruntime/core/stdc/stdint.d. > > If this is worth the complexity or even GCC 16 material I cannot really > tell, though.
One might also be able to arrange for int8_t to be mangled the same as 'char' even when 'signed char' to avoid the ABI change (but retain some of the issues). I'd probably follow what the vendor does, thus Studio 12.6 CC, inter-operability with the "platform default" would be my priority, and see to have work arounds for the fallout that happens elsewhere ... Does Oracle acknowledge the issue and plan to fix in a similar manner? Richard. > > Rainer > > -- > ----------------------------------------------------------------------------- > Rainer Orth, Center for Biotechnology, Bielefeld University
