https://github.com/steven-johnson created
https://github.com/llvm/llvm-project/pull/73702
Apparently the gcc 7.5 compiler for arm32-linux fails to implicitly convert to
std::optional; this inserts an explicit constructor to unbreak that compiler.
[Injection from
https://reviews.llvm.org/rG414
https://github.com/steven-johnson closed
https://github.com/llvm/llvm-project/pull/73702
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
steven-johnson wrote:
FYI: this *appears* to have injected a failure into Halide builds (based on a
bisect) -- there are some bitcode files that we build with LLVM's top-of-tree
Clang then load later on to use for codegen, and these bitcode files are now
missing certain type definitions that a
steven-johnson wrote:
Looks like more-aggressive type stripping broke us. We have a workaround we can
land on the Halide side.
https://github.com/llvm/llvm-project/pull/68882
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.o
steven-johnson wrote:
Note that Halide has removed its dependency on the le32/le64 targets as of
https://github.com/halide/Halide/pull/8344.
https://github.com/llvm/llvm-project/pull/98497
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https:
steven-johnson wrote:
Unfortunately, this utterly breaks Halide, which was in fact relying on both of
these targets.
https://github.com/llvm/llvm-project/pull/98497
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin
steven-johnson wrote:
> LGTM! That partial Le32/Leb64 restore was to give Halide some time.
I understand that you want these gone, but the fact of the matter is that
Halide isn't ready for this change, and so the next integrate of LLVM into
google3 will break the world. We urgently need these
steven-johnson wrote:
> We're trying wasm now.
This looked promising in our GitHub tests, but for reasons that aren't clear
(yet), ~everything compiled this way inside google3 resulted in linker errors
(missing symbols).
https://github.com/llvm/llvm-project/pull/98497
steven-johnson wrote:
After some initial experiments, it looks unlikely that a quick fix is possible
on our side. May I suggest a revert with a (say) 30-day expiration date? That
should give us enough time to come up with a fix on our side (which will
actually be a drop-other-work-and-fix-it a
