On Mon Aug 12, 2024 at 1:05 PM CEST, Faidon Liambotis wrote: > On Sun, Aug 11, 2024 at 12:31:21PM -0400, Reinhard Tartler wrote: > > But you are right, ultimately this decision is up to the maintainer to make. > > Maintainer here :) > > For this particular case, I think removing -Werror made sense as a quick > fix to avoid kicking out the entire WasmEdge -> crun -> podman stack out
I wondered why wasmedge ended up on my radar, but that explains it ;-) > of testing, for what is a bug that existed before (i.e. not a > regression), and was just surfaced by a new compiler version. I had > looked into the bug before Reinhard NMUed and the fix wasn't obvious to > me. So taking a shortcut and prioritizing expediency over correctness > made sense to me, in this particular case. Yeah, makes a lot of sense. > But, at the same time, I agree that addressing this properly means > reporting this upstream so that the underlying bug gets fixed. I see > this has happened now, and upstream is already responsive (as they > usually are!). And what happened is why I like FLOSS so much: a bunch of people working together to solve a problem :-D As you may have seen I was able to verify that the proposed fix indeed solved the GCC-14 compiler issue, so I've DEP-3-ified that patch and attached it here. HTH, Diederik
From: hydai <hy...@secondstate.io> Date: Thu, 15 Aug 2024 13:59:51 +0800 Subject: [Loader] Fix GCC 14 maybe-uninitialized warning (#3659) Origin: upstream, https://github.com/WasmEdge/WasmEdge/commit/e427a71c5a50982c35e124340fc4febcd7600226 Fix #3640 Signed-off-by: hydai <hy...@secondstate.io> --- lib/loader/filemgr.cpp | 5 +++++ 1 file changed, 5 insertions(+) diff --git a/lib/loader/filemgr.cpp b/lib/loader/filemgr.cpp index 14afba28..a6e2b95e 100644 --- a/lib/loader/filemgr.cpp +++ b/lib/loader/filemgr.cpp @@ -49,6 +49,11 @@ Expect<void> FileMgr::setCode(Span<const Byte> CodeData) { // Set code data. See "include/loader/filemgr.h". Expect<void> FileMgr::setCode(std::vector<Byte> CodeData) { reset(); + // Tell GCC 14 that DataHolder has no data. + // Fix the false positive warning, + // which is reported by GCC 14 with `maybe-uninitialized` + assuming(!DataHolder); + DataHolder.emplace(std::move(CodeData)); Data = DataHolder->data(); Size = DataHolder->size(); -- 2.45.2
signature.asc
Description: PGP signature
