Bug#1133663: wasmedge: ftbfs with GCC-16

Hi Faidon!

I'm a newbie contributor to Debian and was looking into fixing this bug as a
useful learning experience. I stumbled across this as crun is removed from
testing because it transitively depends on libwasmedge0, and I use crun via
podman.

It seems the update to 0.17.1 broke the libwasmedge0 ABI:
https://dpaste.com/FGTKVMHSR

After looking into WasmEdge's commit history and general development, it seems
the project is now heavily LLM-coded, and this ABI breakage is perhaps a signal
that there could be better and more stable alternatives.

The only wasmedge reverse dep is crun (Suggests:), and crun supports three
other WASM runtimes: wasmer, wamr, wasmtime. Maybe one of these could be better
positioned as a potential replacement for WasmEdge?

Popcon for libwasmedge0 is 0, so maybe this package could be retired even?

The way I see it, there are a few potential solutions to close this bug:

- Bump wasmedge to 0.16.4 as a minimal-change fix for this FTBFS
- Bump to 0.17.1 as it's the latest upstream
- Backport the fix to 0.16.1
- Disable WASM support in crun to allow it back into test
- Package an alternative runtime and have crun depend on that instead
  - Then perhaps retire wasmedge?

I'm not super experienced so maybe I'm talking rubbish here and am way off the
mark, but I'm happy to help out if one of my potential solutions sounds good to
you (or some other solution entirely)?

Let me know. :)

On Wed, Apr 15, 2026 at 10:44:54AM +0300, Faidon Liambotis wrote:
> Control: forwarded -1 https://github.com/WasmEdge/WasmEdge/issues/4789
> Control: tags -1 + fixed-upstream
>
> On Mon, Apr 13, 2026 at 03:12:35PM +0200, Emanuele Rocca wrote:
> > Package: src:wasmedge
> > Version: 0.16.1+dfsg-1
>
> > wasmedge fails to build in a test rebuild on at least amd64 and arm64 with
> > gcc-16/g++-16, but builds properly with gcc-15/g++-15. The severity of this
> > report will be raised before the forky release.
>
> Thanks for the mass rebuild and for the report!
>
> Apparently this has been fixed upstream for some time, and even reported
> again in the upstream backtracker yesterday by someone else. I've
> confirmed that the fix is part of both the 0.16.2 branch (no tag yet),
> and the 0.17.x alphas. Given GCC 16 is not the default yet, and this is
> not RC, I'm opting to wait until for a WasmEdge release to avoid the
> hassle of backporting it.
>
> Faidon
>