On Mon, Dec 29, 2025, at 4:40 PM, Faidon Liambotis wrote:
> On Sat, Dec 27, 2025 at 07:40:33PM +0100, Fabian Grünbichler wrote:
>> Attached you can find a fix adapting the test Makefile for LLVM 21
>> (seems like similar adaptations were already needed in the past, but
>> not sent upstream?).
>> 
>> I can NMU this if desired, since it will start blocking rustc 1.91.1
>> soonish.
>
> Thanks for the fix as well as well as for offering to NMU, much
> appreciated!
>
> The git tree has gotten a little messy, requiring a three-way merge
> between your earlier NMU, your rust-llvm commit in git tip, and the most
> recent NMU by Matthias (that wasn't checked into git). So I'll do a
> maintainer upload shortly merging all of this together, plus the patch
> from this bug, plus a few unrelated changes. It's passing autopkgtests
> for both architectures in Debusine.

Great, thanks! :)

> A bigger upload incorporating the new 0.15 (or 0.16 at this point)
> upstream release is pending, but I'll do a 0.14 upload in the meantime
> to make sure we don't entangle the rustc migration with potentially more
> risky changes.
>
> Finally, yes, you are absolutely right that these adaptations need to be
> forwarded upstream -- both the sed in the test file that you patched, as
> well as examples-mandelbrot-add-multivalue-reference-types-to-was.patch.
> I'll have a look at that once I rebase to a newer upstream release, but
> feel free to have a look if you have any spare cycles :)
>
> On that note: you're more than welcome to comaintain WasmEdge if you
> would like to, or continue to NMU at your discretion.

I don't have a use case for it atm (my interest in WASM is on the browser
side coming from Rust, and most of that is not yet packaged/enabled in
Debian either ;)).

So I am fine with NMUing or providing patches when I notice breakage :)

Fabian

Reply via email to