Stream: git-wasmtime

Topic: wasmtime / PR #14785 Add `--preload` support for components


view this post on Zulip Wasmtime GitHub notifications bot (Oct 16 2026 at 17:57):

fitzgen opened PR #14785 from fitzgen:component-pre-load to bytecodealliance:main:

This required the ability to wrap a live instance into a wasmtime::component::Linker as well.

<!--
Please make sure you include the following information:

Our development process is documented in the Wasmtime book:
https://docs.wasmtime.dev/contributing-development-process.html

Please review the Bytecode Alliance's AI tool usage policy at
https://github.com/bytecodealliance/governance/blob/main/AI_TOOL_POLICY.md

Please ensure all communication follows the code of conduct:
https://github.com/bytecodealliance/wasmtime/blob/main/CODE_OF_CONDUCT.md
-->

view this post on Zulip Wasmtime GitHub notifications bot (Oct 16 2026 at 17:57):

fitzgen requested dicej for a review on PR #14785.

view this post on Zulip Wasmtime GitHub notifications bot (Oct 16 2026 at 17:57):

fitzgen requested wasmtime-core-reviewers for a review on PR #14785.

view this post on Zulip Wasmtime GitHub notifications bot (Oct 16 2026 at 20:47):

github-actions[bot] added the label wasmtime:api on PR #14785.

view this post on Zulip Wasmtime GitHub notifications bot (Oct 16 2026 at 20:47):

fitzgen updated PR #14785.

view this post on Zulip Wasmtime GitHub notifications bot (Oct 16 2026 at 22:31):

fitzgen updated PR #14785.

view this post on Zulip Wasmtime GitHub notifications bot (Oct 19 2026 at 15:02):

:memo: alexcrichton submitted PR review:

Do you have an immediate use case for this at the moment, and if so could it use wasm-tools compose or wac in the meantime?

I'm hesitant myself to start committing to something here becuase there's a lot that's possible in this space and I'm not sure that this is the best approach. Some thoughts I'd have are:

Basically I'm not confident that we should take --preload as-is and support it within the CLI, and I'm also not confident in the implementation using Val. That's what leads me to ask if a composition works for your use case if you have one. Otherwise though I think it's fine to work through these questions/ideas, but I'm not sure if that's what you were expecting to do here.

view this post on Zulip Wasmtime GitHub notifications bot (Oct 19 2026 at 15:25):

:thumbs_up: dicej submitted PR review:

LGTM, thanks!

It occurred to me that we could precalculate the parameters which have resource borrows to drop based on the types rather than visit them all at runtime, but I suppose we're not trying to optimize this given that composition remains the fast path.

view this post on Zulip Wasmtime GitHub notifications bot (Oct 19 2026 at 15:27):

dicej commented on PR #14785:

Oops, my review crossed with Alex's; he makes some good points about e.g. using WAC just-in-time.

view this post on Zulip Wasmtime GitHub notifications bot (Oct 19 2026 at 16:07):

fitzgen commented on PR #14785:

My immediate goal is to be able to easily run+profile sightglass benchmarks without needing to stub out the bench.start/bench.end imports. I could use wac or whatever, but that is a bit more annoying and less it-just-works-out-of-the-box compared to core wasm, but I guess that is fine as a temporary workaround.

view this post on Zulip Wasmtime GitHub notifications bot (Oct 19 2026 at 16:07):

:cross_mark: fitzgen closed without merge PR #14785.


Last updated: Oct 31 2026 at 23:03 UTC