Stream: git-wasmtime

Topic: wasmtime / PR #10616 component::__internal::InstanceType ...


view this post on Zulip Wasmtime GitHub notifications bot (Apr 18 2025 at 18:51):

pchickey opened PR #10616 from bytecodealliance:pch/component_instancetyped_owned to bytecodealliance:main:

Having refs to a pair of arcs made it less expensive to create an InstanceType, but tied its to a borrow of the ComponentInstance (transitively the Store) which makes some new typechecking operations which have an AsContextMut impossible.

This has no direct benefit in this PR, but it split out of #10610 because it is more easily reviewed in isolation.

<!--
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 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 (Apr 18 2025 at 18:51):

pchickey requested alexcrichton for a review on PR #10616.

view this post on Zulip Wasmtime GitHub notifications bot (Apr 18 2025 at 18:51):

pchickey requested wasmtime-core-reviewers for a review on PR #10616.

view this post on Zulip Wasmtime GitHub notifications bot (Apr 18 2025 at 20:11):

pchickey updated PR #10616.

view this post on Zulip Wasmtime GitHub notifications bot (Apr 21 2025 at 14:51):

alexcrichton submitted PR review.

view this post on Zulip Wasmtime GitHub notifications bot (Apr 21 2025 at 14:51):

alexcrichton created PR review comment:

We talked briefly about this on Zulip, but this is the call that I'm worried about perf-wise. This is used during lifting/lowering of ResourceAny (transitively through the resource_type helper method which delegates to this one). I'd prefer to keep the Arc-cloning off the hot path there if possible, but I need to read up more on the other PR for better understanding the motivation behind this one.

view this post on Zulip Wasmtime GitHub notifications bot (Apr 21 2025 at 16:34):

pchickey created PR review comment:

Agreed, thanks for finding where it was in the hot path, I missed that. If for some reason the approach you outlined in #10610 doesn't cash out I will at very least make sure we can amortize the clone to only happen once per store, so it doesn't hurt resource lifting/lowering.

view this post on Zulip Wasmtime GitHub notifications bot (Apr 21 2025 at 16:34):

pchickey submitted PR review.

view this post on Zulip Wasmtime GitHub notifications bot (Apr 21 2025 at 21:03):

pchickey commented on PR #10616:

Replaced by #10621

view this post on Zulip Wasmtime GitHub notifications bot (Apr 21 2025 at 21:04):

pchickey closed without merge PR #10616.


Last updated: Dec 06 2025 at 06:05 UTC