Stream: git-wasmtime

Topic: wasmtime / PR #8240 egraph: Resolve all aliases at once

view this post on Zulip Wasmtime GitHub notifications bot (Mar 26 2024 at 07:14):

jameysharp opened PR #8240 from jameysharp:resolve-egraph-aliases to bytecodealliance:main:

This way we can use the linear-time alias rewriting pass, and then avoid having to think about value aliases ever again.

view this post on Zulip Wasmtime GitHub notifications bot (Mar 26 2024 at 07:14):

jameysharp requested abrown for a review on PR #8240.

view this post on Zulip Wasmtime GitHub notifications bot (Mar 26 2024 at 07:14):

jameysharp requested wasmtime-compiler-reviewers for a review on PR #8240.

view this post on Zulip Wasmtime GitHub notifications bot (Mar 26 2024 at 20:04):

cfallin submitted PR review:

Nice!

view this post on Zulip Wasmtime GitHub notifications bot (Mar 28 2024 at 21:16):

jameysharp updated PR #8240.

view this post on Zulip Wasmtime GitHub notifications bot (Mar 28 2024 at 21:34):

jameysharp commented on PR #8240:

@cfallin, could you have a look at my additional changes in 70b082781c93a3a3359597e36a9f5ef5f8feeca7 and make sure they make sense?

CI informed me that values deleted due to this alias-rewrite pass were still referenced. While auditing everything in the DataFlowGraph for references to Values, I realized I think it makes sense to merge PCC facts onto the targets of value aliases. So that may affect your work on PCC, hopefully in good ways.

view this post on Zulip Wasmtime GitHub notifications bot (Mar 28 2024 at 22:01):

cfallin submitted PR review:

New changes look good!

view this post on Zulip Wasmtime GitHub notifications bot (Mar 28 2024 at 22:41):

jameysharp merged PR #8240.

Last updated: Nov 22 2024 at 16:03 UTC