iwanders opened PR #10130 from iwanders:issue-10118-improvements
to bytecodealliance:main
:
Improvements discussed in #10118, fyi @cfallin.
Changes are two-fold, commits are split out.
In 681c8c8fb5e286afb58b97f84cd23b0b0af70d54 adds a message to the debug assert stating the reason why the assert failed. Other asserts with string were capitalised, I tried to match it, reads like:
assertion `left == right` failed: Address width of 64 expected, got i32 left: types::I32 right: types::I64
Technically the type assignment and print isn't necessary, but I found it reads more clear than left and right only, though the fmt made it a multi-line :/
In c5f31de428e5f7bb0ea84a64a7e228c9ea9e7574 the size of the first and second argument is changed to 64 bits, as well as the
v3
andv4
, which hold the offset into the array, I also added one sentence above it pointing at the pointer size is assumed to be 64 bits for the conversion of the function to IR. I did the changes manually, confirmed the changes do not trigger the debug assert and verified the created symbol does work to calculate an average.Meta suggestion: The PR template contains:
Our development process is documented in the Wasmtime book:
https://docs.wasmtime.dev/contributing-development-process.htmlThat should really be linked in CONTRIBUTING.md as well, because I ended up looking at the commit history to figure out the commit message preference, only to discover that it'll be squashed anyway :D
iwanders requested wasmtime-compiler-reviewers for a review on PR #10130.
iwanders requested fitzgen for a review on PR #10130.
cfallin submitted PR review:
Thanks!
cfallin has enabled auto merge for PR #10130.
cfallin merged PR #10130.
Last updated: Feb 28 2025 at 03:10 UTC