[pull] master from ruby:master - #1439
Merged
Merged
Conversation
The JIT optimizer sees an `invokebuiltin` and would like to assume that calling it does not write back to the frame. A `Primitive` should never write to the frame by taking a pointer to a local, so this PR removes that functionality. Partial revert of bb89242. ---- * Don't allow writing to locals in cexpr This makes the few remaining cases of `cexpr`-that-looks-at-locals a read-only view. * Remove uses of LOCAL_PTR from ractor.rb Use two new builtin functions instead. * Remove LOCAL_PTR from mk_builtin_loader * Don't bother making the intermediate pointer * Revert bb89242 * Add Primitive.local_self!
Guard the block arg of `foo(&blk)` when profiling shows it is exactly a Proc, and pass it directly as the callee frame's block handler instead of falling back to a dynamic send.
There were some difficulties after ruby/prism#4223 was merged because the excludes were out of sync.
Previously, `setlocal level=0` did not set the local in the frame state during HIR build, unlike `setlocal_WC_0`. `getlocal` and `getlocal_WC_0` had the same issue. I don't know if it's possible to trigger an ZJIT issue with default iseq compile options today, but with ZJIT patches I'm working on that touch locals, it is readily possible. Fix by centralizing the match arms for `getlocal`, `setlocal`, and all of their `*_WC_*` variants.
Closes Shopify#1061. Also add a test to make sure `String#intern` generates the same HIR. ### Benchmarks I had AI create 4 new microbenchmarks for this change and also run `protoboeuf-encode`. They look like this: ```ruby require_relative '../harness/loader' keys = Array.new(2_048) { |i| "zjit-benchmark-key-#{i}" } keys.each(&:to_sym) expected = keys.last.to_sym def zjit_bench_varied(keys) i = 0 result = nil while i < 2_048 result = keys[i].to_sym i += 1 end result end run_benchmark(1) do result = nil 1_000.times { result = zjit_bench_varied(keys) } raise 'incorrect symbol' unless result.equal?(expected) end ``` Let me know if you want those microbenchmarks in the repository and I'll open the PR. Also let me know if you want other macro benchmakrs included. ``` before: ruby 4.1.0dev (2026-09-23T16:47:54Z :detached: d74c656) +ZJIT dev +PRISM [arm64-darwin25] after: ruby 4.1.0dev (2026-09-23T18:10:52Z :detached: 6662568) +ZJIT dev +PRISM [arm64-darwin25] ------------------ ------------ ------------ ------------- ------------ bench before (ms) after (ms) after 1st itr before/after protoboeuf-encode 233.2 ± 2.1% 228.5 ± 0.9% 1.075 1.021 zjit-to-sym-dup 64.8 ± 1.7% 60.2 ± 2.0% 1.071 1.075 zjit-to-sym-new 57.3 ± 4.4% 55.3 ± 3.6% 0.936 1.038 zjit-to-sym-reused 165.5 ± 2.1% 148.9 ± 1.3% 1.245 1.112 zjit-to-sym-varied 196.2 ± 1.8% 179.1 ± 1.3% 1.108 1.096 ------------------ ------------ ------------ ------------- ------------ ```
(ruby/strscan#219) `#pos=` accepts an arbitrary byte offset, so the scan position can fall in the middle of a multi-byte character such as UTF-16BE. Matching from there made Onigmo read a whole character starting at that position, reading past the end of the string. On CRuby this is an out-of-bounds read. On JRuby a search such as `#scan_until` could loop without terminating. On TruffleRuby it aborts with an internal error while building the matched substring. But we don't change the TruffleRuby implementation because we don't have a fast implementation for this. Refuse to match unless the current position is the head of a character. ruby/strscan@a552ea6f49
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to subscribe to this conversation on GitHub.
Already have an account?
Sign in.
7 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
See Commits and Changes for more details.
Created by
pull[bot] (v2.0.0-alpha.4)
Can you help keep this open source service alive? 💖 Please sponsor : )