perf: capture parse Position without boxing the offset Int#876
Merged
stephenamar-db merged 1 commit intoJun 2, 2026
Merged
Conversation
This was referenced May 30, 2026
Motivation: Parser.Pos is invoked for nearly every AST node. It was `Index.map(off => new Position(...))`: fastparse's `Index` stores the offset as an Int in its `successValue: Any` field (boxing it), and the `.map` then unboxes it and allocates a closure — per node. boxToInteger via SharedPackageDefs.Index was a top self-frame in the parse flamegraph on kube-prometheus. Modification: - Rewrite Pos to write the Position object straight into successValue via ctx.freshSuccess(new Position(fileScope, ctx.index)), skipping the Int box/unbox and the map closure. Parse output (positions/errors) is unchanged. Result: JMH ParserBenchmark (parse-only, all test-suite files): 1.669 -> 1.579 ms/op (+5.4%, non-overlapping bands). Native parse_time on kube-prometheus: ~105.6 -> ~100.9 ms (+4.5%, consistent). Output byte-identical. 450/450 tests pass.
He-Pin
force-pushed
the
perf/parser-pos-no-boxing
branch
from
b8a525f to
c231f70
Compare
stephenamar-db
approved these changes
Jun 2, 2026
stephenamar-db
pushed a commit
that referenced
this pull request
Jun 2, 2026
## Motivation `std.manifestJson`, `std.manifestJsonMinified`, and `std.manifestJsonEx` routed through `java.io.StringWriter`, paying `StringBuffer` synchronization per `write`/`flush` on the hot manifestation path. Source-built jrsonnet comparisons showed sjsonnet trailing on object-heavy manifest workloads. ## Modification - Add `StringBuilderWriter`: an unsynchronized `Writer` over a `StringBuilder`. - Add package-private `FastMaterializeJsonRenderer` backed by `StringBuilderWriter`; route the three `std.manifestJson*` builtins through it. Public `MaterializeJsonRenderer` ABI/shape unchanged. - Use an in-place codepoint sort for `sortedVisibleKeyNames` / `maybeSortKeys` (avoids `.sorted` boxing). - Fix codepoint comparison for raw surrogate prefixes; `UnicodeHandlingTests` extended. ## Result Scala Native hyperfine on kube-prometheus, jrsonnet HEAD `2d7eed05`: | Workload (native) | Before | After | Δ | |---|---:|---:|---:| | kube-prometheus, sjsonnet | 158.4 ± 16.8 ms | 143.7 ± 3.2 ms | **−9.3%** | | `manifestJsonEx`, sjsonnet | — | 5.09 ± 1.01 ms | new | ## Test plan - [x] `./mill __.reformat` - [x] `./mill 'sjsonnet.jvm[3.3.7]'.test` — 518/518 pass This PR is the base for the stacked follow-ups #875 (TomlRenderer reuses `StringBuilderWriter`), and the independent #876/#877/#878.
stephenamar-db
pushed a commit
that referenced
this pull request
Jun 22, 2026
…ontexts (#1011) ## Motivation go-jsonnet issue #873 reports that `(function(x, x, x) x)(null, 3, 2)` silently returns `2` instead of rejecting duplicate parameter names. sjsonnet already rejected duplicates in function expressions, but the error message was misleading in local bindings and object methods (e.g. "Expected \")\"" or "Expected \":\""). ## Modification The `params` parser correctly detects duplicates via `flatMapX` + `Fail`. However, in `bind` (`local f(x,x)=...`) and `field` (`{f(x,x):...}`), the optional group `.?` around `params` swallowed the "no duplicate parameter" error and backtracked, producing misleading errors. Two changes fix this: 1. In `field()`: added `~/` (cut) after `"("` to commit to the params path once `"("` is matched: `"(" ~/ params(...) ~ ")"`. 2. In `params()`: set `ctx.cut = true` programmatically when a duplicate is detected, preventing outer `.?` operators from swallowing the error. Also captures `Pos` for each parameter to position the error at the duplicate parameter name (not at the closing paren). ## Result All duplicate parameter contexts now produce the correct error `"Expected no duplicate parameter: x"` with accurate position info: - `(function(x, x, x) x)` -> ParseError (already worked) - `local f(x, x) = x` -> ParseError (was `"Expected )"`) - `{ f(x, x): x }` -> ParseError (was `"Expected :"`) - `{ local f(x, x) = x }` -> ParseError (was `"Expected )"`) - Nested bindings also fixed as a bonus. Tests pass on Scala 3.3.7, 2.13.18, 2.12.21. ## References - google/go-jsonnet#873 ## Cross-implementation comparison: Duplicate function parameters | Implementation | Detection Phase | All Contexts Covered | Error Message Example | |---|---|---|---| | **sjsonnet (this fix)** | Parse time | ✅ All 4 contexts | `Expected no duplicate parameter: x` | | **cpp-jsonnet** | Static analysis (post-parse) | ✅ All 4 (via desugaring) | `Duplicate function parameter: x` | | **go-jsonnet** (master) | ❌ None | ❌ Bug — silently accepts | N/A — last argument wins silently | | **go-jsonnet** ([PR #876](google/go-jsonnet#876), unmerged) | Parse time | ✅ All 4 | `Duplicate function parameter: a` / `Duplicate method parameter: x` | | **jrsonnet** | Static analysis (post-parse) | ✅ All 4 (via desugaring) | `local is already defined in the current frame: x` | ### Behavior by test case | Test Case | sjsonnet (fix) | cpp-jsonnet | go-jsonnet | jrsonnet | |---|---|---|---|---| | `(function(x, x, x) x)(null, 3, 2)` | ✅ ParseError | ✅ StaticError | ❌ Returns `2` | ✅ StaticError | | `local f(x, x) = x; f(1, 2)` | ✅ ParseError | ✅ StaticError | ❌ Accepted | ✅ StaticError | | `{ f(x, x): x }.f(1, 2)` | ✅ ParseError | ✅ StaticError | ❌ Accepted | ✅ StaticError | | `{ local f(x, x) = x, a: f(1, 2) }` | ✅ ParseError | ✅ StaticError | ❌ Accepted | ✅ StaticError | ### Key differences - **Detection timing**: sjsonnet and go-jsonnet (PR #876) detect at parse time; cpp-jsonnet and jrsonnet detect during a separate static analysis pass after parsing. - **Error message specificity**: go-jsonnet PR #876 distinguishes "function parameter" vs "method parameter"; cpp-jsonnet uses "function parameter" uniformly (methods desugar to functions before the check). - **Position accuracy**: sjsonnet points to the duplicate parameter name; cpp-jsonnet points to the entire function AST node. - **go-jsonnet master** has a known bug ([issue #873](google/go-jsonnet#873)) where duplicate parameters are silently accepted with "last argument wins" semantics.
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 join this conversation on GitHub.
Already have an account?
Sign in to comment
2 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.
Motivation
Parser.Posis invoked for nearly every AST node to capture the source offset. It wasIndex.map(off => new Position(...)): fastparse'sIndexstores the offset as anIntin itssuccessValue: Anyfield (boxing it), and the.mapthen unboxes it and allocates a closure — per node.boxToIntegerviaSharedPackageDefs.Indexwas a top self-frame in the parse flamegraph on kube-prometheus.Modification
Positionobject straight intosuccessValueviactx.freshSuccess(new Position(fileScope, ctx.index)), skipping theIntbox/unbox and themapclosure. Parse output (positions / error offsets) is unchanged.Result
Scala Native
--debug-stats parse_timeon kube-prometheus (148 files), interleaved and cooled:→ ~6–8% faster parse; output byte-identical. (JMH
ParserBenchmarkover the test suite showed +5.4% on JVM.)Test plan
./mill __.reformat./mill 'sjsonnet.jvm[3.3.7]'.test— 518/518 pass