reset offset and refactor bench

[kazeburo]

PR Type

Enhancement, Tests

---

Description

• Reset scanner offsets before every scan

• Reuse iteration stop error sentinel

• Add scanner reuse regression coverage

• Refactor benchmarks with realistic log data

---

Diagram Walkthrough

flowchart LR
  Scan["Scanner.Scan"] --> Reset["Reset offset"]
  Reset --> Process["Process input lines"]
  Iter["Scanner.Iter"] --> Sentinel["Shared stop sentinel"]
  Bench["Benchmarks"] --> Data["Generated log data"]
  Data --> Bench

Loading

File Walkthrough

	Relevant files
Bug fix	linebuf.go Reset offsets and reuse iteration sentinel                              linebuf.go Adds a shared iteration-stop error sentinel. Resets scanner offsets before each scan. Preserves iteration error reporting behavior. +6/-3
Tests	linebuf_test.go Expand scanner tests and benchmark realism                              linebuf_test.go Adds scanner reuse offset regression coverage. Replaces synthetic numeric benchmark data. Generates realistic structured log benchmark lines. Reuses scanner configuration across benchmark iterations. +61/-39
Configuration changes	Makefile Increase benchmark execution duration                                        Makefile Extends benchmark duration from two to five seconds. +1/-1

---

[sonarqubecloud]

Quality Gate passed

Issues
0 New issues
0 Accepted issues

Measures
0 Security Hotspots
0.0% Coverage on New Code
0.0% Duplication on New Code

See analysis details on SonarQube Cloud

[github-actions]

Benchmark Comparison

Base: main
PR: feature/offset-reset-and-refactor-bench

BenchmarkScan results

goos: linux
goarch: amd64
pkg: github.com/monitoring-forge/linebuf
cpu: AMD EPYC 9V74 80-Core Processor                
               │ base-bench.txt │             pr-bench.txt              │
               │     sec/op     │    sec/op     vs base                 │
Scan_bufio-4        117.9µ ± 5%    544.2µ ± 4%  +361.40% (p=0.000 n=10)
Scan_linebuf-4      94.49µ ± 1%   520.39µ ± 1%  +450.74% (p=0.000 n=10)
Scan_iter-4         104.1µ ± 6%    524.3µ ± 1%  +403.78% (p=0.000 n=10)
geomean             105.1µ         529.5µ       +403.99%

               │ base-bench.txt │               pr-bench.txt                │
               │      B/op      │     B/op      vs base                     │
Scan_bufio-4       4.000Ki ± 0%   0.000Ki ± 0%  -100.00% (p=0.000 n=10)
Scan_linebuf-4     4.062Ki ± 0%   0.000Ki ± 0%  -100.00% (p=0.000 n=10)
Scan_iter-4        4.078Ki ± 0%   0.000Ki ± 0%  -100.00% (p=0.000 n=10)
geomean            4.047Ki                      ?                       ¹ ²
¹ summaries must be >0 to compute geomean
² ratios must be >0 to compute geomean

               │ base-bench.txt │              pr-bench.txt               │
               │   allocs/op    │ allocs/op   vs base                     │
Scan_bufio-4         1.000 ± 0%   0.000 ± 0%  -100.00% (p=0.000 n=10)
Scan_linebuf-4       2.000 ± 0%   0.000 ± 0%  -100.00% (p=0.000 n=10)
Scan_iter-4          3.000 ± 0%   0.000 ± 0%  -100.00% (p=0.000 n=10)
geomean              1.817                    ?                       ¹ ²
¹ summaries must be >0 to compute geomean
² ratios must be >0 to compute geomean

[github-actions]

PR Reviewer Guide 🔍

Here are some key observations to aid the review process:

⏱️ Estimated effort to review: 2 🔵🔵⚪⚪⚪

🧪 PR contains tests

🔒 No security concerns identified

⚡ No major issues detected

[github-actions]

PR Code Suggestions ✨

No code suggestions found for the PR.