perf: widen simple-absolute http(s) parse fast path

[anonrig]

Summary

Widen the existing try_parse_simple_absolute hot path for absolute http/https (no base URL):

• Bulk host / rest validation (NEON on Apple Silicon where available)
• memchr locate for ? / #
• Fail-closed host checks (is_ipv4, punycode) and path . / .. / %2e rejection
• Cheap digit / [ peek so pure IPv4/IPv6 inputs never enter the fast path

On success, fill url components and the url_aggregator buffer directly (no freelist, no href cache).

Scope

• src/parser.cpp, tests/basic_tests.cpp, small comment in src/implementation.cpp
• No string_pool, no url href cache, no public API signature changes

Independence

Targets main only. Independent of the get-href and href-cache PRs (not stacked).

Test plan

• basic_tests simple-absolute cases
• Full CI
• Optional: Release benchdata filter for aggregator/url href

[codecov]

Codecov Report

❌ Patch coverage is 83.05085% with 20 lines in your changes missing coverage. Please review.
✅ Project coverage is 61.06%. Comparing base (d154358) to head (68e3d77).

Files with missing lines	Patch %	Lines
src/parser.cpp	83.05%	5 Missing and 15 partials ⚠️

Additional details and impacted files

@@           Coverage Diff           @@
##             main    #1199   +/-   ##
=======================================
  Coverage   61.06%   61.06%           
=======================================
  Files          38       38           
  Lines        6939     6943    +4     
  Branches     3231     3239    +8     
=======================================
+ Hits         4237     4240    +3     
- Misses        749      752    +3     
+ Partials     1953     1951    -2     

☔ View full report in Codecov by Harness.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[codspeed-hq]

Merging this PR will improve performance by 10.01%

⚡ 10 improved benchmarks
✅ 19 untouched benchmarks
⏩ 4 skipped benchmarks¹

Performance Changes

	Benchmark	BASE	HEAD	Efficiency
⚡	Bench_DNS_AdaURL	83.7 ms	70.7 ms	+18.34%
⚡	url_search_params_AdaURL	126.4 µs	109 µs	+15.93%
⚡	BenchData_BasicBench_AdaURL_aggregator_href	66.8 ms	57.9 ms	+15.41%
⚡	BenchData_BasicBench_AdaURL_href	98.1 ms	85.1 ms	+15.27%
⚡	Bench_DNS_Aggregator	67.6 ms	58.7 ms	+15.18%
⚡	Bench_IPv6_AdaURL	3.3 ms	3.1 ms	+5.31%
⚡	Bench_BasicBench_AdaURL_href	27 µs	25.8 µs	+4.7%
⚡	Bench_IPv6_Aggregator	4.1 ms	3.9 ms	+4.54%
⚡	BBC_BasicBench_AdaURL_href	18.3 µs	17.6 µs	+3.93%
⚡	Bench_BasicBench_AdaURL_aggregator_href	22.1 µs	21.4 µs	+3.08%

Tip

Curious why this is faster? Comment @codspeedbot explain why this is faster on this PR, or directly use the CodSpeed MCP with your agent.

---

_{Comparing perf/simple-absolute-parse (68e3d77) with main (d154358)}

Footnotes

1. 4 benchmarks were skipped, so the baseline results were used instead. If they were deleted from the codebase, click here and archive them to remove them from the performance reports. ↩

[anonrig]

clangcl (Debug) failed with gtest discovery timeout (5s) on basic_tests / wpt_url_tests after link — empty output, no assert. Release clangcl and MSVC Debug passed.

This looks like Windows Debug discovery flakiness under load rather than a hang in try_parse_simple_absolute (no global init; NEON is offline on that job). Re-run if it persists.

CodSpeed for this PR is green with improvements and no IPv4 aggregator regressions.

[Copilot]

Pull request overview

This PR widens Ada’s existing try_parse_simple_absolute fast path for absolute http/https URLs (no base URL) to validate and materialize components more efficiently, aiming to reduce overhead in common “already-normalized” inputs while fail-closing to the full state machine for edge cases (credentials, IP literals, punycode, and dot-segment normalization triggers).

Changes:

• Adds bulk host scanning and bulk “rest” validation (NEON where available), plus memchr-based ?/# location to reduce per-byte branching in the hot path.
• Tightens fast-path eligibility and fallthrough conditions (credentials/port rejection, IPv4/punycode checks, dot-segment / %2e path normalization detection).
• Extends test coverage for fast-path fallthrough and correctness around credentials and IPv4-like host edge cases.

Reviewed changes

Copilot reviewed 3 out of 3 changed files in this pull request and generated 1 comment.

File	Description
tests/basic_tests.cpp	Adds regressions/coverage for credentialed URLs and additional simple-absolute fast-path edge cases.
src/parser.cpp	Implements widened simple-absolute http(s) fast path with bulk scanning/validation and direct component/buffer filling.
src/implementation.cpp	Documents why try_parse should not be redundantly called from parse() due to performance regressions on non-simple inputs.

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[Copilot]

Pull request overview

Copilot reviewed 4 out of 4 changed files in this pull request and generated 1 comment.

Suppressed comments (1)

src/parser.cpp:53

• k_rest is now unused in this file (the fast path uses rest_is_clean/k_rest_ok instead). Leaving an unused constexpr table at namespace scope can trigger -Wunused-const-variable warnings in some builds; either remove it or mark it [[maybe_unused]].

// 0 = ok, 1 = ?/#, 2 = reject. '%' is allowed (already-encoded); "%2e" is
// checked later via path_needs_norm so serialization stays exact.
constexpr std::array<uint8_t, 256> k_rest = []() consteval {
  std::array<uint8_t, 256> t{};