feat/parse url

[nahime0]

Summary

Adds PHP-compatible parse_url() on both the AOT compiler path and the Magician eval() runtime, plus the predefined PHP_URL_* constants.

• AOT: registry builtin (src/builtins/string/parse_url.rs) → RuntimeFnId::ParseUrl → Mixed-returning __rt_parse_url scanner (AArch64 + x86_64) with associative-array and component-selector results, Fresh ownership, and catchable ValueError for invalid selectors.
• Magician: eval_builtin! home file with a pure-Rust scanner aligned to PHP 8.4 ext/standard/url.c quirks (scheme-relative URLs, IPv6 brackets, empty present components, control-byte substitution, port range checks).
• Constants: PHP_URL_SCHEME … PHP_URL_FRAGMENT (0–7) wired through checker, prescan, name resolver, autoload interpreter, Magician predefined constants, README, and docs.
• Parity: shared fixture corpus in tests/fixtures/parse_url_cases.json exercised by Magician unit tests and AOT codegen tests; coverage also includes component shapes, named/callable call sites, invalid URLs (false), selector ValueError, and heap-debug ownership.

Behavior notes

Call	Success	Missing component	Invalid URL
parse_url($url) / negative component	assoc array of present keys (port as int)	key omitted	false
PHP_URL_PORT	int	null	false
other PHP_URL_*	string (may be "")	null	false

Selectors greater than PHP_URL_FRAGMENT raise:

parse_url(): Argument #2 ($component) must be a valid URL component identifier, N given

Any negative component selects the full-array form (PHP 8.4). Invalid URLs return false even when the selector would otherwise be invalid.

Test plan

• cargo build
• cargo test --test codegen_tests parse_url
• cargo test --test error_tests parse_url
• cargo test -p elephc-magician parse_url
• git diff --check
• Optional: ELEPHC_PHP_CHECK=1 cargo test --test codegen_tests parse_url
• CI matrix: macos-aarch64, linux-aarch64, linux-x86_64

Docs / examples

• docs/php/strings.md, generated builtin pages, internals runtime note
• examples/parse-url/main.php
• CHANGELOG.md [Unreleased] entry

[nahime0]

@greptile-apps review this

[greptile-apps]

Greptile Summary

Adds PHP-compatible parse_url() support to native compilation and the Magician eval runtime.

• Registers the builtin, predefined PHP_URL_* constants, checker return types, and runtime lowering.
• Implements matching Rust and AArch64/x86_64 scanners with component selection, associative-array results, and catchable selector errors.
• Adds shared compatibility fixtures, runtime ownership coverage, documentation, and an example.

Confidence Score: 5/5

The PR appears safe to merge based on the inspected parsing, ABI, ownership, exception, and cross-target behavior.

The native and eval implementations consistently materialize PHP-compatible result shapes, preserve the expected runtime layouts, and are exercised by a shared edge-case corpus; no actionable changed-code failure remains.

Important Files Changed

Filename	Overview
crates/elephc-magician/src/interpreter/builtins/string/parse_url.rs	Adds the eval builtin, PHP-compatible byte scanner, result materialization, and fixture-driven parity tests; no actionable defect was established.
src/builtins/string/parse_url.rs	Registers the AOT builtin and refines static selector result types while retaining Mixed for dynamic selectors.
src/codegen/lower_inst/builtins/strings.rs	Lowers URL and selector operands into the architecture-specific runtime ABI without an established register-preservation defect.
src/codegen_support/runtime/strings/parse_url.rs	Emits scanner support, static component keys, and catchable ValueError construction for both targets.
src/codegen_support/runtime/strings/parse_url/aarch64.s	Implements AArch64 parsing, component copying, Mixed boxing, and associative-hash construction with consistent ABI usage.
src/codegen_support/runtime/strings/parse_url/x86_64.s	Implements the corresponding x86_64 scanner and result construction with no concrete architecture divergence found.
tests/fixtures/parse_url_cases.json	Provides shared PHP-derived coverage for standard forms, invalid authorities and ports, empty components, IPv6, and control-byte substitution.
Cargo.lock	Adds the existing serde_json package to the Magician dependency list without changing the advisory-flagged dependency versions.

Flowchart

%%{init: {'theme': 'neutral'}}%%
flowchart TD
  A["PHP parse_url(url, component)"] --> B{"Execution path"}
  B -->|AOT| C["EIR RuntimeFnId::ParseUrl"]
  C --> D["Target-aware lowering"]
  D --> E["AArch64 / x86_64 scanner"]
  B -->|eval| F["Magician builtin registry"]
  F --> G["Rust URL scanner"]
  E --> H{"Parse result"}
  G --> H
  H -->|Invalid URL| I["false"]
  H -->|Negative selector| J["Associative component array"]
  H -->|Selector 0–7| K["string / int / null"]
  H -->|Selector > 7| L["ValueError"]

Loading

_{Reviews (1): Last reviewed commit: "docs: add parse_url implementation plan" | Re-trigger Greptile}

[greptile-apps]

Too many files changed for review (199 files, 100 file limit).

Bypass the limit by tagging @greptile-apps to review.