rust: support SCCACHE_BASEDIRS across checkout roots

[pamarcos]

First, I want to point that I've used GPT-5.6 Sol to assist me with the work. However, I've reviewed the code and tested it.

Rust compilation currently includes checkout-specific paths in its cache key, so the existing SCCACHE_BASEDIRS support does not provide cross-checkout cache hits for Rust crates.

This adds Rust support while preserving location sensitivity for path inputs that are not safely covered by compiler remapping. In particular, normalization requires an effective --remap-path-prefix covering the working directory with the default or all remap scope. Without that compiler-side remapping, the working directory and other path-sensitive inputs remain exact.

This work is based on #2678 (Matt remains the author of the imported baseline commit). I also found that there's a similar PR at johnkferguson#1. Not sure why a PR upstream was never created.

So, I decided to create this PR because the other ones seem not to be followed up, but I'm happy to close this as a duplicate if the others are planned to be merged.

Implementation details

• normalizes only the source side of parsed --remap-path-prefix values, preserving the destination,
  mapping precedence, component matching, and final remap scope;
• hashes the effective remapped identity of every source file and preserves original identity when
  multiple sources map to the same virtual path;
• keeps env!, option_env!, profiles, plugins, and unremapped paths location-sensitive;
• normalizes a small allowlist of path-valued CARGO_* variables, including
  CARGO_INSTALL_ROOT, only when the working directory is remapped, there are no explicit externs,
  and crate search paths contain no dynamic libraries;
• frames normalized values and advances the Rust cache-key version from 6 to 7;
• falls back to local compilation when a distributed path transformer cannot safely preserve a user
  remap; and
• declines SCCACHE_BASEDIRS prefix stripping for non-ASCII Windows values.

This deliberately avoids scanning arbitrary argument, linker, environment, or replacement-path bytes. Those values may affect compiler output without being rewritten by rustc, so stripping them from the key could return an artifact containing paths or semantics from another checkout.

Full provenance and related discussions

Direct lineage

Reference	Relevance
#35	Original request for a CCACHE_BASEDIR equivalent.
#104	Early unmerged basedir implementation.
#2521	Mikhail Shiryaev's merged SCCACHE_BASEDIRS foundation for C/C++.
#2270	Made Rust --remap-path-prefix compilations cacheable.
#2651	Stopped separately hashing CARGO_ENCODED_RUSTFLAGS, which often carries the remap.
#2652	Weihang Lo's direct request for Rust support and warning that key-only normalization is unsafe.
vercel#10	Matt Mastracci's original Rust prototype.
#2678	Matt Mastracci's direct upstream implementation and the baseline for this PR.

Independent implementations and experiments

Reference	Relevance
johnkferguson#1	Glob/worktree experiment, per-token argument diagnosis, CARGO_INSTALL_ROOT, and reported 98.6% hit rate.
manoelcalixto#1	Independent opt-in linked-worktree implementation with extensive safety and CI coverage.
zackees/zccache#229	Related request for safe Rust cache sharing across source roots.
zackees/zccache#237	Root-covering remap gate and path-sensitive Rust integration tests in zccache.
https://github.com/moriyoshi/sccache-wrapper	Standalone normalized-artifact cache wrapper for cross-worktree reuse.
NVIDIA/OpenShell#2379	Operational evidence of zero Rust cross-worktree hits before normalization.
NVIDIA/OpenShell#2475	Follow-up experiment using SCCACHE_BASEDIRS.

Cache-key and artifact-safety context

Reference	Relevance
#196	Historical report of shared Rust dependencies missing across projects.
#206	Earlier path-independent Rust key attempt, left unmerged because artifacts differed by path.
#345	Added Rust working-directory hashing because the path is embedded in rlibs.
#986	Discussion of whether and why the working directory belongs in cache keys.
#2494	Discussion of which CARGO_* variables should affect Rust keys.
#2495	Alternative proposal to stop hashing redundant or irrelevant CARGO_* variables.
#2673	Analogous parsed source-side normalization for C/C++ prefix-map arguments.
#2711	Related C/C++ argument normalization and per-argument key framing.
#2763	Demonstrates the debugger/source-path consequence of serving an artifact from another checkout.
#2765	Daniel Colascione's direct request to normalize congruent Rust --remap-path-prefix values.
#2595	Worktree workflow and dynamic basedir discovery requirements.
#2737	Windows basedir matching bug report.
#2736	Windows escaped-separator fix in shared basedir handling.
rust-lang/cargo#10915	Upstream evidence that moving CARGO_HOME changes path-sensitive artifacts/fingerprints.
rust-lang/cargo#12137	Cargo trim-paths and remap-scope tracking issue.

Benchmarks building ClickHouse

Running on a AMD Ryzen 9 9950X with 96G

Scenario	Average build time	Sample SD	C/C++ hits	C/C++ misses	Rust hits	Rust misses
ccache	142.73s	2.03s	16,614	2	Not wrapped	Not wrapped
sccache with Rust	54.91s	0.63s	16,730	2	822	0
sccache with Rust, client-side	154.72s	0.27s	16,730	2	808	0
sccache C/C++ only	144.03s	0.78s	16,614	2	Not wrapped	Not wrapped

The interesting part here is that using the new client-side architecture, it slows down a lot. I'll probably create a separate issue to point that.

Closes: #2652
Closes: #2765
Related: #2678
Related: #2595

[sylvestre]

Long comment #0 written by a llm isn't useful... Please make it shorter

[codecov-commenter]

Codecov Report

❌ Patch coverage is 98.51412% with 10 lines in your changes missing coverage. Please review.
✅ Project coverage is 74.23%. Comparing base (c037e11) to head (928dbf6).
⚠️ Report is 2 commits behind head on main.

Files with missing lines	Patch %	Lines
src/compiler/rust.rs	98.39%	10 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #2794      +/-   ##
==========================================
- Coverage   74.32%   74.23%   -0.10%     
==========================================
  Files          71       72       +1     
  Lines       40789    40555     -234     
==========================================
- Hits        30318    30107     -211     
+ Misses      10471    10448      -23     

☔ 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.

[pamarcos]

Long comment #0 written by a llm isn't useful... Please make it shorter

Roger. I cut most of it. Left the implementation details available in case someone's interested, though.