[POC] {cwd} placeholder in --remap-path-prefix

[ashi009]

Warning

Proof-of-concept, do not merge. Validates feasibility + CI ahead of a Major Change Proposal, since it adds placeholder semantics to the stable --remap-path-prefix flag. Companion to #157348 (the minimal, unstable-only fix that can land now).

Idea

Let a --remap-path-prefix FROM contain a {cwd} placeholder:

--remap-path-prefix={cwd}=.
--remap-path-prefix={cwd}/lib/std=std

The FROM is stored verbatim in remap_path_prefix, and the placeholder is expanded to the working directory only when the FilePathMapping is built (apply time). So:

• The stored value is the placeholder text ({cwd}/sub), not the absolute cwd → its dependency hash is stable across build directories → the incremental cache survives (fixes the cwd half of no-op remap_path_prefix change invalidates incremental cache #132132).
• No new type and no change to the field, tracking, decoder, or rustdoc — storing the template verbatim keeps {cwd} (placeholder) and {{cwd}} (escaped literal) distinct in storage, so a plain PathBuf is sufficient.

-Zremap-cwd-prefix=V is reimplemented as exactly sugar for --remap-path-prefix={cwd}=V, giving a path to sunset that unstable flag.

Escaping

format!-style: a path component {cwd} is the placeholder; {{/}} are literal {/}, so a literal {cwd} directory is written {{cwd}}.

Why {cwd} (vs <cwd>, %cwd%, \(cwd))

• format! syntax — native to Rust, self-escaping ({{/}}), doesn't look path-like.
• <cwd> / \(cwd): shell redirection / backslash issues (\ is the Windows path separator).
• %cwd%: collides with cmd.exe %VAR% (%CD%).

(Hand-typed it's quoted like any tool-level token; build systems pass argv with no shell, so it arrives verbatim.)

Open questions for the MCP

• Placeholder syntax + escaping ({cwd}, {{/}} here).
• Unknown placeholder handling — kept literal here; erroring may be better.
• Whole-component vs substring placeholders (whole-component here).
• Interaction with --remap-path-scope; relationship to RFC 3127.
• Stable-surface change → requires team sign-off (MCP).

Tests

{cwd}/sub stored-verbatim + applied; escaped {{cwd}} stays literal; -Zremap-cwd-prefix → {cwd} equivalence.

r? compiler

[rustbot]

Thanks for the pull request, and welcome! The Rust team is excited to review your changes, and you should hear from @chenyukang (or someone else) some time within the next two weeks.

Please see the contribution instructions for more information. Namely, in order to ensure the minimum review times lag, PR authors and assigned reviewers should ensure that the review label (S-waiting-on-review and S-waiting-on-author) stays updated, invoking these commands when appropriate:

• @rustbot author: the review is finished, PR author should check the comments and take action accordingly
• @rustbot review: the author is ready for a review, this PR will be queued again in the reviewer's queue

Why was this reviewer chosen?

The reviewer was selected based on:

• compiler expanded to 73 candidates
• Random selection from 18 candidates