docs: stop publishing the internal plans directory on a public repo

[JustinKovacich]

Closes #158.

main on this public repository carried docs/simple_someip/plans/ — eight
design and plan documents written for ourselves, named by internal phase and
pull-request number. This removes them.

What the PR does

docs: drop the in-repo pointers into the internal plans directory — four
citations of paths under docs/simple_someip/plans/ (three source comments,
one CI comment) removed, so nothing is left pointing at files that no longer
exist. The surrounding prose already carried the reasoning; the citations only
told a public reader where our internal documents live. One of them — the
future-size budget comment in src/client/mod.rs — also said "compare against
the baseline", meaning the baseline file; it now points at the constants
immediately below it.

docs: stop publishing docs/simple_someip/plans/ on a public repo — the
eight files, git rm'd. docs/ held nothing else, so the directory is gone.

Two things worth knowing

They were also inside every published crate. [package] declares no
include/exclude (the exclude in the manifest is [workspace]'s, for
tools/size_probe), so cargo package picked the whole directory up.
Before: cargo package --list | grep -c '^docs/' → 8. After → 0.

Someone already decided this, and only half-landed it. .gitignore has
listed docs/simple_someip/plans/ since 6925d56 (2026-07-16, "gitignore
superpowers plan/spec docs; untrack the #142 pair"). That commit untracked the
newer pair and noted that already-tracked plans stay tracked. These eight
landed 2026-06-10 and earlier, so the rule never reached them. With them gone,
the existing rule keeps them gone — no new ignore entry needed.

What this is not

It stops the publishing; it does not redact. The documents remain in this
repository's git history, and in the crates.io releases already cut, which are
immutable. If the goal is that they stop being retrievable rather than that
they stop being browsable, this PR is not enough and the issue thread is the
place to say so.

They also still have value to us. Nothing here preserves a copy anywhere
internal — git show 0baa3f1:docs/simple_someip/plans/<file> recovers any of
them if we want to move rather than drop.

Not touched

Cargo.toml still cites "the bare_metal_plan_v2 TriCore delta" in a dependency
comment — an internal plan name in a public, published manifest. Same class of
leak, different document, and it is not dangling, so I left it. Worth a
follow-up if we care.

Testing

• cargo fmt --all --check — clean.
• cargo clippy --workspace --no-default-features --features $ALLOC_FEATURES -- -D warnings -D clippy::pedantic — clean (CI's alloc/host lane).
• RUSTDOCFLAGS=-Dwarnings cargo doc --no-deps --no-default-features --features $ALLOC_FEATURES — clean.
• cargo package --list — no docs/ entries.

Note for anyone running these locally: --all-features does not build on this
crate and never did (bare-metal-runtime is nightly and mutually exclusive
with the alloc features) — use the CI feature sets, and include bare_metal
for the doc build or bare_metal_tasks cfg's out and its intra-doc links fail.

Review status

Not yet reviewed by a human.

🤖 Generated with Claude Code

[codecov-commenter]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ All tests successful. No failed tests found.

@@           Coverage Diff           @@
##             main     #159   +/-   ##
=======================================
  Coverage   81.72%   81.72%           
=======================================
  Files          48       48           
  Lines       16178    16178           
=======================================
  Hits        13222    13222           
  Misses       2956     2956           

Files with missing lines	Coverage Δ
src/client/mod.rs	82.02% <ø> (ø)
src/transport.rs	70.49% <ø> (ø)