fix(truncate): enforce VFS limits before resize

[chaliy]

What

Reject truncate target sizes that exceed the active VFS max_file_size/max_total_bytes before resizing the in-memory file buffer.

Why

truncate -s 4G /tmp/p previously passed the builtin's hard-coded 4 GiB check and attempted to allocate/zero-fill before write_file enforced default 10 MB VFS limits.

How

• Replace the hard-coded truncate cap with ctx.fs.limits()-derived validation.
• Use usize::try_from before allocation to avoid truncating oversized lengths on narrow platforms.
• Add a regression test proving oversized targets fail before the write path.
• Update TM-DOS-005 threat-model mitigation text.

Risk

• Low
• Could reject truncate requests above the configured VFS limit earlier than before; this matches VFS quota semantics and avoids allocation DoS.

Checklist

• Tests added or updated
• Backward compatibility considered

Verification

• CARGO_BUILD_JOBS=1 cargo test -p bashkit --lib builtins::truncate::tests::rejects_target_above_vfs_limit_before_write --all-features failed before the fix with cannot write ... file too large, then passed after the fix.
• CARGO_BUILD_JOBS=1 cargo test -p bashkit --lib builtins::truncate::tests --all-features
• just pre-pr
• cargo run -p bashkit-cli -- -c "truncate -s 4G /tmp/p; echo exit=\$?; test ! -e /tmp/p && echo no-file"

[cloudflare-workers-and-pages]

Deploying with    Cloudflare Workers

The latest updates on your project. Learn more about integrating Git with Workers.

Status	Name	Latest Commit	Preview URL	Updated (UTC)
✅ Deployment successful! View logs	bashkit	cd5fc5a	Commit Preview URL	May 08 2026, 08:21 PM