refactor(e2b): keep provider layer substrate-only

This commit is contained in:
Tianyi Cui
2026-07-29 19:07:11 +08:00
parent 65b345c96a
commit 917a7493f7
28 changed files with 280 additions and 578 deletions

View File

@@ -6,10 +6,9 @@ E2B implementation of the [`@deepseek-ai/dsh-fs`](../../fs/fs/README.md) provide
## Behavior
- **Remote identity and metadata** — relative paths resolve as POSIX paths against the caller cwd or `ctx.e2b.cwd`; `realpath -m` supplies canonical target identity without requiring the final file to exist. `stat`, no-follow `lstat`, and stable one-level directory listings project E2B metadata into the filesystem seam. Versions are opaque hashes of E2B metadata plus a per-write extended attribute.
- **Remote identity and metadata** — relative paths resolve as POSIX paths against the caller cwd or `ctx.e2b.cwd`; GNU `realpath -mz` supplies canonical target identity without requiring the final file to exist, and ASCII/base64 plus strict NUL framing preserves newline and multibyte paths across the decoded SDK transport. `stat`, no-follow `lstat`, and stable one-level directory listings project E2B metadata into the filesystem seam; listings reuse returned metadata and resolve symbolic-link entries sequentially. Versions are opaque hashes of E2B metadata plus a per-write extended attribute.
- **Execution-world paths** — canonical targets expose absolute POSIX process paths, percent-encoded `file:` URIs, and provider-owned containment checks, so generic subprocess consumers never parse E2B target ids or apply host path rules.
- **UTF-8 reads** — whole reads and streamed reads preserve cross-chunk decoding, reject invalid UTF-8, and use the seam's 8192-byte NUL sample for binary detection. The model-facing tool still owns size selection and line windowing.
- **Stable bounded reads** — a dependency-free Node helper walks directory descriptors with no-follow opens and reads one held regular-file descriptor through the byte cap. Generic LSP queries therefore reject parent swaps, non-files, invalid UTF-8, and growth past the configured document limit before server startup.
- **Atomic mutations** — writes create a random sibling staging directory, change it to mode `0700` before uploading content, preserve an existing file's POSIX mode, and publish the staged file through E2B's same-filesystem atomic rename. The rename response supplies the committed version, so no fallible metadata request follows the commit point. E2B creates missing parent directories. Literal edits LF-normalize for matching, restore dominant CRLF storage, and serialize mutations per canonical target within the host process. Optional create/version guards keep the base seam's observed-state semantics.
- **Failures and cancellation** — E2B not-found, permission, abort, and other controller failures map to the existing `FsError` vocabulary. Cancellation is best-effort at earlier SDK request boundaries and checked immediately before rename. The signal is not forwarded into the rename RPC, so cancellation cannot interrupt the atomic commit; a successful rename is the commit point.
@@ -27,5 +26,6 @@ No direct invalidation; the named consumer owns any request-prefix changes.
- **No host synchronization** — an empty E2B cwd stays empty until a tool, command, template, or external process populates it; local files are neither uploaded nor reflected back.
- **Mutation coordination is host-process-local** — another harness connection or remote command can race the adapter; version guards detect only metadata changes represented by E2B.
- **Reads reopen canonical targets by path** — a concurrent remote path replacement between resolution and stream opening is not fenced by a stable file handle; no observed product defect justifies a provider-specific bounded-read protocol in this POC.
- **Whole-file mutation costs remain** — overwrite diffs and literal edits read complete files into host memory, and every operation incurs E2B controller latency.
- **Custom templates must support the used Linux, Node, procfs, and envd features** — `realpath`, `chmod`, `mv`, same-filesystem POSIX rename, streaming reads, file metadata extended attributes, `/proc/self/fd`, and no-follow descriptor opens are required; unsupported templates fail rather than degrade silently.
- **Custom templates must support the used Linux/GNU and E2B filesystem features** — `realpath -mz`, `base64 -w0`, `chmod`, same-filesystem rename, streaming reads, and file metadata extended attributes are required; unsupported templates fail rather than degrade silently.