This commit adds support for IMAP servers that don't provide UIDVALIDITY,
such as Tencent Enterprise Mail (腾讯企业邮箱).
Changes:
- Added `generate_synthetic_uidvalidity()` function that creates a stable
hash-based UIDVALIDITY from the mailbox name
- Modified `reconcile_mailboxes()` to use synthetic UIDVALIDITY when the
server doesn't provide one
- Servers without UIDVALIDITY can now sync all mailboxes including system
folders (Sent Messages, Drafts, Deleted Messages)
- Incremental sync is supported via the synthetic UIDVALIDITY
- Added warning logs to indicate when synthetic UIDVALIDITY is in use
- Updated mailbox metadata to store the resolved UIDVALIDITY
Fixes issues with:
- Tencent Enterprise Mail (腾讯企业邮箱)
- Other non-compliant IMAP servers
- Mailboxes that don't properly support UIDVALIDITY"
When the request returns an empty body, choose to skip it and print the account ID and UID information, leaving it for the user to investigate themselves. Otherwise, the IMAP download process will be blocked by this.
- Treat MIME parts with Content-ID but no Content-Disposition as inline
- Add account_ids filter to CLI export search to avoid pulling all accounts
- Skip failed emails during export instead of aborting the entire batch
- Switch memdb durability from Full to Batch(100) with 10s flush worker
- Offload BlobManager fjall writes to spawn_blocking
- Wrap Tantivy commit operations in block_in_place
- Flush memdb WAL on graceful shutdown
InternalDate and IngestAt were renamed for the wire with #[serde(rename)] only. SortBy derives poem_openapi::Enum, which does not honour serde attributes, so the REST deserializer exposed them under their Rust identifiers instead of the intended INTERNAL_DATE / INGEST_AT — inconsistent with the existing DATE/SIZE values and rejecting the documented names with HTTP 400. Add #[oai(rename = ...)] alongside the serde rename.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
POST /api/v1/search-messages previously filtered and sorted only on the
sender-controlled Date: header. Add support for two server-controlled
timestamps that are already indexed as FAST i64 fields:
- internal_date (IMAP INTERNALDATE)
- ingest_at (Bichon's archival time)
EmailSearchFilter gains internal_date_since/before and ingest_since/before
range bounds, mirroring the existing `since`/`before` Date: handling.
SortBy gains InternalDate and IngestAt variants (wire values INTERNAL_DATE
and INGEST_AT), mirroring the existing DATE/SIZE sort handling.
The envelope and attachment Tantivy schemas already declare these fields
as INDEXED | STORED | FAST, so no re-index or migration is required.
Attachments carry no IMAP INTERNALDATE, so the attachment search maps the
InternalDate sort to the attachment's own date field as a defined fallback.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
The download-message REST endpoint returned 404 [30000] "Original email
content not found" whenever an indexed message's raw content blob was
missing from the blob store, leaving the message permanently
unrecoverable even though it still existed on the IMAP server.
Make the endpoint self-healing: when the content blob is absent, fetch
that one message on demand from the IMAP server via
UID FETCH <uid> (BODY.PEEK[]), repopulate the detached blob, and return
the content. The 404 is now only produced when the on-demand fetch
itself fails (mailbox gone, UID gone, connection failure, or the fetched
bytes no longer match the archived content_hash).
- Add ImapExecutor::fetch_single_message_body: examines the mailbox
read-only and fetches one message by UID, reusing the existing
BODY.PEEK[] fetch command.
- Add reattach_eml_content_self_healing / recover_message_blob in the
envelope extractor: fast-path delegates to reattach_eml_content when
the blob exists; otherwise recovers it via IMAP and re-stores the
stripped EML + attachments through the existing blob queue.
- Make store::blob::get_reader async and route it through the
self-healing path; update the single caller (download_message).
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
cleanup_unused_content decides whether to delete a deduplicated blob by
running a Tantivy Count of envelopes referencing each content_hash. The
searcher it used reflected only the committed index state at the time it
was built, so an envelope that shared a content hash but was still
sitting uncommitted in the IndexWriter buffer (for example added by the
background ingest task before the delete operation acquired the writer
lock) was invisible to the count. The count read 0 and a
still-referenced blob was deleted, permanently 404ing that envelope's
download-message.
Pass the locked IndexWriter into cleanup_unused_content and fatal_commit
it immediately before creating the searcher. create_searcher already
reloads the reader, so the Count is now evaluated against a fully
committed, freshly-reloaded index state. The barrier is local to the GC
path and self-contained, independent of what the caller committed.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
- Add 10-minute cache after fetching mailbox list from mail accounts
- Display progress during initial mailbox retrieval
- Prevent timeouts when handling large mailbox lists
Lei Zhu