Return-Path: X-Original-To: notmuch@notmuchmail.org Delivered-To: notmuch@notmuchmail.org Received: from localhost (localhost [127.0.0.1]) by arlo.cworth.org (Postfix) with ESMTP id 695456DE0360 for ; Sat, 2 Apr 2016 07:16:11 -0700 (PDT) X-Virus-Scanned: Debian amavisd-new at cworth.org X-Spam-Flag: NO X-Spam-Score: -0.314 X-Spam-Level: X-Spam-Status: No, score=-0.314 tagged_above=-999 required=5 tests=[AWL=-0.314] autolearn=disabled Received: from arlo.cworth.org ([127.0.0.1]) by localhost (arlo.cworth.org [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id mWQiHFtSQ_J5 for ; Sat, 2 Apr 2016 07:16:03 -0700 (PDT) Received: from che.mayfirst.org (che.mayfirst.org [209.234.253.108]) by arlo.cworth.org (Postfix) with ESMTP id 25C5B6DE02D9 for ; Sat, 2 Apr 2016 07:15:48 -0700 (PDT) Received: from fifthhorseman.net (dhcp-a244.meeting.ietf.org [31.133.162.68]) by che.mayfirst.org (Postfix) with ESMTPSA id D20DCF99C for ; Sat, 2 Apr 2016 10:15:46 -0400 (EDT) Received: by fifthhorseman.net (Postfix, from userid 1000) id A9EBC20898; Sat, 2 Apr 2016 11:15:41 -0300 (BRT) From: Daniel Kahn Gillmor To: Notmuch Mail Subject: [PATCH v2 3/7] fix thread breakage via ghost-on-removal Date: Sat, 2 Apr 2016 11:15:37 -0300 Message-Id: <1459606541-23889-3-git-send-email-dkg@fifthhorseman.net> X-Mailer: git-send-email 2.8.0.rc3 In-Reply-To: <1459606541-23889-1-git-send-email-dkg@fifthhorseman.net> References: <1459445693-3900-1-git-send-email-dkg@fifthhorseman.net> <1459606541-23889-1-git-send-email-dkg@fifthhorseman.net> X-BeenThere: notmuch@notmuchmail.org X-Mailman-Version: 2.1.20 Precedence: list List-Id: "Use and development of the notmuch mail system." List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Sat, 02 Apr 2016 14:16:11 -0000 ghost-on-removal the solution to T590-thread-breakage.sh that just adds a ghost message after removing each message. It leaks information about whether we've ever seen a given message id, but it's a fairly simple implementation. Note that _resolve_message_id_to_thread_id also introduces new message_ids to the database, so i think just searching for a given message ID may introduce the same metadata leakage. This differs from v1 of this changeset in that we implement the change in _notmuch_message_delete, a more "internal" function. --- lib/database.cc | 2 +- lib/message.cc | 29 ++++++++++++++++++++++++++--- 2 files changed, 27 insertions(+), 4 deletions(-) diff --git a/lib/database.cc b/lib/database.cc index 3b342f1..d5733c9 100644 --- a/lib/database.cc +++ b/lib/database.cc @@ -2557,7 +2557,7 @@ notmuch_database_remove_message (notmuch_database_t *notmuch, if (status == NOTMUCH_STATUS_SUCCESS && message) { status = _notmuch_message_remove_filename (message, filename); - if (status == NOTMUCH_STATUS_SUCCESS) + if (status == NOTMUCH_STATUS_SUCCESS) _notmuch_message_delete (message); else if (status == NOTMUCH_STATUS_DUPLICATE_MESSAGE_ID) _notmuch_message_sync (message); diff --git a/lib/message.cc b/lib/message.cc index 8d72ea2..e414e9c 100644 --- a/lib/message.cc +++ b/lib/message.cc @@ -1037,20 +1037,43 @@ _notmuch_message_sync (notmuch_message_t *message) message->modified = FALSE; } -/* Delete a message document from the database. */ +/* Delete a message document from the database, leaving a ghost + * message in its place */ notmuch_status_t _notmuch_message_delete (notmuch_message_t *message) { notmuch_status_t status; Xapian::WritableDatabase *db; + const char *mid, *tid; + notmuch_message_t *ghost; + notmuch_private_status_t private_status; + notmuch_database_t *notmuch; + + mid = notmuch_message_get_message_id (message); + tid = notmuch_message_get_thread_id (message); + notmuch = message->notmuch; status = _notmuch_database_ensure_writable (message->notmuch); if (status) return status; - db = static_cast (message->notmuch->xapian_db); + db = static_cast (notmuch->xapian_db); db->delete_document (message->doc_id); - return NOTMUCH_STATUS_SUCCESS; + + /* and reintroduce a ghost in its place */ + ghost = _notmuch_message_create_for_message_id (notmuch, mid, &private_status); + if (private_status == NOTMUCH_PRIVATE_STATUS_SUCCESS) { + /* this is deeply weird, and we should not have gotten into + this state. is there a better error message to return + here? */ + return NOTMUCH_STATUS_DUPLICATE_MESSAGE_ID; + } else if (private_status == NOTMUCH_PRIVATE_STATUS_NO_DOCUMENT_FOUND) { + private_status = _notmuch_message_initialize_ghost (ghost, tid); + if (! private_status) + _notmuch_message_sync (ghost); + } + notmuch_message_destroy (ghost); + return COERCE_STATUS (private_status, "Error converting to ghost message"); } /* Transform a blank message into a ghost message. The caller must -- 2.8.0.rc3