Removing a relay now takes effect immediately:
- the profile stops fetching and advertising it,
- secondary devices immediately apply the removal through the transport sync,
Upgrading removes unpublished relays and triggers keyupdates.
BREAKING CHANGE: set_transport_unpublished() is removed: UIs call delete_transport() when the user removes a relay.
BREAKING CHANGE: list_transports_ex() and the TransportListEntry type are removed: use list_transports().
BREAKING CHANGE: delete_transport() no longer refuses to remove the primary transport: it refuses only to remove the last one and re-elects the sending transport as needed.
BREAKING CHANGE: TransportsModified is now also emitted on the device modifying the transports, not only on devices applying the synced change.
Deprecated: DC_STR_PHASING_OUT
`relay_addrs()` fell back to the contact address even when it is empty,
which happens for key-contacts created from a sync message or for the
self-contact, putting an empty string into the SMTP recipient list.
When the published relay list changes, key-contacts are informed with an
unsigned message carrying the re-signed key, encrypted to a chunk of contacts
at a time. It is shaped like a receipt notification naming no message, so that
cores which know nothing about keyupdates trash it as well.
See the src/keyupdate.rs module docs for the design.
An unsigned message carries no intended recipient fingerprints,
so recipients of an encrypted unsigned message
learn nothing about other recipients from the PGP packets.
No functional changes:
Add a relay_addrs helper, share the protected headers and self-key rendering
of non-chat messages, and move insert_into_smtp from securejoin to smtp.
all published deltachat-rpc-server wheels so far fail "wheel tags"
and probably other tools.
Translate "dev" cargo-versioning to PEP440-versioning
to make CI dev wheel builds reproducable at least for the 11 non-mac targets.
this is meant to help configuration event consumers (Python, UIs)
to not hang waiting for configuration outcomes.
One test case is added that fails on main.
Requests now register before testing for shutdown,
so either the reader loop or the caller answers them.
Also failed start() winds down its threads, ignoring a broken pipe on stdin.
also core src/imap.rs retries connecting (2s with backoff) but direct_imap is only used in (CI) tests
and we can just keep trying every second until pytest timeout hits.
As we are using a dedicated CI relay instance, we don't need to be careful with creating test addresses.
But re-using test addresses in a test running in a single worker can leak events/messages which compromises test isolation.
Instead of encountering random flaky tests because let's just not cache addresses to maximize test isolation.
previously only 10 messages were sent but if joinining did not succeed (due to slow CI sometimes) nothing would be received and the test would hang.
Also cleans up the tests to not set webxdc realtime config (true by default) and restructure helpers to reduce redundant code.
Devices no longer implicitely use the From address of sync messages
to determine their primary transport. Receivers have no concept of it and own devices
may disagree on which relay is reachable because of VPN or different networks:
1. Make setting a primary transport (`configured_addr`) a per-device non-synced operation.
2. Transport rows (add/remove/unpublish) keep syncing like before.
3. A device reelects a primary if a sync message unpublished/removed the current primary
if there is a better candidate.
4. `TransportsModified` event is emitted at most once on an incoming transport sync message.
Users will notice the change in that changing primary transport in settings/advanced/relays
will not synchronize to other devices anymore.
Came across this while investigating more test_transport_synchronization flakiness,
sometimes missing TransportsModified events or getting a missing configured_addr.
The underlying problem was that stopping IO was triggered immediately during
receiving sync messages, potentially *canceling* the processing of the sync message,
effectively de-syncing the device's view on transports.
register device token if XDELTAPUSH IMAP capability is available (current relay setup provides it)
or if "maxsmtprecipients" IMAP metadata key is available (relays since May 2026 provide it),
allowing chatmail relay setups to drop XDELTAPUSH capability marker with the next release
while retaining push notification support.
- treat an Autocrypt header as a kind of Autocrypt-Gossip: addr=<self>
letting it participate in the existing cooperative key gossipping mechanics.
- speed up processing of incoming autocrypt-gossip headers by using
1 sql commit for N gossip headers instead of the current N commits.
unsuspecting consumers will otherwise require execnet to run pytest with deltachat-rpc-client installed.
"import py" is discouraged for pytest these days but still supported in compatibility mode.
BREAKING CHANGE: provider lookup APIs were removed from CFFI and JSON-RPC.
also removes offline provider database code and generated provider data,
provider-specific fields in configure/transport paths, and REPL providerinfo.
Heartbeat notifications are only used on iOS for classic mail servers
but both code and privacy wise (notification server sees IP addresses from those users)
not something we want to support any longer (was discussed with bjoern).
Remove is_chatmail flag, and rely on IMAP metadata
advertising the recipients limit (usually 1000),
falling back to 50 (or fewer in some exceptional cases)
just as before when is_chatmail was false.
Also, server metadata is now keyed per transport,
which in the future eases collecting ICE servers
from all relays (instead of just any first relay).
BREAKING CHANGE: removed oauth2 module, dc_get_oauth2_url FFI function, DC_LP_AUTH flags and configured/serverflags, and the oauth2 parameter/field from SMTP/IMAP clients, JSON-RPC interfaces, and CLI tools.
also contains regenerated provider data after dropping oauth in the update script.
Rpc.start() now calls get_system_info() after launching the server
to verify it started successfully. If the server exits early (e.g.
due to an invalid accounts directory), the core error message from
stderr is captured and included in the raised JsonRpcError.
The reader_loop now unblocks pending RPC requests when the server
closes stdout, so callers never hang on a dead server.
Export JsonRpcError from the deltachat_rpc_client package.
Add test_early_failure verifying that Rpc.start() raises with
the actual core error message for invalid accounts directories.
holger krekel
