The greeting is now always read manually,
even for STARTTLS connections,
so the errors returned on failure to read form the stream
are the same regardless of the connection type.
This change introduces new config options
`proxy_enabled` and `proxy_url`
that replace `socks5_*`.
Tested with deltachat-repl
by starting it with
`cargo run --locked -p deltachat-repl -- deltachat-db` and running
```
> set proxy_enabled 1
> set proxy_url ss://...
> setqr dcaccount:https://chatmail.example.org/new
> configure
```
Previously for each connection candidate (essentially host and port
pair) after resolving the host to a list of IPs Delta Chat iterated IP
addresses one by one. Now for IMAP and SMTP we try up to 5 IP addresses
in parallel. We start with one connection and add more connections
later. If some connection fails, e.g. we try to connect to IPv6 on IPv4
network and get "Network is unreachable" (ENETUNREACH) error, we replace
failed connection with another one immediately.
Co-authored-by: Hocuri <hocuri@gmx.de>
Apparently some providers fail TLS connection
with "no_application_protocol" alert
even when requesting "imap" protocol for IMAP connection
and "smtp" protocol for SMTP connection.
Fixes <https://github.com/deltachat/deltachat-core-rust/issues/5892>.
Single function smtp::connect::connect_stream
returns a stream of a single `dyn` type
that can be a TLS, STARTTLS or plaintext
connection established over SOCKS5 or directly.
If a time value doesn't need to be sent to another host, saved to the db or otherwise used across
program restarts, a monotonically nondecreasing clock (`Instant`) should be used. But as `Instant`
may use `libc::clock_gettime(CLOCK_MONOTONIC)`, e.g. on Android, and does not advance while being in
deep sleep mode, get rid of `Instant` in favor of using `SystemTime`, but add `tools::Time` as an
alias for it with the appropriate comment so that it's clear why `Instant` isn't used in those
places and to protect from unwanted usages of `Instant` in the future. Also this can help to switch
to another clock impl if we find any.
a27e84ad89 "fix: Delete received outgoing messages from SMTP queue"
can break sending messages sent as several SMTP messages because they have a lot of recipients:
`pub(crate) const DEFAULT_MAX_SMTP_RCPT_TO: usize = 50;`
We should not cancel sending if it is such a message and we received BCC-self because it does not
mean the other part was sent successfully. For this, split such messages into separate jobs in the
`smtp` table so that only a job containing BCC-self is canceled from `receive_imf_inner()`. Although
this doesn't solve the initial problem with timed-out SMTP requests for such messages completely,
this enables fine-grained SMTP retries so we don't need to resend all SMTP messages if only some of
them failed to be sent.
* do not use ratelimiter for bots
i tried some other approaches as having optional ratelimiter
or handle `can_send()` for bots differently,
but all that results in _far_ more code and/or indirections -
esp. as the "bot" config can change and is also persisted -
and the ratelimiter is created
at a point where the database is not yet available ...
of course, all that could be refactored as well,
but this two-liner also does the job :)
* update CHANGELOG
New ratelimiter module counts number of sent messages and calculates
the time until more messages can be sent.
Rate limiter is currently applied only to MDNs. Other messages are
sent without rate limiting even if quota is exceeded, but MDNs are not
sent until ratelimiter allows sending again.
Switches from rusqlite to sqlx to have a fully async based interface
to sqlite.
Co-authored-by: B. Petersen <r10s@b44t.com>
Co-authored-by: Hocuri <hocuri@gmx.de>
Co-authored-by: link2xt <link2xt@testrun.org>
This changes the internal stock strings API to be more strongly typed,
ensuring that the caller can not construct the stock string in the
wrong way.
The old approach left it to the callers to figure out how a stock
string should be created, now each stock string has their specific
arguments and callers can not make mistakes. In particular all the
subtleties and different ways of calling stock_system_msg() disappear.
This could not use a trait for stock strings, as this would not allow
having per-message typed arguments. So we needed a type per message
with a custom method, only by convention this method is .stock_str().
The type is a enum without variants to avoid allowing someone to
create the type.
Sadly the fallback string and substitutions are still far away from
each other, but it is now only one place which needs to know how to
construct the string instead of many.
Previously MX records were queried only for OAuth 2 configuration and
did not affect the list of servers tried. User was required to manually
configure the servers for Google Workspace (former GSuite) domains.
Now MX records are queried during configuration. If provider is found in
offline database, its ID, corresponding to the filename, is saved as
`configured_provider`.
`configured_provider` is also set during database migration if email
address uses the domain from the provider database, but no MX querying
is done.
providers may have a maximum for the SMTP RCPT TO: header,
therefore, an outgoing message is split into several chunks.
the chunk size defaults to 50, however, if known to be larger or smaller
by a dedicated provider, this setting may be changed.
this does not affect the MIME To: headers,
if a provider has a maximum here, this is change would not help.
Previously DC_EVENT_ERROR_NETWORK events were emitted for each failed
attempt during autoconfiguration, even if eventually configuration
succeeded. Such error reports are not useful and often confusing,
especially if they report failures to connect to domains that don't
exist, such as imap.example.org when mail.example.org should be used.
Now DC_EVENT_ERROR_NETWORK is only emitted when attempting to connect
with existing IMAP and SMTP configuration already stored in the
database.
Configuration failure is still indicated by DC_EVENT_CONFIGURE_PROGRESS
with data1 set to 0. Python tests from TestOnlineConfigurationFails
group are changed to only wait for this event.
