It's a more stable interface, if `shift()` suddenly will need to wait for smth like finalisation of
parallel running critical sections assuming there are no time jumps, we don't need to fix all its
usages.
If `delete_device_after` is configured, that period should be counted for webxdc instances from the
last status update, otherwise nothing prevents from deleting them. Use `msgs.timestamp_rcvd` to
store the last status update timestamp, it anyway isn't used for anything except displaying a
detailed message info. Also, as `ephemeral::select_expired_messages()` now also checks
`timestamp_rcvd`, we have an improvement that a message is guaranteed not to be deleted for the
`delete_device_after` period since its receipt. Before only the sort timestamp was checked which is
derived from the "sent" timestamp.
The limit is better enforced by webxdc distributors,
e.g. xdc store bots or actually email providers
to allow for experimentation with large frameworks
or porting existing apps and testing them
before reducing their size.
Besides that, the comment on WEBXDC_SENDING_LIMIT was outdated,
it was not updated when the limit was increased to 640 kB.
If a Delta Chat message has the Message-ID already existing in the db, but a greater "Date", it's a
resent message that can be deleted. Messages having the same "Date" mustn't be deleted because they
can be already seen messages moved back to INBOX. Also don't delete messages having lesser "Date" to
avoid deleting both messages in a multi-device setting.
send_webxdc_status_update JSON-RPC call
and corresponding Rust call sometimes fail in CI with
---
database is locked
Caused by:
Error code 5: The database file is locked
---
Adding more context to send_webxdc_status_update() errors
to better localize the error origin.
It can be used e.g. as a default in the file saving dialog. Also display the original filename in
the message info. For these purposes add Param::Filename in addition to Param::File and use it as an
attachment filename in sent emails.
Webxdc update messages may contain
long lines that get hard-wrapped
and may corrupt JSON if the message
is not encrypted.
base64-encode the update part
to avoid hard wrapping.
This is not necessary for encrypted
messages, but does not introduce
size overhead as OpenPGP messages
are compressed.
receive_imf() calls add_parts()
which INSERTs or UPDATEs the message using UPSERT [1].
It then uses last_insert_rowid() to get
the ID of the inserted message.
However, it is incorrect to use last_insert_rowid()
if an UPDATE was executed instead of INSERT.
The solution is to use `RETURNING id` clause
to make the UPSERT statement return message ID in any case [2].
The fix is tested in test_webxdc_update_for_not_downloaded_instance()
and with a debug_assert!.
[1] https://www.sqlite.org/lang_UPSERT.html
[2] https://sqlite.org/forum/forumpost/9ce3bc1c4a85c15f
Message.set_text() and Message.get_text() are modified accordingly
to accept String and return String.
Messages which previously contained None text
are now represented as messages with empty text.
Use Message.set_text("".to_string())
instead of Message.set_text(None).
Moved custom ToSql trait including Send + Sync from lib.rs to sql.rs.
Replaced most params! and paramsv! macro usage with tuples.
Replaced paramsv! and params_iterv! with params_slice!,
because there is no need to construct a vector.
To handle backups the UIs have to make sure they do stop the IO
scheduler and also don't accidentally restart it while working on it.
Since they have to call start_io from a bunch of locations this can be
a bit difficult to manage.
This introduces a mechanism for the core to pause IO for some time,
which is used by the imex function. It interacts well with other
calls to dc_start_io() and dc_stop_io() making sure that when resumed
the scheduler will be running or not as the latest calls to them.
This was a little more invasive then hoped due to the scheduler. The
additional abstraction of the scheduler on the context seems a nice
improvement though.
We do not make all transactions
[IMMEDIATE](https://www.sqlite.org/lang_transaction.html#deferred_immediate_and_exclusive_transactions)
for more parallelism -- at least read transactions can be made DEFERRED to run in parallel
w/o any drawbacks. But if we make write transactions DEFERRED also w/o any external locking,
then they are upgraded from read to write ones on the first write statement. This has some
drawbacks:
- If there are other write transactions, we block the thread and the db connection until
upgraded. Also if some reader comes then, it has to get next, less used connection with a
worse per-connection page cache.
- If a transaction is blocked for more than busy_timeout, it fails with SQLITE_BUSY.
- Configuring busy_timeout is not the best way to manage transaction timeouts, we would
prefer it to be integrated with Rust/tokio asyncs. Moreover, SQLite implements waiting
using sleeps.
- If upon a successful upgrade to a write transaction the db has been modified by another
one, the transaction has to be rolled back and retried. It is an extra work in terms of
CPU/battery.
- Maybe minor, but we lose some fairness in servicing write transactions, i.e. we service
them in the order of the first write statement, not in the order they come.
The only pro of making write transactions DEFERRED w/o the external locking is some
parallelism between them. Also we have an option to make write transactions IMMEDIATE, also
w/o the external locking. But then the most of cons above are still valid. Instead, if we
perform all write transactions under an async mutex, the only cons is losing some
parallelism for write transactions.
