fix: do not run more than one housekeeping at a time

With multiple transports there are multiple inbox loops on the same profile `Context`. 
They tend to start running housekeeping at the same time, e.g. when deleting
a message with an attachment, and then `remove_unused_files()`
tries to remove the same files that are already deleted by another thread
and logs errors.

src/context.rs

@@ -239,6 +239,9 @@ pub struct InnerContext {
     pub(crate) oauth2_mutex: Mutex<()>,
     /// Mutex to prevent a race condition when a "your pw is wrong" warning is sent, resulting in multiple messages being sent.
     pub(crate) wrong_pw_warning_mutex: Mutex<()>,
+    /// Mutex to prevent running housekeeping from multiple threads at once.
+    pub(crate) housekeeping_mutex: Mutex<()>,
+
     pub(crate) translated_stockstrings: StockStrings,
     pub(crate) events: Events,
 
@@ -478,6 +481,7 @@ impl Context {
             generating_key_mutex: Mutex::new(()),
             oauth2_mutex: Mutex::new(()),
             wrong_pw_warning_mutex: Mutex::new(()),
+            housekeeping_mutex: Mutex::new(()),
             translated_stockstrings: stockstrings,
             events,
             scheduler: SchedulerState::new(),

src/sql.rs

@@ -828,6 +828,10 @@ async fn incremental_vacuum(context: &Context) -> Result<()> {
 
 /// Cleanup the account to restore some storage and optimize the database.
 pub async fn housekeeping(context: &Context) -> Result<()> {
+    let Ok(_housekeeping_lock) = context.housekeeping_mutex.try_lock() else {
+        // Housekeeping is already running in another thread, do nothing.
+        return Ok(());
+    };
     // Setting `Config::LastHousekeeping` at the beginning avoids endless loops when things do not
     // work out for whatever reason or are interrupted by the OS.
     if let Err(e) = context