Merge pull request #1356 from deltachat/feat/async-jobs

asyncify core & remove manual thread handling

.github/workflows/code-quality.yml

@@ -10,7 +10,7 @@ jobs:
       - uses: actions-rs/toolchain@v1
         with:
           profile: minimal
-          toolchain: nightly-2020-03-12
+          toolchain: nightly-2020-03-19
           override: true
       - uses: actions-rs/cargo@v1
         with:
@@ -25,7 +25,7 @@ jobs:
       - uses: actions-rs/toolchain@v1
         with:
           profile: minimal
-          toolchain: nightly-2020-03-12
+          toolchain: nightly-2020-03-19
           override: true
       - run: rustup component add rustfmt
       - uses: actions-rs/cargo@v1
@@ -39,7 +39,7 @@ jobs:
       - uses: actions/checkout@v1
       - uses: actions-rs/toolchain@v1
         with:
-            toolchain: nightly-2020-03-12
+            toolchain: nightly-2020-03-19
             components: clippy
             override: true
       - uses: actions-rs/clippy-check@v1

ASYNC-API-TODO.txt

@@ -0,0 +1,54 @@
+
+Delta Chat ASYNC  (friedel, bjoern, floris, friedel)
+
+- smtp fake-idle/load jobs gerade noch alle fuenf sekunden , sollte alle zehn minuten (oder gar nicht)
+
+APIs:
+    dc_context_new                 # opens the database
+    dc_open                             # FFI only
+      -> drop it and move parameters to dc_context_new()
+
+    dc_configure                      # note: dc_start_jobs() is NOT allowed to run concurrently
+    dc_imex NEVER goes through the job system
+    dc_imex import_backup needs to ensure dc_stop_jobs()
+
+    dc_start_io                # start smtp/imap and job handling subsystems
+    dc_stop_io                 # stop smtp/imap and job handling subsystems
+    dc_is_io_running           # return 1 if smtp/imap/jobs susbystem is running
+
+    dc_close                            # FFI only
+      -> can be dropped
+    dc_context_unref
+
+    for ios share-extension:
+    Int dc_direct_send()  -> try send out without going through jobs system, but queue a job in db if it needs to be retried on failure
+        0: message was sent
+        1: message failed to go out, is queued as a job to be retried later
+        2: message permanently failed?
+
+EVENT handling:
+    start a callback thread and call get_next_event() which is BLOCKING
+    it's fine to start this callback thread later, it will see all events.
+    Note that the core infinitely fills the internal queue if you never drain it.
+
+    FFI-get_next_event() returns NULL if the context is unrefed already?
+
+   sidenote: how python's callback thread does it currently:
+         CB-thread runs this while loop:
+         while not QUITFLAG:
+             ev = context.get_next_event( )
+             ...
+         So in order to shutdown properly one has to set QUITFLAG
+         before calling dc_stop_jobs() and dc_context_unref
+
+    event API:
+        get_data1_int
+        get_data2_int
+        get_data3_str
+
+
+- userdata likely only used for the callbacks, likely can be dropped, needs verification
+
+
+- iOS needs for the share app to call "try_send_smtp" wihtout a full dc_context_run and without going
+