Status Update
Comments
ap...@google.com <ap...@google.com> #3
lblb636@ Thanks for reporting this bug. I just want to confirm that the issue is the call device
button being not responsive according to the uploaded video. It seems like this issue only happened on Windows.
ap...@google.com <ap...@google.com> #4
And why only from specific API of Android?
On API 28 for example, it doesn't occur...
Please fix this issue for all API versions of the emulator.
ap...@google.com <ap...@google.com> #5
Another similar issue that's related
ap...@google.com <ap...@google.com> #6
RE#4 I need to try this on Windows laptop first and I couldn't repro on MacOS or Linux. But I think this is most likely related to the modem simulator we introduced for API 31. Will need to further investigate.
ap...@google.com <ap...@google.com> #7
Why isn't the emulator tested more on Windows OS?
Windows OS is more popular than Linux and MacOS combined (on desktop) ...
ap...@google.com <ap...@google.com> #8
ap...@google.com <ap...@google.com> #9
I reproduced this bug on Windows with emulator version 31.2.10-8420304
ap...@google.com <ap...@google.com> #10
il...@google.com <il...@google.com> #11
Found the root cause when the "call device" button was unresponsive. It became unresponsive because the inbound call failed to initialize. Basically, the modem simulator's shared FD failed. When receive_inbound_call is invoked, it is supposed to make connection to the "host server" via a shared fd but it is flaky on Windows unfortunately.
modem-simulator: receive_inbound_call:101 inbound call from 6505551212
modem-simulator: process_msgs:296 waiting for new messages ...
modem-simulator: main_host_thread:501 got connection at host server at 65094
modem-simulator: start_calling_thread:254 making a connection to main host server 65094
modem-simulator: main_host_thread:507 Detected close from the other side
modem-simulator: start_calling_thread:262 nothing read quit
modem-simulator: main_host_thread:494 looping at main host server at 65094
modem-simulator: start_calling_thread:284 done with this thread
Description
Navigation Compose has APIs that would be a better fit as part of the base Navigation library. Instead of allowing Navigation Compose to go to 1.0.0 stable with the libraries and attempting to deprecate them later when they are added to base Navigation, we should instead add those APIs to the next version of Navigation and move Navigation Compose in sync with that.
This means that Navigation Compose will be part of Navigation 2.4.0 instead of an 1.0.0.