Repository navigation
|
Per this commit's description, a module that hooks system_server should no longer requires a reboot. In my test a new module won't be active until I reboot, but Idk if that's the intended behavior or if the hot reload doesn't do its supposed things correctly. My device is lineageOS A15 with Vector 2.2 release build. |
Replies: 1 comment
|
Your test matches the intended behavior. Hot reload only replaces the code of a module that is already loaded in a running process. It does not add or remove a module from system_server. The PR you linked describes it this way: it is "replacing a module's code inside a running process", and a target is a process that already received the module from the daemon (#757). The reload is also only offered to modules that are already loaded in system_server. So:
The manager's scope screen documents this in a comment: "system_server reads its module list once, when it starts ... True for leaving as well as joining – a module already loaded into system_server stays loaded until that process goes." ( After you change the System Framework scope, the manager shows a dialog, "The framework has not picked this up yet" ( Vector/manager/src/main/res/values/strings.xml Lines 442 to 444 in efb8288 |
Your test matches the intended behavior. Hot reload only replaces the code of a module that is already loaded in a running process. It does not add or remove a module from system_server.
The PR you linked describes it this way: it is "replacing a module's code inside a running process", and a target is a process that already received the module from the daemon (#757). The reload is also only offered to modules that are already loaded in system_server. So:
hotReloadModule()or automatically if it setsautoHotReload=trueinmodule.propand agrees inonHotRel…