![]() ![]() The FileMaker script step 'perform script on server', the option 'wait for completion' is crucial here. ![]() The developer should make sure everything is handled in the backend. With that timeout option set to 0 we have the possibility to perform a taks asynchronous which is to me a huge benefit. To me, it's up to the user to make sure he handles everything correct. Well we all know that system commands are extremely powerful and should be used with caution. Is there a use case for this that requires it not to timeout, that can't be solved via changing the command that you're using? We can change the Mac behaviour to launch a shell first, which seems to negate some of the issues, but could cause other issues. We can change the docs and effectively remove the 0 option, and force the user to do a timeout, which forces the closing of the command that you launched. It would be possible to modify the code match the documented behaviour but this would re-introduce a memory leak, which is why we changed behaviour in the first place. Letting the user launch multiple processes, which never complete and can't be properly closed by the plugin is a cause of issues, not really memory, but more file descriptors and other things, but effectively a leak. Starting a command that you just then try to quit kind of defeats the purpose of the command in the first place. Notably, with a 0 second timeout, the current implementation does not return immediately leaving the command running, it terminates the command (or tries to). ![]() Second: The changes in 4.1.3 are a modified version of the previous (2/3.x) implementation which does not address the changes to the specification/documented behaviour in 4.x. First: caffeinate is unusual but the behaviour will occur with other process with the current implementation. ![]()
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |