Closed bollwyvl closed 8 years ago
I'm a little concerned about the concept of auto-pushing on save. We do automatic saves in the revisioncontrol extension, for example, which might cause unexpected publication. Have we considered the possibility of a simple status widget (dirty bit) that indicates the existence of changes that haven't been uploaded?
Ok, let's table this for now!
This came up in the meeting last week as kind of a poor-man's multi-user collaboration. Thoughts and ramblings below...
yield
y business, doing a publish blocks pretty much everything, whereas File->Save is handled asynchronously. This is probably a good thing to do anyway, and would probably just require a little refactoring to put the expensive method into a ThreadPoolExecutor or whatever pattern which has likely been done on other extensions, though we may have to do some custom kernel message.Cmd+S
people like me aren't making meeelions of notebooks. by the same token, it's always the last save the counts, so perhaps it's more like a tailing debounce, i.e. "publish every time i hit save, unless it's been less than five seconds... but check again in 10 seconds"@damianavila @malev @mmarchetti @ijstokes