On 6 June 2015 at 00:22, Karl Berry <[email protected]> wrote: > Better to upset whatever terminal settings the user had beforehand > > Clearly, doing the stty on this system is desirable. > > Doing the stty on all systems is what I'm worried about, i.e., > upsetting things that were fine before. > > I don't know how we're supposed to check > if the shell is broken or not. > > No need to check if it's broken specifically. It's enough to check if > we're on that system (simple uname) and then do the stty or switch to > another shell or skip the test.
What if the same broken shell/behaviour exists on other systems, or if someone updates the shell to work on that system? I suppose we could first check for stty, then look for "-echo" in the output of stty, and then after a timed out read, look for it again and see if it changed, and only then run "stty sane". Alas this shell portability stuff is a massive time sink and every new OS seems to bring new problems.
