xev@hstbme.mit.edu (Xev Gittler) (06/04/88)
I know there was a whole discussion on this topic before, but I missed it. Could someone tell me the reason that elm ignores HUP's at some points, and mre importantly, is it bad to change it not to ignore them? Xev Gittler xev@hstbme.mit.edu, or xev@athena.mit.edu
mb@ttidca.TTI.COM (Michael Bloom) (06/06/88)
In article <5615@bloom-beacon.MIT.EDU> xev@hstbme.mit.edu (Xev Gittler) writes: > ... Could someone tell me the reason that elm ignores HUP's at some >points, and mre importantly, is it bad to change it not to ignore >them? I think it may be "bad" NOT to change it, considering that on many systems terminal opens can be invalidated (after which programs that do ignore HUPS may subsequently loop on apparent EOF's). While there may be critical sections to protect, it would be better to hold the HUP rather than to ignore it. You can do this for systems without sighold or sigblock by using a dummy handler that just notes that the HUP occurred and deal with it later.