info-vax@ucbvax.ARPA (04/15/85)
From: *Hobbit* <AWalker@RUTGERS.ARPA> The discussion about RMS's synchronous behavior reminded me of another question. It used to be, under 3.x, that if you opened a mailbox [in DCL] and read from it or wrote to it, your process would LEF until the I/O was completed [like something read the mailbox at the other end]. This would happen if you did something like issued two writes; the second one would hang until the first one was read out again, emptying the mailbox. Now, under 3.x, you could ^Y out of this condition and regain control. It was a fix from 2.x in which your process would get hopelessly hung until someone else read/wrote the mailbox. Now, under 4.0, you *again* get hung if you do this! Why did they remove this??!? Is there an easy fix I'm missing? [Don't tell me to $set control=y; I've transcended to more desperate hacks.] There are times when I mess with mailboxes a lot, and being able to punt a write was real handy. Now, I have to fire up a subprocess and kill it if it gets hung. _H* -------