tom@ssd.csd.harris.com (Tom Horsley) (02/26/90)
>If you could tell me if it keeps working with only gimme and sp volatilized, >I'd be much obliged. I'm not planning on volatilizing the others. It seems to work fine with just these two variables volatilized (especially once I fixed perl.h to say "#if defined(__STDC__) || defined(_CX_UX)" instead of "#ifdef defined(__STDC__) || defined(_CX_UX)" (which is what I started with). Amusing anecdote follows: I did have some trouble testing this however. I spent all day trying to figure out why the base.term test was failing, sometimes it would work, the next time I would compile with different options and it would fail. It was only after I watched the exact same version of perl fail the test once and pass it the next time that I got desperate enough to type "ls -l /dev/null". That was when I discovered that /dev/null was an ordinary file (NOT a character special device). It seems that someone on our system managed (somehow, no one is quite sure how) to remove /dev/null. Naturally, the next time something re-directed its output to /dev/null, that re-created it, only now it was an ordinary file. Each time some other task re-directed output to /dev/null, it would change contents randomly, sometimes it would be zero length, sometimes not. Since base.term expects to read /dev/null and get an EOF (a fairly reasonable assumption) I was getting non-deterministic behavior from perl. On a confusion index of 1 to 10, this one has got to rate an 11. -- ===================================================================== domain: tahorsley@ssd.csd.harris.com USMail: Tom Horsley uucp: ...!novavax!hcx1!tahorsley 511 Kingbird Circle or ...!uunet!hcx1!tahorsley Delray Beach, FL 33444 ======================== Aging: Just say no! ========================