There is a bug in the "alarm" function in IRIX 4.0.5 (I’m not sure if it’s in
earlier versions). If you pass a value in the range 2148 to 2385 inclusive
to "alarm", the alarm goes off immediately (instead of 36 to 40 minutes
later). Other values seem OK, although we can’t pretend to have made a
thorough check.
The person who discovered this bug spent a lot of time finding it, so I
though I’d post it in case it saves someone else some time. This has been
reported to the TAC (call ID 170161), and apparently it’s fixed in 5.0.
—
John Hogg; h…@ucs.ubc.ca; Computing Services, University of British Columbia
In <1ovembINN…@iskut.ucs.ubc.ca> h…@ucs.ubc.ca (John Hogg) writes:
| There is a bug in the "alarm" function in IRIX 4.0.5 (I’m not sure if it’s in
| earlier versions). If you pass a value in the range 2148 to 2385 inclusive
| to "alarm", the alarm goes off immediately (instead of 36 to 40 minutes
| later). Other values seem OK, although we can’t pretend to have made a
| thorough check.
|
| The person who discovered this bug spent a lot of time finding it, so I
| though I’d post it in case it saves someone else some time. This has been
| reported to the TAC (call ID 170161), and apparently it’s fixed in 5.0.
Yep. I just checked the code, and somebody blew it on the calculation
of seconds to clock ticks, and the upper bound of the check is 2386,
when it should have been 2148 (it’s all done with a bunch of macros
and defined values, and somebody missed the effect).
—
Let no one tell me that silence gives consent, | Dave Olson
because whoever is silent dissents. | Silicon Graphics, Inc.
Maria Isabel Barreno | ol…@sgi.com