[ILUG] Strange problem with oggdec, which became a question about time.

[Niall O Broin]

On Wednesday 25 February 2004, johnm at rte.ie (John Moylan) wrote:

>Use && instead of ;
>with && a command will only run after the first command has succeeded.
>
>date && time oggdec -o 123.wav dante.ogg && date

That makes no difference at all provided the time command doesn't fail,
which it won't.

But I suspect I've put my finger on it - the fast box I was testing on
is a dual HT Xeon. I expect

command_A; command_B; command_C

to do those commands, in that order, starting one after the other has
finished. But maybe on such a box scheduling happens differently.
TBH I don't see how - it seems it would break shell semantics. And it
doesn't seem to work like that. For instance,

date; sleep 2; sleep 2; sleep 2; date

shows the expected 6 second elapsed time. But yet when I do

date ;touch X1; time oggdec -o 123.wav dante.ogg ;touch X2; date

the output of the date commands and the time commands and of
ls --full-time X?  shows me that the oggdec took about 1 second -
despite the fact that more wallclock time has definitely elapsed.


This is well weird.



Niall