>>> I have asked around in the kernel hacking circles when I had
>>> to choose the values and got told that 32 bits are enough for
>>> ino64_t. So live with it.
>>
>> No. I hope you realize now that it was a mistake to define
>> kernel-related things without waiting for the kernel implementation.
>
> Sure and Linus should have used a 64bit off_t when he wrote Linux 0.12.
> Hindsight is a wonderous thing.
Hindsight is a wonderous thing, and that is almost exactly the point.
The library developers had no way of knowing future needs. Given such
inability to guess the future, they should not have tried.
They did try though, and they failed. One can not blame them for the
failure itself. It is the very attempt to predict the future that
was a serious mistake.
As a result:
1. current systems translate system calls
2. near-future systems may translate new system calls
3. eventually the library API must be hacked
Had the developers not attempted to predict the future:
1. current systems would run fast
2. eventually the library API must be hacked
The long-term result is the same, but erroneous predictions of
the future cause so much additional suffering. It is better to
not predict at all than to risk such errors, because everyone
must pay the cost of mistakes via wasted memory and CPU time.
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/