[PATCH] unify itoa

Rob Landley rob at landley.net
Tue Jul 11 15:36:44 UTC 2006


On Monday 10 July 2006 10:13 am, Michael S. Zick wrote:
> On Mon July 10 2006 04:46, Denis Vlasenko wrote:
> > I think I'm starting to waste your (and my) time.
> > Alpha is almost history now...
>
> Almost for some,
>
> (HP) Vax and MicroVax - no new machines, remarket only
> (HP) Alpha Servers - new orders only until Oct. 2006
> (HP) PA-RISC, 64-bit - alive and well - no 64-bit userland
> under Linux - 64-bit Linux kernel alive and well.

I always assumed that PA-RISC used HP-PA processors.  (Which I vaguely recall 
Linux being ported to by The Puffin Group, years ago...)

> Two on-going efforts to keep in mind...
>
> Giving parisc-Linux a 64-bit userland is on-going, but you
> will not see these in an embedded system.

I still think PA-RISC is an architecture from HP.  And the parisc-linux 
website would appear to agree with me: http://www.parisc-linux.org/

> If an adaption of BB gets shoved into the kernel as some
> propose - then handling 64-bit & 128-bit types may come up again.

*blink* *blink*

No.

And I'm not against handling 64 bit and 128 bit types, I'm just saying 
that "int" is 32 bit (not 64 or 128 bit) on any platform we care to support.  
Long is 64 bits on some platforms of interest to us, and long long is 128 bit 
on some, 64 bit on others.

> But still, sizeof(unsigned int) is set by choosing the compiler
> to use for the code.  There are only two widely used compilers
> for pa-risc - HP and GCC - those are 32-bit unsigned int for
> both compilers, both 32-bit and 64-bit machines.
> (The hardware handles the masking and sign extension.)
>
> Now if you wanted to know about the sizeof(long double)...
> Thank goodness, you don't.

We largely avoid floating point.  Where we use it, we usually have a CONFIG 
option, ala SORT.

Rob
-- 
Never bet against the cheap plastic solution.



More information about the busybox mailing list