Re: Wallclock C-api functions

[Sergei Gavrikov <w3sg at SoftHome dot net>]

Bart, thank you for your explanation. It's pity what "eCos WallClock" is
an absent section in eCos Reference). And the cyg_libc_time_settime() is
just that what I wanted. The wallclock2.cxx test contains only two calls
from the cyg_libc_time_* set. Thanking you, "nm" did talk me about other
ones.

Thanks!

	Sergei


On Sun, Sep 02, 2007 at 11:30:08AM +0100, Bart Veer wrote:
> >>>>> "Sergei" == Sergei Gavrikov <w3sg@SoftHome.net> writes:
> 
>     Sergei> RFC
> 
>     Sergei> I would want to manage the eCos wallclock (to set/get it's
>     Sergei> time) using a C API, but, it seems that isn't possible
>     Sergei> (correct me, please, if I did miss something). Can we fix
>     Sergei> that issue using same intervention in the C++ world?
> 
> Reading the wallclock is already supported by the standard C library
> routines, time() etc., in CYGPKG_LIBC_TIME. Setting the wallclock is
> possible via the eCos extension cyg_libc_time_settime() in the same
> package. Hence accessing the wallclock via a C API is already
> straightforward, and adding equivalent functions to the generic
> wallclock package does not gain us anything.
> 
> There is an argument for moving the two core functions, time() and
> cyg_libc_time_settime() into CYGPKG_IO_WALLCLOCK so that they are
> readily available to configurations which do not involve the C
> library, but I do not think the gains (if any) are worth the
> disruption.
> 
> Bart
> 
> -- 
> Bart Veer                                   eCos Configuration Architect
> eCosCentric Limited    The eCos experts      http://www.ecoscentric.com/
> Barnwell House, Barnwell Drive, Cambridge, UK.      Tel: +44 1223 245571
> Registered in England and Wales: Reg No 4422071.