the std.c.time.sleep(x) method argument is actually in *seconds*; this is
really not very useful <g>

Can we get it changed to represent ms instead? That is; use SleepEx() rather
than Sleep() on Win32 ...

(I know many of these functions don't guarantee granularity below around 20ms, but that's much better than a granularity of 1000ms :-)

- Kris

Kris wrote:

> the std.c.time.sleep(x) method argument is actually in *seconds*; this is
> really not very useful <g>
> 
> Can we get it changed to represent ms instead? That is; use SleepEx() rather
> than Sleep() on Win32 ...
<snip>

Better for backward and forward compatibility would be to keep it representing seconds, and have it take a float rather than an int.

That way:
- there'd be no need to change every existing program so that it doesn't run 1000 times too fast
- when an API function comes along that takes microseconds, it'll fit in neatly with Phobos as it stands.  At that point, the millisecond'll be a rather arbitrary choice of unit.  The second, OTOH, is the basic SI unit of time, as well as being a unit in which most people are likely to think, and so remains a reasonable choice.

Stewart.

-- 
My e-mail is valid but not my primary mailbox, aside from its being the unfortunate victim of intensive mail-bombing at the moment.  Please keep replies on the 'group where everyone may benefit.

Stewart Gordon wrote:

> Kris wrote:
> 
>> the std.c.time.sleep(x) method argument is actually in *seconds*; this is
>> really not very useful <g>
<snip>

Oops, I overlooked a rather basic aspect!

It is in std.c.  This means that it is an interface to the C RTL.  So unless the RTL changes, std.c.time.sleep isn't going to change.

Still, we can always invent a new function with the semantics I've suggested....

Stewart.

-- 
My e-mail is valid but not my primary mailbox, aside from its being the unfortunate victim of intensive mail-bombing at the moment.  Please keep replies on the 'group where everyone may benefit.

Wait! I take it all back. If I'd bothered to read a bit further I'd have
seen the unix-style msleep() and usleep().

Sorry :-(


I marked it as a buglet since I imagine it was actually *meant* to be ms ... I've never seen a runtime library sleep() function that doesn't get any finer than 1000ms intervals. It's almost worthless as it is <g>  Naturally, there's a first time for everything.

The fact that it's still that way may indicate that no-one has really used it yet? It's certainly not used in any of the phobos code.

The windows.d has this instead: export void Sleep(DWORD dwMilliseconds);
But that's not portable.

- Kris


"Stewart Gordon" <smjg_1998@yahoo.com> wrote in message news:c78hoc$kmg$1@digitaldaemon.com...
> Kris wrote:
>
> > the std.c.time.sleep(x) method argument is actually in *seconds*; this
is
> > really not very useful <g>
> >
> > Can we get it changed to represent ms instead? That is; use SleepEx()
rather
> > than Sleep() on Win32 ...
> <snip>
>
> Better for backward and forward compatibility would be to keep it representing seconds, and have it take a float rather than an int.
>
> That way:
> - there'd be no need to change every existing program so that it doesn't
> run 1000 times too fast
> - when an API function comes along that takes microseconds, it'll fit in
> neatly with Phobos as it stands.  At that point, the millisecond'll be a
> rather arbitrary choice of unit.  The second, OTOH, is the basic SI unit
> of time, as well as being a unit in which most people are likely to
> think, and so remains a reasonable choice.
>
> Stewart.
>
> --
> My e-mail is valid but not my primary mailbox, aside from its being the unfortunate victim of intensive mail-bombing at the moment.  Please keep replies on the 'group where everyone may benefit.

"Kris" <someidiot@earthlink.dot.dot.dot.net> wrote in message news:c78mhk$so4$1@digitaldaemon.com...
> Wait! I take it all back. If I'd bothered to read a bit further I'd have
> seen the unix-style msleep() and usleep().
>
> Sorry :-(
>
>
> I marked it as a buglet since I imagine it was actually *meant* to be ms
...
> I've never seen a runtime library sleep() function that doesn't get any finer than 1000ms intervals. It's almost worthless as it is <g>
Naturally,
> there's a first time for everything.
>
> The fact that it's still that way may indicate that no-one has really used it yet? It's certainly not used in any of the phobos code.
>

Ive used msleep.

I think these functions taken from C arent they?

Phill

"Phill" <phill@pacific.net.au> wrote in message news:c7d44c$1kl0$1@digitaldaemon.com...
> Ive used msleep.
>
> I think these functions taken from C arent they?

They're not taken from C, they hook up directly to the C versions in the C runtime library.

"Walter" <newshound@digitalmars.com> wrote in message news:c7fb0q$25cg$1@digitaldaemon.com...
>
> "Phill" <phill@pacific.net.au> wrote in message news:c7d44c$1kl0$1@digitaldaemon.com...
> > Ive used msleep.
> >
> > I think these functions taken from C arent they?
>
> They're not taken from C, they hook up directly to the C versions in the C runtime library.
>
Thats what I meant sorry, it was just a poor choice of words on my part :o(

Phill wrote:
> 
> Ive used msleep.
> 
> I think these functions taken from C arent they?

They're optional portions of POSIX.


Sean

Sean Kelly wrote:
> Phill wrote:
>>
>> Ive used msleep.
>>
>> I think these functions taken from C arent they?
> 
> They're optional portions of POSIX.

Oops... and I thought I was replying to a recent post :-)


Sean