RE: [MiNT] DATE/TIME cookies

["Julian Reschke" <reschke@medicaldataservice.de>]

> On Wed, 17 Feb 1999 16:39:54 +0100 (EET), Konrad Kokoszkiewicz wrote:
>
> KK>> OK, so what if FreeMiNT allocated two word blocks in a
> global memory and
> KK>> ´mirrored´ the internal datestamp and timestamp variables into those
> KK>> vars in global memory?
>
> KK>Two problems:
> KK>
> KK>1) (Julian already pointed it out) internal date and timestamp of MiNT
> KK>   must be in TOS format. Which won´t be the case, the internal
> KK>   implementation of timekeeping variables is currently changing.
>
> there will always be the code for converting the internal variables into
> GEMDOS format (for Tgetdate/Tgettime). This code could be called once
> per second to update the global variables DATE/TIME cookies would point
> to.
>
> KK>The real problem is that calling MiNT from interrupt (whether MiNT is
> KK>reentrant or not) would greatly increase system load
>
> I agree. So please tell me how to allocate 4 bytes in globally
> accessible memory (read-only is OK) and I´ll write the code for
> DATE/TIME cookies. Then you can say if it´s OK to integrate it or not.

It still escapes me why we are even discussing this.

a) You can easily write a corner clock application as accessory or MiNT
background process. No need for a TSR.

b) Even IF it is a TSR, there is absolutely NO reason to retrieve the system
time more than once a second.

Please do not try to fix something which is fine. Just use a better approach
for your utility.

Regards, jr