Re: [Freemint-discuss] Mint handling of file timestamps

[Helmut K. <hel...@is...>]

Am 10.11.2017, 03:58 Uhr, schrieb Thorsten Otto:

>> Shall the file-time be stored in utc or local time on disk?
>
> Since it is a plain tos/msdos filesystem (no vfat), it can only be  
> stored in
> local time.

So if you send the image to someone in China he will see your local times,  
but for a TOS-FS there might be no other way.

>
> I think the only possible solution is to add some flag to Dopendir, to  
> tell
> Dxreaddir that it should return the timestamps as UTC so the application  
> can
> do the calculation. Of course, that would also require changes to the
> applications using it directly (like teradesk), and mintlib for all  
> others.
> For other cases, like stat() and Fdatime(), there are already  
> alternatives
> that should work, like Fstat64() and fcntl(FUTIME_UTC).

Yes, store the times in UTC. Maybe mintlib can set a flag to tell the  
kernel about it's preference, and old apps would still work as before.

As far as I see, tools like ls expect utc and call localtime to get the  
tm-content in local-time representation.

So it should be like this:

TOS-FS: file-times are local-time
Others: file-times are utc.

I see on getstat64,xfs_xdd.c:

     /* no native UTC extension
      * -> convert to unix UTC
      */
     stat->atime.high_time = 0;
     stat->atime.time = unixtime (xattr.atime, xattr.adate) + timezone;
     stat->atime.nanoseconds = 0;

That would mean that MiNT always tries to returns local time? And timezone  
is a constant not taking account of the time-offset of the actual time?