Date:	Mon, 17 Apr 2000 11:53:45 +0200 (IST)
From:	Eli Zaretskii <eliz AT is DOT elta DOT co DOT il>
X-Sender:	eliz AT is
To:	djgpp-workers AT delorie DOT com
Subject:	Re: timezone files: the solution
In-Reply-To:	<Pine.A41.4.05.10004171025310.67430-100000@ieva05.lanet.lv>
Message-ID:	<Pine.SUN.3.91.1000417114916.14867Z-100000@is>
MIME-Version:	1.0
Reply-To:	djgpp-workers AT delorie DOT com
Errors-To:	nobody AT delorie DOT com
X-Mailing-List:	djgpp-workers AT delorie DOT com
X-Unsubscribes-To:	listserv AT delorie DOT com

On Mon, 17 Apr 2000, Andris Pavenis wrote:

> > > I'm getting broken djtzn203.zip also when building it under Linux (on 
> > > ix86 of course)
> > 
> > x86 is not the issue here (the timezone files are
> > endian-independent).  The question is: how does Linux define time_t?
> > If it's a signed type, you will see the same problems as on Irix,
> > because the timezone files are generated by zic compiled for the host,
> > so it uses host's time_t.
> 
> Tested: glibc-2.1.3 defines time_t to 'long int'

This explains why you get broken timezone files.

> > > C:\DJGPP\TEST\tz>tztest
> > > Fri, 14 Apr 2000 17:58:05 +7200
> > > Fri, 14 Apr 2000 17:58:05 EET
> > > 
> > > Which seems to be Ok.
> > 
> > Thanks for testing.
> > 
> 
> One more comment:
> 
> In Linux (glibc-2.1.3) I got following output from the same test program:
> 
> Mon, 17 Apr 2000 10:32:14 +0200
> Mon, 17 Apr 2000 10:32:14 EET
> 
> which is rather different from ones I'm getting with DJGPP. Which one 
> is correct?

The first result is correct, since the timezone files were produced with 
DJGPP-compiled zic.

Using timezone files produced by DJGPP with zoc compiled with glibc 
will produce incorrect results, for the same reasons that using timezone 
files produced by glibc-linked zic with DJGPP will work incorrectly.

The time_t type is incompatible in these two libraries.

If you want compatible timezone files, add -s to the invocation of 
host-zic in zoneinfo/src/makefile and rebuild djdev.

- Raw text -