Date: Sun, 21 Oct 2001 13:35:58 +0200 (IST)
From: Eli Zaretskii <eliz AT is DOT elta DOT co DOT il>
X-Sender: eliz AT is
To: Richard Dawe <rich AT phekda DOT freeserve DOT co DOT uk>
cc: DJGPP workers <djgpp-workers AT delorie DOT com>
Subject: Re: Some build fixes for the test suite
In-Reply-To: <3BD2AAB2.CAF5D96B@phekda.freeserve.co.uk>
Message-ID: <Pine.SUN.3.91.1011021133036.22550C-100000@is>
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
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
Precedence: bulk


On Sun, 21 Oct 2001, Richard Dawe wrote:

> Some of the tests do not currently build with gcc 2.95.3. Below is a diff
> that fixes some of the tests.

Thanks; I have some comments.

> The only objection I can see is that the putpath test now strdup's a bunch
> of strings, but does not free them when the test exits. But these strings
> are freed automatically on exit, right?

All memory allocated by a program is freed when the program exits.

>       if (f->_cnt > last_count)
>       {
> !       printf("Count: %5d  Fill: %5d  Bufsiz: %5d  SoFar: %5d\n",
>              f->_cnt, f->_fillsize, f->_bufsiz, so_far);
>       }
>       last_count = f->_cnt;
> --- 12,18 ----
>     {
>       if (f->_cnt > last_count)
>       {
> !       printf("Count: %5d  Fill: %5ld  Bufsiz: %5ld  SoFar: %5d\n",
>              f->_cnt, f->_fillsize, f->_bufsiz, so_far);

If we are nitpicking, lets pick all the nits ;-)  Let's cast the size_t 
members to long (actually, unsigned long sounds better).  That way, we 
don't need to do anything if some day the size_t typedef changes.

> *** tests/libc/ansi/time/strftime.c     1997/11/02 15:23:56     1.1
> --- tests/libc/ansi/time/strftime.c     2001/10/14 21:01:37
> *************** compare (const char *fmt, const struct t
> *** 28,33 ****
> --- 28,35 ----
>   int
>   main (void)
>   {
> + #undef TZ
> +   char TZ[] = "TZ=GMT0";

I don't understand the reason for this one.  Can you explain?

> !   putenv(strdup("DJDIR=c:/djgpp2"));
> !   putenv(strdup("HOME=c:\\home"));

Yuck!  Is this because of the "const char *" vs "char *" nuisance?  If 
so, won't simple cast or `unconst' do?