Index | Thread | Search

From:
YASUOKA Masahiko <yasuoka@openbsd.org>
Subject:
Re: games/nblood diff for stdio opaque
To:
tb@openbsd.org
Cc:
jsg@jsg.id.au, ports@openbsd.org
Date:
Tue, 15 Jul 2025 19:20:37 +0900

Download raw body.

Thread
On Tue, 15 Jul 2025 09:08:46 +0200
Theo Buehler <tb@openbsd.org> wrote:
> On Tue, Jul 15, 2025 at 02:18:19PM +0900, YASUOKA Masahiko wrote:
>> On Tue, 15 Jul 2025 11:50:43 +1000
>> Jonathan Gray <jsg@jsg.id.au> wrote:
>> > On Tue, Jul 15, 2025 at 10:23:00AM +0900, YASUOKA Masahiko wrote:
>> >> Hi,
>> >> 
>> >> A fix like the diff following is needed after stdio opaque diff.  The
>> >> diff can be applied in advance.
>> >> 
>> >> ok?
>> > 
>> > Shouldn't this also change REVISION?
>> 
>> Yes.  The diffs cause a functionally change to the program.
>> I updated the diff
>> 
> 
>> +@@ -872,9 +872,11 @@ void maybe_redirect_outputs(void)
>> +     if (fp)
>> +     {
>> +         setvbuf(fp, 0, _IONBF, 0);
>> +-        *stdout = *fp;
>> +-        *stderr = *fp;
>> +     }
>> ++
>> ++    fp = freopen("stdout.txt", "w", stderr);
>> ++    if (fp)
>> ++        setvbuf(fp, 0, _IONBF, 0);
> 
> I don't see a good way of solving this right now, but does this not
> leak a FILE *?
> 
> Maybe it's not a problem if the function is only called once.

The function is called just once from main().  So the problem is
acceptable.