Message-ID: <000b01c086af$d9195a10$b0e979c3@internet>
From: "Robert van der Boon" <rjvdboon AT europe DOT com>
To: <djgpp-workers AT delorie DOT com>
References: <Pine DOT SUN DOT 3 DOT 91 DOT 1010123121545 DOT 9288F-100000 AT is>
Subject: Re: patch.exe (fwd)
Date: Thu, 25 Jan 2001 10:17:58 +0100
MIME-Version: 1.0
Content-Type: text/plain;
	charset="iso-8859-1"
Content-Transfer-Encoding: 7bit
X-Priority: 3
X-MSMail-Priority: Normal
X-Mailer: Microsoft Outlook Express 5.50.4133.2400
X-MimeOLE: Produced By Microsoft MimeOLE V5.50.4133.2400
Reply-To: djgpp-workers AT delorie DOT com

> > _doserrno == 0xB7 (network segment error...)
> > So treating 0xB7 && fileexist the same as 0x2 && fileexist gives the
> > following
> > patch to _rename.c, and then it works:
> > !    && (r.x.ax == 5 || ((r.x.ax == 2 || r.x.ax==0xB7) &&
> > __file_exists(old))))
> Why do you think it is necessary to test if the old file exists?  Does
> W2K return 0xb7 even if the file does not exist?

I should think before typing, what I meant was:
 !    && (r.x.ax == 5 || (r.x.ax == 2 && __file_exists(old)) ||
(r.x.ax==0xB7 && __file_exists(new)))
It gives 0xB7 only if the new file exists..., but I didn't want to widen the
scope of the test, that's why the __file_exist(new) is there.
If the performance penalty of __file_exist() is too big, I think that test
can
go, and we'll treat 0xB7 exactly as 0x5.

> > Another thing I noticed was that the following text (or better: all
texts
> > without \n):
> > "Reversed (or previously applied) patch detected!  Assume -R? [n]"
> > don't show up _before_ you press a valid key, but only _after_ you
> > press y/n/space. It seems like fflush isn't working either
> Yes, someone posted reports to that effect in the past.  Feel like
> digging into this one as well? ;-)

I wouldn't have a clue where to start...

> > but it does work if I'm in GDB.
> GDB uses termios, which redirects all screen writes through Int 29h.
> So it looks like only DOS I/O is affected.

Yep.

> > "W2K is a pain in the lower posterior" (tm)
> Tell me about it.  Another report from the trenches seems to indicate
> that you cannot redirect stderr separately from stdout.  Could you try
> this with `redir' and see if that's really so?

Maybe later this week.

> Thanks a lot for your help.

You're welcome, always.

Bye now,
 Robert