X-Authentication-Warning: delorie.com: mail set sender to djgpp-bounces using -f
X-Recipient: djgpp AT delorie DOT com
Date: Sun, 28 Jun 2009 21:59:37 +0300
From: Eli Zaretskii <eliz AT gnu DOT org>
Subject: Re: Emacs 23.0.95 pretest is available
In-reply-to: <5a7eb0db-1b56-4f7a-811d-bb7abdaef0cc@b14g2000yqd.googlegroups.com>
X-012-Sender: halo1 AT inter DOT net DOT il
To: djgpp AT delorie DOT com
Message-id: <83k52w421y.fsf@gnu.org>
References: <83eit5zsit DOT fsf AT gnu DOT org> <5a7eb0db-1b56-4f7a-811d-bb7abdaef0cc AT b14g2000yqd DOT googlegroups DOT com>
Reply-To: djgpp AT delorie DOT com
Errors-To: nobody AT delorie DOT com
X-Mailing-List: djgpp AT delorie DOT com
X-Unsubscribes-To: listserv AT delorie DOT com
Precedence: bulk

> From: Rugxulo <rugxulo AT gmail DOT com>
> Date: Sat, 27 Jun 2009 19:41:56 -0700 (PDT)
> 
> "make install" took approx. 4 1/2 mins.

Is it good or bad?

If bad, what part took most of the time?

> However, on XP or FreeDOS, if using GCC 2.95.3 (oldy moldy which I'm
> fond of), it seems to hang at loadup.el (I think)

What do you see if you interrupt it at this stage (you will need to
run it under GDB for this to work).

> "C-x 5 1" etc. (frames) seem to be giving problems (FreeDOS or Vista).

Ouch!  Thanks for reporting this, patch to fix it below.

> Either way, Make (3.79.1 or 3.81) seemed to have problems
> finding test-distrib.c

Please show the signs of these problems.

> instead of the weird shell hack that is trying to use "e:/emacs"
> instead

Sorry, I don't follow: what weird hack, and why does it use e:/emacs?

> "C-x i" (can't find Info dir in FreeDOS w/ SFNs, but Vista w/ LFNs works
> okay).

I believe you mean "C-h i".

Do you have INFOPATH set on any of these machines?  Please look in
djgpp.env: it sets things up so that Emacs expects to find the Info
files that come with Emacs in %DJDIR%/gnu/emacs/info; if you installed
Emacs in some other directory, you will need to set INFOPATH to point
to its `info' subdirectory.

If all of this does not help, please show your value of the variable
Info-directory-list after you type "C-h i".

> I also tried with 3.4.4 for laughs just in case some of it was
> GCC version specific, but no apparent differences (well, no silly
> alignment warnings, 2.95.3 says "... too much, defaulting to 4 " a
> lot) beyond a slightly bigger .EXE (well, and the above loadup issue
> which 3.4.4 doesn't have, but it doesn't have 8-byte alignment either,
> so the above download is using GCC 4.4.0).

I use GCC 3.4.4 and don't have any alignment issues.

Thanks again for testing.  Here's the patch for the problem with
"C-x 5 1" (already in the Emacs repository):


2009-06-28  Eli Zaretskii  <eliz AT gnu DOT org>

	* term.c (create_tty_output) [MSDOS]: #ifdef away.
	(tty_free_frame_resources) [MSDOS]: Add a DOS-specific version.


--- src/term.c~	2009-06-06 23:42:16 +0300
+++ src/term.c	2009-06-28 21:34:54 +0300
@@ -3185,6 +3185,7 @@ DEFUN ("gpm-mouse-stop", Fgpm_mouse_stop
 			    Initialization
  ***********************************************************************/
 
+#ifndef MSDOS
 /* Initialize the tty-dependent part of frame F.  The frame must
    already have its device initialized. */
 
@@ -3218,6 +3219,20 @@ tty_free_frame_resources (struct frame *
   xfree (f->output_data.tty);
 }
 
+#else  /* MSDOS */
+
+/* Delete frame F's face cache. */
+
+static void
+tty_free_frame_resources (struct frame *f)
+{
+  if (! FRAME_TERMCAP_P (f) && ! FRAME_MSDOS_P (f))
+    abort ();
+
+  if (FRAME_FACE_CACHE (f))
+    free_frame_faces (f);
+}
+#endif	/* MSDOS */
 
 /* Reset the hooks in TERMINAL.  */