mac-emacs-users

[Mac-emacs-users] Mac OS X Background Color Bug Fix

[Andrew C. <ak...@sh...>]

Hi,

Just checked in a bug fix that corrects a problem with Emacs on Mac OS
X not properly using the background color set in default-frame-alist.

If the background color you set only shows for areas with text and a
white background is displayed for areas without text, this patch
should be the solution.

This fix was sent by Steven Tamm.  Amos at UT Dallas helped test it.

Thank you both!

Andrew.

[Mac-emacs-users] Mac OS Command Key as <ALT> Patch

[Andrew C. <ak...@sh...>]

Hi,

I have added a small change to src/macterm.c that will allow Emacs on
the Mac OS to recognize the Mac command key as <ALT> when the Mac
option key is used as <META>.  This is useful for people who want to
map cut, copy, and paste to the familiar command-X, command-C, and
command-V, for example.

This will not affect people using the Mac command key as <META>.

Andrew.

Re: [Mac-emacs-users] command key patch in 21.2

[Rodney S. <rsp...@mc...>]

Andrew:

That's the one.  I figured that if there is an addressable key, that 
people will want to use it.
I know I do :o)  It's nice to be able to use command-x, command-c, 
command-v and command-q for
cut, copy, paste and quit with all applications.

Thanks, Rodney

On Monday, June 17, 2002, at 07:51 PM, Andrew Choi wrote:

>> I also have a question about a patch.  When I compiled 21.2, I
>> noticed that the definition of the command key was not there.  Of
>> course, it wasn't hard to fix that (well, not hard since Andrew told
>> me how to do it :o).  Was that because I downloaded 21.2 before it
>> got there?  If not, are there future plans to include it?  Is there
>> some reason that it was not included?  Thanks.
>
> Do you mean the code that lets the command key function as ALT when
> the option key is used as META?
>
> If so I can easily add it back.
>

Re: [Mac-emacs-users] command key patch in 21.2

[Andrew C. <ak...@sh...>]

> I also have a question about a patch.  When I compiled 21.2, I
> noticed that the definition of the command key was not there.  Of
> course, it wasn't hard to fix that (well, not hard since Andrew told
> me how to do it :o).  Was that because I downloaded 21.2 before it
> got there?  If not, are there future plans to include it?  Is there
> some reason that it was not included?  Thanks.

Do you mean the code that lets the command key function as ALT when
the option key is used as META?

If so I can easily add it back.

Re: [Mac-emacs-users] "browse-url" support for Macintosh

[Andrew C. <ak...@sh...>]

> * pi...@cs... [2002-06-13 16:27:07]
> > There is no reason to patch browse-urk for this. The following also works:
> >
> > (if (eq system-type 'darwin)
> >     (setq browse-url-generic-program
> > 	  "openurl"))
> >
> > where openurl is the shellscript:
> >
> > #!/bin/sh
> > osascript -e 'open location "'$*'"'
> 
> Yes, this is similar to the first version that I had (mine called
> osascript directly from browse-url).
> 
> > I think the do-applescript solution requires a patched emacs, doesn't it?
> 
> Yes, though that patch isn't something specific to browse-url - it
> would be a good thing to do generally.  The advantage of the
> do-applescript version is that it doesn't require the creation of
> extra processes.  It should be a little faster, and will consume less
> resources (forking emacs to exec the shell and/or osascript seems like
> something best avoided where possible).
> 
> dme.

Let's test it and add the modification to browse-url.el if it works
well.  It would be nice to have it already setup with some default.  I
wonder if we need to handle `mailto:' specially however.

Re: [Mac-emacs-users] getaddrinfo() related problems

[Andrew C. <ak...@sh...>]

> In using Emacs with Gnus, I see frequent failures to look up
> hostnames.  In particular, news.gmane.org almost always fails.
> Reading in other places that the getaddrinfo() implementation in
> MacOSX is poor, I applied the following patch to avoid using it.
> Things seem much better now - I've not had a single failure since.
> 
> Does anyone else see similar problems ?
> 
> Index: darwin.h
> ===================================================================
> RCS file: /cvsroot/emacs/emacs/src/s/darwin.h,v
> retrieving revision 1.1
> diff -u -r1.1 darwin.h
> --- darwin.h	26 Apr 2002 23:39:06 -0000	1.1
> +++ darwin.h	13 Jun 2002 07:04:36 -0000
> @@ -1,5 +1,5 @@
>  /* System description header file for Darwin (Mac OS X).
> -   Copyright (C) 2001 Free Software Foundation, Inc.
> +   Copyright (C) 2001, 2002 Free Software Foundation, Inc.
>  
>  This file is part of GNU Emacs.
>  
> @@ -276,6 +276,8 @@
>  /* Do not define matherr in floatfns.c.  */
>  #define NO_MATHERR
>  
> +/* Up to 10.1.5 getaddrinfo() is suspect (dm...@dm...) */
> +#undef HAVE_GETADDRINFO
>  
>  /* This prevents a compilation error in xfaces.c: struct kboard * is
>     used in a function protocol the first time this type appears in the

Thank you for this fix.  It should be harmless since it just tells
Emacs to use its own emulation of getaddrinfo.  Can people please try
this out?  Does anyone experience a similar problem?  What are the
symptoms when the look up of a hostname fails?

Re: [Mac-emacs-users] "browse-url" support for Macintosh

[Andrew C. <ak...@sh...>]

> Here is a patch to browse-url.el which adds support for a MacOS
> browser function.  It uses Applescript to ask the default browser to
> open URLs.
> 
> The previous patch to persuade do-applescript to work is required.
> 
> This is only tested under MacOSX.  It seems as though it should work
> under previous versions of MacOS, but I don't have them to test.
> What is the value of "system-type" there ?

Thank you for the patch.  Can people please try this out before I
check it in?

I also need a solution for Mac OS 8/9.  This should not be hard.  I'll
look at it when I can.

Re: [Mac-emacs-users] Re: do-applescript

[Andrew C. <ak...@sh...>]

> Thanks for the pointer.  Using the patch below (which is based on
> that in the thread you indicated), do-applescript now works for me
> just fine.
> 
> Could this be committed to CVS ?

Thank you for the patch.  This has now been checked in.

Re: [Mac-emacs-users] Swapping Caps and Ctrl from within Emacs

[Gregory P. K. <gre...@lm...>]

The latest version of uControl[http://www.gnufoo.org/macosx/] has been
released. This version should work on the newer TiBooks (they had a
newer kernel build), and on 10.1.5 (The previous version caused a nasty
kernel panic).

This is a great little utility that allows us TiBook owners to enjoy our
modern Emacs lifestyle.

Share and Enjoy!

Gregory

On Tue, 2002-06-04 at 14:39, Matthias Scheidegger wrote:
    Can I swap Caps Lock and Ctrl from within Emacs somehow? I've seen a trick to
    tweak the behaviour of the Option key, so I hope there's something similar for
    Caps & Ctrl.
    I can't swap them outside Emacs since I've got a new TiBook. At least the USB
    keyboard tweak and the uControl library don't work on it (no USB keyboard / too
    recent build of OS X). I'd be grateful for any pointers to other possibilities
    of course.
    

[Mac-emacs-users] Mac OS X do-applescript Bug Fix

[Andrew C. <ak...@sh...>]

Hi,

I have checked in a change to src/mac.c that will enable the built-in
function do-applescript to work correctly on the Mac OS X build.

Andrew.

[Mac-emacs-users] Japanese Input on Mac OS X Bug Fix

[Andrew C. <ak...@sh...>]

Hi,

Toshikazu Onishi provided a fix for crashes one encounters when using
Mac OS Japanese Input with Emacs on Mac OS X.

The fix simply involves changing a call to FrontWindow in
XTread_socket in the file src/macterm.c to FrontNonFloatingWindow.  If
you are experiencing similar problems, please try it.  I have checked
the change into the CVS trunk.  It should not affect the use of other
Keyboard layouts.

Andrew.

[Mac-emacs-users] Mac OS X Font Bug Fix and Jaguar Patch Committed

[Andrew C. <ak...@sh...>]

Hi,

I have just committed to the trunk of the CVS the patch I posted
earlier that fixes the problem with the use of bold and italic fonts
on Mac OS X.  Now you should be able to use the italic, bold, or
bold-italic variant of any font installed on your system.  Some fonts
will look nicer than others.

I have also included a few minor changes that allows Emacs to be built
on Mac OS X 10.2 preview.  You should still be able to build on 10.1
of course.

Andrew.

Re: [Mac-emacs-users] Background Color problems in New frames

[Andrew C. <ak...@sh...>]

> Steve, did it for you?  I just added that and did a make; sudo make
> install; sudo ditto mac/Emacs.app /Applications/Emacs.app.  Background 
> color in new frames still gets out of alignment.  I don't need to
> do a make clean and recompile everything, do I?  I was testing it
> with same setup, i.e.:
> 
> (cond
>  ((eq window-system 'mac)
>   (setq default-frame-alist
>         '(;;(font . "-*-andale mono-*-*-*-*-10-*-*-*-*-*-mac-roman")
>           ;;(font . "-*-courier-*-*-*-*-10-*-*-*-*-*-mac-roman")
>           (font . "-*-monaco-*-*-*-*-10-*-*-*-*-*-mac-roman")
>           (mouse-color . "midnightblue")
>           (foreground-color . "midnightblue")
>           (background-color . "antiquewhite1")))))
> 
> I was also fiddling with the fonts.  Playing with the font selector
> for the Terminal app it *seems* like I should be able to try that
> "andale mono" one, but Emacs sure doesn't like it.  This is from a
> CVS update (from HEAD) last night.  I *thought* the latest font
> stuff was added to macterm.c, but not sure.  
> 
> -- 
> Amos

He sent a fix a while earlier.  I'll try to post that to the list
later today.

Andrew.

Re: [Mac-emacs-users] Background Color problems in New frames

[Amos <amo...@ut...>]

>>>>> On Fri, 14 Jun 2002 21:10:04 -0600,
>>>>> Andrew Choi <ak...@sh...> (ac) writes:

ac> Hi Steven,

ac> Please try adding the line

ac>   XClearWindow (FRAME_MAC_DISPLAY (f), FRAME_MAC_WINDOW (f));

ac> just before the line

ac>   expose_frame (f, 0, 0, 0, 0);

ac> in the function do_window_update in the file macterm.c and see if that
ac> solves the problem.

Steve, did it for you?  I just added that and did a make; sudo make
install; sudo ditto mac/Emacs.app /Applications/Emacs.app.  Background 
color in new frames still gets out of alignment.  I don't need to
do a make clean and recompile everything, do I?  I was testing it
with same setup, i.e.:

(cond
 ((eq window-system 'mac)
  (setq default-frame-alist
        '(;;(font . "-*-andale mono-*-*-*-*-10-*-*-*-*-*-mac-roman")
          ;;(font . "-*-courier-*-*-*-*-10-*-*-*-*-*-mac-roman")
          (font . "-*-monaco-*-*-*-*-10-*-*-*-*-*-mac-roman")
          (mouse-color . "midnightblue")
          (foreground-color . "midnightblue")
          (background-color . "antiquewhite1")))))

I was also fiddling with the fonts.  Playing with the font selector
for the Terminal app it *seems* like I should be able to try that
"andale mono" one, but Emacs sure doesn't like it.  This is from a
CVS update (from HEAD) last night.  I *thought* the latest font
stuff was added to macterm.c, but not sure.  

-- 
Amos

Re: [Mac-emacs-users] Background Color problems in New frames

[Andrew C. <ak...@sh...>]

> I've only been using Emacs 21 or Mac OS X for a little while, but I'm 
> pretty far along in hacking it.  However, I'm having some trouble 
> keeping my background color correct.  The behavior I'm seeing is that 
> the background color for areas without text is white, regardless of what 
> I specify in the default.  However the background color for areas with 
> text is what I specify.  I'm not sure whether it's emacs 21 or specific 
> to the mac.  It seems to only happen in newly created frames, and it 
> goes away if I specifically call (set-background-color)
> 
> Here's the code I'm using to specify the frame parameters:
> 
> (cond ((eq window-system 'mac)
>         (setq default-frame-alist
> 	     '((mouse-color . "midnightblue")
> 	       (foreground-color . "midnightblue")
> 	       (background-color . "antiquewhite1")))
>         ))
> 
> Thanks for any help,
> -Steven

Hi Steven,

Please try adding the line

  XClearWindow (FRAME_MAC_DISPLAY (f), FRAME_MAC_WINDOW (f));

just before the line

  expose_frame (f, 0, 0, 0, 0);

in the function do_window_update in the file macterm.c and see if that
solves the problem.

Andrew.

[Mac-emacs-users] Background Color problems in New frames

[Steven T. <ste...@ma...>]

I've only been using Emacs 21 or Mac OS X for a little while, but I'm 
pretty far along in hacking it.  However, I'm having some trouble 
keeping my background color correct.  The behavior I'm seeing is that 
the background color for areas without text is white, regardless of what 
I specify in the default.  However the background color for areas with 
text is what I specify.  I'm not sure whether it's emacs 21 or specific 
to the mac.  It seems to only happen in newly created frames, and it 
goes away if I specifically call (set-background-color)

Here's the code I'm using to specify the frame parameters:

(cond ((eq window-system 'mac)
        (setq default-frame-alist
	     '((mouse-color . "midnightblue")
	       (foreground-color . "midnightblue")
	       (background-color . "antiquewhite1")))
        ))

Thanks for any help,
-Steven

[Mac-emacs-users] command key patch in 21.2

[Rodney S. <rsp...@po...>]

I also have a question about a patch.  When I compiled 21.2, I noticed that the
definition of the command key was not there.  Of course, it wasn't hard to fix 
that (well, not hard since Andrew told me how to do it :o).  Was that because I 
downloaded 21.2 before it got there?  If not, are there future plans to include 
it?  Is there some reason that it was not included?  Thanks.

Rodney Sparapani              Medical College of Wisconsin
Sr. Biostatistician           Patient Care & Outcomes Research (PCOR)
rsp...@mc...              http://www.mcw.edu/pcor
Was 'Name That Tune' rigged?  WWLD -- What Would Lombardi Do

Re: [Mac-emacs-users] "browse-url" support for Macintosh

[<dm...@dm...>]

* pi...@cs... [2002-06-13 16:27:07]
> There is no reason to patch browse-urk for this. The following also works:
>
> (if (eq system-type 'darwin)
>     (setq browse-url-generic-program
> 	  "openurl"))
>
> where openurl is the shellscript:
>
> #!/bin/sh
> osascript -e 'open location "'$*'"'

Yes, this is similar to the first version that I had (mine called
osascript directly from browse-url).

> I think the do-applescript solution requires a patched emacs, doesn't it?

Yes, though that patch isn't something specific to browse-url - it
would be a good thing to do generally.  The advantage of the
do-applescript version is that it doesn't require the creation of
extra processes.  It should be a little faster, and will consume less
resources (forking emacs to exec the shell and/or osascript seems like
something best avoided where possible).

dme.

Re: [Mac-emacs-users] "browse-url" support for Macintosh

[Piet v. O. <pi...@cs...>]

>>>>> dm...@dm... (d) writes:

d> Here is a patch to browse-url.el which adds support for a MacOS
d> browser function.  It uses Applescript to ask the default browser to
d> open URLs.

There is no reason to patch browse-urk for this. The following also works:

(if (eq system-type 'darwin)
    (setq browse-url-generic-program
	  "openurl"))

where openurl is the shellscript:

#!/bin/sh
osascript -e 'open location "'$*'"'

I think the do-applescript solution requires a patched emacs, doesn't it?
-- 
Piet van Oostrum <pi...@cs...>
URL: http://www.cs.uu.nl/~piet [PGP]
Private email: P.v...@hc...

[Mac-emacs-users] getaddrinfo() related problems

[<dm...@dm...>]

In using Emacs with Gnus, I see frequent failures to look up
hostnames.  In particular, news.gmane.org almost always fails.
Reading in other places that the getaddrinfo() implementation in
MacOSX is poor, I applied the following patch to avoid using it.
Things seem much better now - I've not had a single failure since.

Does anyone else see similar problems ?

Index: darwin.h
===================================================================
RCS file: /cvsroot/emacs/emacs/src/s/darwin.h,v
retrieving revision 1.1
diff -u -r1.1 darwin.h
--- darwin.h	26 Apr 2002 23:39:06 -0000	1.1
+++ darwin.h	13 Jun 2002 07:04:36 -0000
@@ -1,5 +1,5 @@
 /* System description header file for Darwin (Mac OS X).
-   Copyright (C) 2001 Free Software Foundation, Inc.
+   Copyright (C) 2001, 2002 Free Software Foundation, Inc.
 
 This file is part of GNU Emacs.
 
@@ -276,6 +276,8 @@
 /* Do not define matherr in floatfns.c.  */
 #define NO_MATHERR
 
+/* Up to 10.1.5 getaddrinfo() is suspect (dm...@dm...) */
+#undef HAVE_GETADDRINFO
 
 /* This prevents a compilation error in xfaces.c: struct kboard * is
    used in a function protocol the first time this type appears in the

[Mac-emacs-users] "browse-url" support for Macintosh

[<dm...@dm...>]

Here is a patch to browse-url.el which adds support for a MacOS
browser function.  It uses Applescript to ask the default browser to
open URLs.

The previous patch to persuade do-applescript to work is required.

This is only tested under MacOSX.  It seems as though it should work
under previous versions of MacOS, but I don't have them to test.
What is the value of "system-type" there ?

Index: browse-url.el
===================================================================
RCS file: /cvsroot/emacs/emacs/lisp/net/browse-url.el,v
retrieving revision 1.21
diff -u -r1.21 browse-url.el
--- browse-url.el	18 Mar 2002 04:24:09 -0000	1.21
+++ browse-url.el	13 Jun 2002 06:57:46 -0000
@@ -1,6 +1,6 @@
 ;;; browse-url.el --- pass a URL to a WWW browser
 
-;; Copyright (C) 1995, 96, 97, 98, 99, 2000, 2001
+;; Copyright (C) 1995, 96, 97, 98, 99, 2000, 2001, 2002
 ;;   Free Software Foundation, Inc.
 
 ;; Author: Denis Howe <db...@do...>
@@ -52,6 +52,7 @@
 ;; browse-url-default-windows-browser MS-Windows browser
 ;; browse-url-gnome-moz               GNOME interface to Mozilla
 ;; browse-url-kde                     KDE konqueror (kfm)
+;; browse-url-default-macintosh-browser	Macintosh browser
 
 ;; [A version of the Netscape browser is now free software
 ;; <URL:http://www.mozilla.org/>, albeit not GPLed, so it is
@@ -227,9 +228,13 @@
 
 ;;;###autoload
 (defcustom browse-url-browser-function
-  (if (memq system-type '(windows-nt ms-dos))
-      'browse-url-default-windows-browser
-    'browse-url-default-browser)
+  (cond
+   ((memq system-type '(windows-nt ms-dos))
+    'browse-url-default-windows-browser)
+   ((eq system-type 'darwin)
+    'browse-url-default-macintosh-browser)
+   (t
+    'browse-url-default-browser))
   "*Function to display the current buffer in a WWW browser.
 This is used by the `browse-url-at-point', `browse-url-at-mouse', and
 `browse-url-of-file' commands.
@@ -262,6 +267,8 @@
 			 :value browse-url-default-windows-browser)
 	  (function-item :tag "GNOME invoking Mozilla"
 			 :value browse-url-gnome-moz)
+	  (function-item :tag "Default Macintosh Browser"
+			 :value browse-url-default-macintosh-browser)
 	  (function-item :tag "Default browser"
 			 :value browse-url-default-browser)
 	  (function :tag "Your own function")
@@ -1260,6 +1267,14 @@
   (message "Sending URL to KDE...")
   (apply #'start-process `(,(concat "KDE" url) nil ,browse-url-kde-program
 			   ,@browse-url-kde-args ,url)))
+
+;;;###autoload
+(defun browse-url-default-macintosh-browser (url &optional new-window)
+  "Ask the default browser to load a URL.
+Default to the URL around or before point."
+  (interactive (browse-url-interactive-arg "Macintosh URL: "))
+  (message "Sending URL...")
+  (do-applescript (concat "open location \"" url "\"")))
 
 (provide 'browse-url)
 

[Mac-emacs-users] Re: do-applescript

[<dm...@dm...>]

* ak...@sh... [2002-06-11 21:10:15]
> [...I couldn't get "do-applescript" to work...]
> Please take a look at this message in the mailing list archive:
>   http://sourceforge.net/mailarchive/forum.php?thread_id=257726&forum_id=5271
> I don't think the patch described in it has been tested and included
> into the CVS code yet.

Thanks for the pointer.  Using the patch below (which is based on
that in the thread you indicated), do-applescript now works for me
just fine.

Could this be committed to CVS ?

> You can probably also use the Lisp function call-process with the
> osascript command.

This also worked, though I preferred to use the internal function.

Index: mac.c
===================================================================
RCS file: /cvsroot/emacs/emacs/src/mac.c,v
retrieving revision 1.3
diff -u -r1.3 mac.c
--- mac.c	6 May 2002 06:48:57 -0000	1.3
+++ mac.c	13 Jun 2002 06:48:21 -0000
@@ -2443,6 +2443,9 @@
 
   *result = 0;
 
+  if (!as_scripting_component)
+    initialize_applescript();
+
   error = AECreateDesc (typeChar, script, strlen(script), &script_desc);
   if (error)
     return error;
@@ -2499,10 +2502,10 @@
         }
       HUnlock (result_desc.dataHandle);
 #endif /* not TARGET_API_MAC_CARBON */
+      AEDisposeDesc (&result_desc);    
     }
 
   AEDisposeDesc (&script_desc);
-  AEDisposeDesc (&result_desc);    
 
   return osaerror;
 }
@@ -2527,7 +2530,7 @@
   if (status)
     {
       if (!result)
-        error ("AppleScript error %ld", status);
+        error ("AppleScript error %d", status);
       else
         {
           /* Unfortunately only OSADoScript in do_applescript knows how

Re: [Mac-emacs-users] do-applescript

[Andrew C. <ak...@sh...>]

> I'm using current CVS emacs built on 10.1.5 with the current (beta ?)
> development kit.  I've applied the patch to fix boxed/bold/italic
> chars.
> 
> Is there any guidance on using the "do-applescript" function ?
> 
> I've tried passing a string which is an applescript script, the name
> of a file which contains a script, ...  These all generate errors.
> 
> I changed the call to "error" so that it no longer complains about %ld
> as a format string, but now I just see large negative numbers.
> 
> Any help would be appreciated.

Please take a look at this message in the mailing list archive:

  http://sourceforge.net/mailarchive/forum.php?thread_id=257726&forum_id=5271

I don't think the patch described in it has been tested and included
into the CVS code yet.

You can probably also use the Lisp function call-process with the
osascript command.

Andrew.

[Mac-emacs-users] do-applescript

[<dm...@dm...>]

I'm using current CVS emacs built on 10.1.5 with the current (beta ?)
development kit.  I've applied the patch to fix boxed/bold/italic
chars.

Is there any guidance on using the "do-applescript" function ?

I've tried passing a string which is an applescript script, the name
of a file which contains a script, ...  These all generate errors.

I changed the call to "error" so that it no longer complains about %ld
as a format string, but now I just see large negative numbers.

Any help would be appreciated.

Re: [Mac-emacs-users] Running Ruby interpreter in subshell (OS X)

[Piet v. O. <pi...@cs...>]

>>>>> Brian Marick <ma...@te...> (BM) writes:

BM> I'm running "GNU Emacs 21.1.30.3 (powerpc-apple-darwin5.2.2) of 2002-02-10 on
BM> localhost". I'm pretty sure I got it as the precompiled binary from here:
BM> <http://www.cse.cuhk.edu.hk/~akochoi/emacs/21.1/emacs-21.1-mac-bin.smi.bin>

BM> When I start emacs, it connects to the network. What's it doing? Have I
BM> installed something I oughtn't have?

It could be that the ruby interpreter expects to be connected through a
pty. The emacs 21.1.30 image didn't have pty support.
-- 
Piet van Oostrum <pi...@cs...>
URL: http://www.cs.uu.nl/~piet [PGP]
Private email: P.v...@hc...