Re: [dwm] xprop patch

2008-12-13 Thread Frederic Chardon 
Hello,

It is now useless to close stdin:

--- orig/dwm.c  2008-12-13 18:05:32.0 +0100
+++ dwm/dwm.c   2008-12-13 18:09:16.0 +0100
@@ -343,7 +343,6 @@
Arg a = {.ui = ~0};
Layout foo = { , NULL };

-   close(STDIN_FILENO);
view(a);
lt[sellt] = foo;
while(stack)


Regards

Fred

Re: [dwm] xprop patch

2008-12-12 Thread Antoni Grzymala 
Anselm R Garbe dixit (2008-12-09, 19:44):

 If you consider that Neale's patch makes it upstream, what do you think about:
 
 while true
 do
 dwm
 done
 
 in .xinitrc to restart dwm?

How do I exit back to xdm cleanly then?

-- 
[a]

Re: [dwm] xprop patch

2008-12-10 Thread yy 
2008/12/9 Anselm R Garbe [EMAIL PROTECTED]:
 2008/12/9 Neale Pickett [EMAIL PROTECTED]:
 I very much like this patch.  I realized right away that I would never
 again need to restart dwm when I work on my status script.  When it
 dies, I can just start it up again without restarting DWM.  Someone
 could even have multiple programs running to update the status text.  It
 removes the need for the readin variable: you either set the property
 or you don't.

 I dislike X properties, however I like your patch a lot and consider
 it making mainstream, since it simplifies dwm a lot and fixes a lot of
 issues. dwm depends on X anyways, so for the status text mechanism to
 happen, it's the simpler method.

 This would be 40 if the property were hard coded to XA_WM_NAME.  I left
 in a comment to use a property named DWM_STATUS instead of XA_WM_NAME,
 so people could play with it easily, but I think XA_WM_NAME is
 ultimately the right way to do it.

 XA_WM_NAME is the way to go.


I have the same opinion. Now that I see the patch it looks so clear
that I also vote for its inclusion in main. I was completely wrong
about this, my apologies to Mr.Pickett ;)



-- 


- yiyus || JGL .

Re: [dwm] xprop patch

2008-12-09 Thread Benoit T 
On Tue, Dec 09, 2008 at 10:41:35AM -0700, Neale Pickett wrote:
 I very much like this patch.  I realized right away that I would never
 again need to restart dwm when I work on my status script.  When it
 dies, I can just start it up again without restarting DWM.  Someone
 could even have multiple programs running to update the status text.  It
 removes the need for the readin variable: you either set the property
 or you don't.
 
 The patch removes 39 SLOC:

gotta admit Neale has 2 points :)

i like the idea of not having to restart dwm when hacking on the status
script

conversely, when hacking on dwm itself, i like being able to restart dwm
without restarting my x session, yet i want the session to exit when dwm
exits, ie. dwm  xterm in .xsession is not what i want.

here is a respawn patch. it is most useful in conjunction with a
local install in the makefile, copying the newly built dwm binary over
the currently running one, inside my $HOME rather than in /usr/local

the patch costs 5 loc in the source + 1 loc in config.h
the patch is not completely portable due to use of ``environ''. i hope
that even the BSDs have that nowadays, but probably not through defining
_GNU_SOURCE, which is glibc specific.

cheers

-- 
Benoit Triquet benoit.triquet at gmail.com
 .''`.
: :' :  We are debian.org. Lower your prices, surrender your code.
`. `'   We will add your hardware and software distinctiveness to
  `-our own. Resistance is futile.
diff -ruN orig/dwm-5.3.1/config.def.h dwm-5.3.1/config.def.h
--- orig/dwm-5.3.1/config.def.h	2008-12-06 10:33:03.0 +0100
+++ dwm-5.3.1/config.def.h	2008-12-09 20:12:45.0 +0100
@@ -80,7 +80,8 @@
 	TAGKEYS(XK_7,  6)
 	TAGKEYS(XK_8,  7)
 	TAGKEYS(XK_9,  8)
-	{ MODKEY|ShiftMask, XK_q,  quit,   {0} },
+	{ MODKEY|ShiftMask, XK_q,  quit,   {.i = 0 } },
+	{ MODKEY|ShiftMask|ControlMask, XK_q,  quit,   {.i = 1 } },
 };
 
 /* button definitions */
diff -ruN orig/dwm-5.3.1/dwm.c dwm-5.3.1/dwm.c
--- orig/dwm-5.3.1/dwm.c	2008-12-06 10:33:03.0 +0100
+++ dwm-5.3.1/dwm.c	2008-12-08 20:33:16.0 +0100
@@ -23,6 +23,7 @@
  *
  * To understand everything else, start reading main().
  */
+#define _GNU_SOURCE
 #include errno.h
 #include locale.h
 #include stdarg.h
@@ -238,6 +240,7 @@
 static DC dc;
 static Layout *lt[] = { NULL, NULL };
 static Window root, barwin;
+static char **args;
 /* configuration, allows nested code to access above variables */
 #include config.h
 
@@ -1026,6 +1032,8 @@
 
 void
 quit(const Arg *arg) {
+	if(arg-i)	/* restart dwm, if fails just fall through and exit */
+		execve(args[0], args, environ);
 	readin = running = False;
 }
 
@@ -1733,6 +1741,7 @@
 	if(!(dpy = XOpenDisplay(0)))
 		die(dwm: cannot open display\n);
 
+	args = argv;
 	checkotherwm();
 	setup();
 	scan();

Re: [dwm] xprop patch

2008-12-09 Thread Anselm R Garbe 
2008/12/9 Benoit T [EMAIL PROTECTED]:
 On Tue, Dec 09, 2008 at 10:41:35AM -0700, Neale Pickett wrote:
 I very much like this patch.  I realized right away that I would never
 again need to restart dwm when I work on my status script.  When it
 dies, I can just start it up again without restarting DWM.  Someone
 could even have multiple programs running to update the status text.  It
 removes the need for the readin variable: you either set the property
 or you don't.

 The patch removes 39 SLOC:

 gotta admit Neale has 2 points :)

 i like the idea of not having to restart dwm when hacking on the status
 script

 conversely, when hacking on dwm itself, i like being able to restart dwm
 without restarting my x session, yet i want the session to exit when dwm
 exits, ie. dwm  xterm in .xsession is not what i want.

 here is a respawn patch. it is most useful in conjunction with a
 local install in the makefile, copying the newly built dwm binary over
 the currently running one, inside my $HOME rather than in /usr/local

 the patch costs 5 loc in the source + 1 loc in config.h
 the patch is not completely portable due to use of ``environ''. i hope
 that even the BSDs have that nowadays, but probably not through defining
 _GNU_SOURCE, which is glibc specific.

If you consider that Neale's patch makes it upstream, what do you think about:

while true
do
dwm
done

in .xinitrc to restart dwm?

Kind regards,
--Anselm

Re: [dwm] xprop patch

2008-12-09 Thread Benoit T 
On Tue, Dec 09, 2008 at 07:44:21PM +, Anselm R Garbe wrote:
 2008/12/9 Benoit T [EMAIL PROTECTED]:
  conversely, when hacking on dwm itself, i like being able to restart dwm
  without restarting my x session, yet i want the session to exit when dwm
  exits, ie. dwm  xterm in .xsession is not what i want.
 
 If you consider that Neale's patch makes it upstream, what do you think about:
 
 while true
 do
 dwm
 done
 
 in .xinitrc to restart dwm?

One of my requirements above was that quitting dwm should quit the X
session. Here you would have to kill .xsession. This can certainly be
done by killing dwm's parent in quit() but then the behavior when dwm is
launched from an xterm would be counter-intuitive!


Also, I initially had the requirement of being able to exec dwm at the
end of .xsession (and have a restart function). But you don't do that
with dwm reading from stdin, so no way to save one shell process.

With Neale's patch I can finally fulfill that requirement :)

One might point out that we still want one shell process for the status
script, and that shell could be the one interpreting .xsession and thus
we do not need exec'ing dwm in the first place.

I know of 2 reasons why we still want a separate process for the status
script:
- I can't think of a way to update the status script if it is sourced by
.xsession rather than forked.
- the status script might break (after all I am hacking it :) and if
sourced by .xsession could prevent the entire X session from starting up.


Speaking of a separate process, people should probably check the status
from xsetroot and exit when the X connection closes, to make it more
robust in case they detach the status script from the parent X process.

cheers

-- 
Benoit Triquet benoit.triquet at gmail.com
 .''`.
: :' :  We are debian.org. Lower your prices, surrender your code.
`. `'   We will add your hardware and software distinctiveness to
  `-our own. Resistance is futile.

Re: [dwm] xprop patch

2008-12-09 Thread Evan Gates 
You can do
while true
do
dwm || exit
done

This way when you kill dwm correctly (with modkey+shift+q as default)
it respawns, but if you kill it any other way, the Xsession finishes.
You could even add a key binding that does killall dwm or something of
the sort for when you want to end the Xsession.

-Evan

On Tue, Dec 9, 2008 at 3:44 PM, Benoit T [EMAIL PROTECTED] wrote:
 On Tue, Dec 09, 2008 at 07:44:21PM +, Anselm R Garbe wrote:
 2008/12/9 Benoit T [EMAIL PROTECTED]:
  conversely, when hacking on dwm itself, i like being able to restart dwm
  without restarting my x session, yet i want the session to exit when dwm
  exits, ie. dwm  xterm in .xsession is not what i want.

 If you consider that Neale's patch makes it upstream, what do you think 
 about:

 while true
 do
 dwm
 done

 in .xinitrc to restart dwm?

 One of my requirements above was that quitting dwm should quit the X
 session. Here you would have to kill .xsession. This can certainly be
 done by killing dwm's parent in quit() but then the behavior when dwm is
 launched from an xterm would be counter-intuitive!


 Also, I initially had the requirement of being able to exec dwm at the
 end of .xsession (and have a restart function). But you don't do that
 with dwm reading from stdin, so no way to save one shell process.

 With Neale's patch I can finally fulfill that requirement :)

 One might point out that we still want one shell process for the status
 script, and that shell could be the one interpreting .xsession and thus
 we do not need exec'ing dwm in the first place.

 I know of 2 reasons why we still want a separate process for the status
 script:
 - I can't think of a way to update the status script if it is sourced by
 .xsession rather than forked.
 - the status script might break (after all I am hacking it :) and if
 sourced by .xsession could prevent the entire X session from starting up.


 Speaking of a separate process, people should probably check the status
 from xsetroot and exit when the X connection closes, to make it more
 robust in case they detach the status script from the parent X process.

 cheers

 --
 Benoit Triquet benoit.triquet at gmail.com
  .''`.
 : :' :  We are debian.org. Lower your prices, surrender your code.
 `. `'   We will add your hardware and software distinctiveness to
  `-our own. Resistance is futile.

Re: [dwm] xprop patch

2008-12-09 Thread Anselm R Garbe 
2008/12/9 Neale Pickett [EMAIL PROTECTED]:
 I very much like this patch.  I realized right away that I would never
 again need to restart dwm when I work on my status script.  When it
 dies, I can just start it up again without restarting DWM.  Someone
 could even have multiple programs running to update the status text.  It
 removes the need for the readin variable: you either set the property
 or you don't.

I dislike X properties, however I like your patch a lot and consider
it making mainstream, since it simplifies dwm a lot and fixes a lot of
issues. dwm depends on X anyways, so for the status text mechanism to
happen, it's the simpler method.

 This would be 40 if the property were hard coded to XA_WM_NAME.  I left
 in a comment to use a property named DWM_STATUS instead of XA_WM_NAME,
 so people could play with it easily, but I think XA_WM_NAME is
 ultimately the right way to do it.

XA_WM_NAME is the way to go.

Kind regards,
--Anselm

Re: [dwm] xprop patch

2008-12-09 Thread Jeremy Jay 
Or even simpler =)

while dwm
do
  sleep 1
done

- Jeremy



On Tue 09 Dec 2008 - 04:26PM, Evan Gates wrote:
 You can do
 while true
 do
 dwm || exit
 done
 
 This way when you kill dwm correctly (with modkey+shift+q as default)
 it respawns, but if you kill it any other way, the Xsession finishes.
 You could even add a key binding that does killall dwm or something of
 the sort for when you want to end the Xsession.
 
 -Evan
 
 On Tue, Dec 9, 2008 at 3:44 PM, Benoit T [EMAIL PROTECTED] wrote:
  On Tue, Dec 09, 2008 at 07:44:21PM +, Anselm R Garbe wrote:
  2008/12/9 Benoit T [EMAIL PROTECTED]:
   conversely, when hacking on dwm itself, i like being able to restart dwm
   without restarting my x session, yet i want the session to exit when dwm
   exits, ie. dwm  xterm in .xsession is not what i want.
 
  If you consider that Neale's patch makes it upstream, what do you think 
  about:
 
  while true
  do
  dwm
  done
 
  in .xinitrc to restart dwm?
 
  One of my requirements above was that quitting dwm should quit the X
  session. Here you would have to kill .xsession. This can certainly be
  done by killing dwm's parent in quit() but then the behavior when dwm is
  launched from an xterm would be counter-intuitive!
 
 
  Also, I initially had the requirement of being able to exec dwm at the
  end of .xsession (and have a restart function). But you don't do that
  with dwm reading from stdin, so no way to save one shell process.
 
  With Neale's patch I can finally fulfill that requirement :)
 
  One might point out that we still want one shell process for the status
  script, and that shell could be the one interpreting .xsession and thus
  we do not need exec'ing dwm in the first place.
 
  I know of 2 reasons why we still want a separate process for the status
  script:
  - I can't think of a way to update the status script if it is sourced by
  .xsession rather than forked.
  - the status script might break (after all I am hacking it :) and if
  sourced by .xsession could prevent the entire X session from starting up.
 
 
  Speaking of a separate process, people should probably check the status
  from xsetroot and exit when the X connection closes, to make it more
  robust in case they detach the status script from the parent X process.
 
  cheers
 
  --
  Benoit Triquet benoit.triquet at gmail.com
   .''`.
  : :' :  We are debian.org. Lower your prices, surrender your code.
  `. `'   We will add your hardware and software distinctiveness to
   `-our own. Resistance is futile.