You are not logged in.
@Cloudef: Sorry for getting back to you so late; I was out of town for a bit, and haven't bothered to troubleshoot yet. I'll get you the info as soon as I can.
Offline
Hey Cookie!
Would you be able to include the usless gap mod in the multi-monitor branch?
Offline
@absolutionx,
$ git clone git://github.com/c00kiemon5ter/monsterwm.git -b multi-monitor
$ cd monsterwm
$ git merge origin/uselessgaps
.. save the file that appears
.. and you're ready to go
$ make
$ $EDITOR config.h
$ sudo make clean install # .. etc
.:[ git me! ] :.
Offline
@absolutionx,
$ git clone git://github.com/c00kiemon5ter/monsterwm.git -b multi-monitor $ cd monsterwm $ git merge origin/uselessgaps .. save the file that appears .. and you're ready to go $ make $ $EDITOR config.h $ sudo make clean install # .. etc
You're my hero! Thanks so much
Offline
keep in mind that the multi-monitor branch is not well tested, and especially after the latest changes, it is very probable that it will need some work (cause I messed with some things ). If you get any bugs/weird behavior, report it
.:[ git me! ] :.
Offline
keep in mind that the multi-monitor branch is not well tested, and especially after the latest changes, it is very probable that it will need some work (cause I messed with some things ). If you get any bugs/weird behavior, report it
Will do. I was originally using Cloudef's multi-monitor until I just discovered that you were working on a similar version of his MM as well.
I'm relatively new to DWM/MonsterWM but I'll let you know if I find anything that doesn't work.
Offline
Yeah I'm having trouble with the MM branch crashing X when using the kill client command.
I checked messages.log and found this
Jun 21 23:03:59 localhost dbus[289]: [system] Reloaded configuration
Jun 21 23:13:31 localhost kernel: [ 692.959928] monsterwm[1266]: segfault at 7fff8c444ff8 ip 0000000000402299 sp 00007fff8c444fe0 error 6 in monsterwm[400000+5000]
Jun 21 23:16:01 localhost kernel: [ 842.304258] monsterwm[1418]: segfault at 7fff2d3daff0 ip 00007f7204154070 sp 00007fff2d3db008 error 6 in libX11.so.6.3.0[7f720410f000+133000]
Jun 21 23:17:45 localhost kernel: [ 945.927044] monsterwm[1639]: segfault at 7fff3f326ff8 ip 000000000040226d sp 00007fff3f327000 error 6 in monsterwm[400000+5000]
Jun 21 23:37:45 localhost -- MARK --
Jun 21 23:44:00 localhost kernel: [ 2516.710912] monsterwm[2004]: segfault at 7fffd4319ff0 ip 00007fc0d8747070 sp 00007fffd431a008 error 6 in libX11.so.6.3.0[7fc0d8702000+133000]
Offline
Offline
I guess the XCB branch is not a priority now right?
Offline
XCB needs to get updated. I will probably make another branch and try to keep it up to date.
btw, OMG @Cloudef, @kuraku, one of you,
please try CLICK_TO_FOCUS with Numlock off(!) and tell me if that's the issue :S :S
Last edited by c00kiemon5ter (2012-06-22 21:07:07)
.:[ git me! ] :.
Offline
Hey all, I'm still catching up on this thread, but I just started playing with monsterwm - it's great.
I actually started with Cloudef's XCB version, but I had a significant problem which I thought about reporting on Cloudef's github - but if this may be taken over by someone else I'll bring it up here: the xcb version consistently goes haywire whenever a file dialog opens. For example in luakit or firefox when I opt to "save linked file as" a rectanlge where the file dialog should be comes up, but focus bounces rapidly between the (empty) file dialog and the browser window. Neither window is at all responsive, and it's even a challenge to get to another window to run a killall <browser>. The Xlib monsterwm works just fine in this respect, so this must be something different in the XCB port. I did change back and forth between the xlib and xcb versions a few times, and this was a consistent (100% of the time) problem in xcb, and never had a problem in xlib.
Also, if useful, I had some trouble with the click to focus in both versions. I was content with going with the focus-follows-mouse option which works flawlessly. However, whenever I try the click-to-focus option, clicking to focus "works", but then no client ever receives the click event. Note that this was the version from ~1 week ago, so recent changes may have fixed that.
Last edited by Trilby (2012-06-22 21:14:45)
"UNIX is simple and coherent..." - Dennis Ritchie, "GNU's Not UNIX" - Richard Stallman
Offline
Hey all, I'm still catching up on this thread, but I just started playing with monsterwm - it's great.
I actually started with Cloudef's XCB version, but I had a significant problem which I thought about reporting on Cloudef's github - but if this may be taken over by someone else I'll bring it up here: the xcb version consistently goes haywire whenever a file dialog opens. For example in luakit or firefox when I opt to "save linked file as" a rectanlge where the file dialog should be comes up, but focus bounces rapidly between the (empty) file dialog and the browser window. Neither window is at all responsive, and it's even a challenge to get to another window to run a killall <browser>. The Xlib monsterwm works just fine in this respect, so this must be something different in the XCB port. I did change back and forth between the xlib and xcb versions a few times, and this was a consistent (100% of the time) problem in xcb, and never had a problem in xli
Also, if useful, I had some trouble with the click to focus in both versions. I was content with going with the focus-follows-mouse option which works flawlessly. However, whenever I try the click-to-focus option, clicking to focus "works", but then no client ever receives the click event. Note that this was the version from ~1 week ago, so recent changes may have fixed that.
heh, the previous post should answer most of your questions .
xcb needs updating, and click_to_focus is the next thing on my todo.
I think I figured out why it didn't work everywhere.
.:[ git me! ] :.
Offline
alright, I pushed a change that should fix CLICK_TO_FOCUS behavior.
whenever you get the time, @kuraku and/or @Cloudef please test
next thing I'll into will probably be the xcb version, once I get some time to read a bit on xcb
Last edited by c00kiemon5ter (2012-06-23 13:12:20)
.:[ git me! ] :.
Offline
Click to foucs works great here now, thanks.
Is there a way to force a window back into tiling aside from changing the layout?
I add that and one other function bound to mouse button 3 and mod4+space respectively
static void un_float() {
if (current->isfloating) current->isfloating = False;
tile(); update_current(current);
}
static void next_mode() {
switch(mode) {
case TILE: mode = BSTACK; break;
case BSTACK: mode = MONOCLE; break;
case MONOCLE: mode = GRID; break;
case GRID: mode = TILE; break;
}
tile(); update_current(current);
desktopinfo();
}
Let me know if I'm reinventing the wheel with these.
"UNIX is simple and coherent..." - Dennis Ritchie, "GNU's Not UNIX" - Richard Stallman
Offline
thnx for CLICK_TO_FOCUS feedback
you can retile all floating clients by reselecting the current layout. you cannot however retile just one floating client, unless you move all other floating windows to another desktop, or introduce a new function as you did.
you don't need to reselect the current client, it's alread current.
you should also check if there is indeed a client to change it's state.
static void un_float() {
if (curr && curr->isfloating) curr->isfloating = False;
tile();
}
next_mode cycles the layouts. that isn't implemented, so you're not reinventing the wheel
you can do it though more easily as:
static void cyclemodes(const Arg *arg) {
switch_mode(&(Arg){.i = (MODES + mode + arg->i) % MODES});
}
so now you have
{ MOD1, XK_n, cyclemodes, {.i = +1}}, /* next layout */
{ MOD1, XK_p, cyclemodes, {.i = -1}}, /* prev layout */
Last edited by c00kiemon5ter (2012-06-26 17:29:04)
.:[ git me! ] :.
Offline
Cool thanks. I figured it could be cleaner, that was just the first pass to see if I could get it to work.
"UNIX is simple and coherent..." - Dennis Ritchie, "GNU's Not UNIX" - Richard Stallman
Offline
alright, I pushed a change that should fix CLICK_TO_FOCUS behavior.
whenever you get the time, @kuraku and/or @Cloudef please test
It is working \o/
Offline
Offline
I just updated monsterwm with a couple of (ehm probably many) structual changes, nothing visible, and should be working just like before. If I missed something, please report.
I was thinking that counting lines and having line limits was a bit kiddish, but now monsterwm is about 20loc shorter currently, monsterwm.c is 599loc on master and 553loc on core branch now.
the only change you should be worried about is changing
static key keys[] = {
to ^
static Key keys[] = {
that is, use a capital 'K' for the keys array, in your config.h.
I think I can make some more changes and save some more lines
maybe I'll do that, when I'll have time to test it.
on other news:
- XCB is on the wait list still
- I'm thinking of adding a feature on a new branch to share a window in multiple desktops, not sure how this will be implemented yet, but I'll see if I can think of something that doesnt require too much rework of the current structs.
- there is a possibility, that I get an unused second monitor, if that happens I'll go after multi-monitor bugs
Last edited by c00kiemon5ter (2012-06-26 17:24:12)
.:[ git me! ] :.
Offline
@c00kiemon5ter
When i close any client, with monsterwm from latest commit, with shift+alt+c, every client on all workspaces disappears. I changed "key" to "Key" and my config.h is same as before: http://sprunge.us/SQUB?c
I managed to catch some tty1 messages and they mentioned monsterwm, memory and xcb? Not sure what was that about but i'm unable to get that since i dont know where is that log stored.
Besides that, i'm unable to use it further to check any other things in latest build. I will revert back to the one from 24.06 (it works perfectly atm for me).
Offline
thanks for the report
I see what you say, but this should be happening on the previous version too.
Open three windows and shift-alt-c the first one (the head/master window).
I think it's a double free case, will fix it asap.
.:[ git me! ] :.
Offline
thanks for the report
I see what you say, but this should be happening on the previous version too.
Open three windows and shift-alt-c the first one (the head/master window).
I think it's a double free case, will fix it asap.
Oh yeah, i expirienced this bug before. Do you remember when i described you, on IRC, those 2 bugs that i found? You fixed one already and this is second one
Offline
Sorry for the double post but i have to report that mobar is not wokring like it should. I compiled it but i see only bright green bar and the only time that i actually saw app squares is when i launch firefox. First time was okay until i changed width/height and colors. After that i was unable to change anything. Any help on this?
Edit:
It seems to be working now. Strange, i had to turn off my computer to make it work. Maybe there were several processes in the bg.
Last edited by kuraku (2012-06-28 18:09:18)
Offline
hey all,
I wasted two days trying to understand what was wrong with the code (kuraku's report). I knew where the problem was, I knew what would fix it, but I still cannot understand why that code was not working. I created three test cases, emulating that behavior, and everything worked fine, I shared code with other people and none found it weird or could think of it misbehaving for some reason .. but the problem was there. I have verified the line that causes it, the exact call, but I cannot understand it. So, I gave up and went back to the old code for that part (2 lines more ). gah, ffrrrphhht >.<
so, the new version should be ok. I tested it as much as I could, and it should be working good.
Fixed a couple of glitches (ie if one had two windows on a desktop, one tiled, one floating and move or closed the floating one, the tiled one would lose its borders but would not get resized to filled that gap, so it looked like it had uselessgaps around it).
Windows that spawn floating(or on FLOAT mode), now appear in the center of the screen, by default. I think it's better than having them spawn on the top left corner.
Fixed an issue for tiled to floating clients not getting on top of other floating clients
Reduced some calls that were not needed.
I think there's something else I wanted to say, but I forgot what that was <.<
------------------------------------------------------------------------------------
for mopag, I spawn it like:
.xinitrc:
....
exec wmrun.sh
wmrun.sh:
ff="/tmp/monsterwm.fifo"
[ -p "$ff" ] || mkfifo -m 600 "$ff"
mopag < "$ff" &
monsterwm | tee -a "$ff"
the 'tee' there lets the output from monsterwm also show up on the console. I do that because if I'm debugging or fiddling with something I want to looking at what the wm thinks it's going on. If you dont want that, as @kuraku asked, one can just skip that.
monsterwm > "$ff"
I think I'll change the scripts in the gist to match this.
or if one doesnt even want a pipe or output at all, just mute it
monsterwm 2>&1 1>/dev/null
if you want to run some_sorta_bar/dzen2 etc along with mopag, (or some_sorta_bar along with dzen2 etc), do
ffm="/tmp/monsterwm.fifo.m"
ffs="/tmp/monsterwm.fifo.s"
[ -p "$fmf" ] || mkfifo -m 600 "$ffm"
[ -p "$fms" ] || mkfifo -m 600 "$ffs"
mopag < "$ffm" &
filter_for_ssb < "$ffs" | some_sorta_bar
monsterwm | tee -a "$ffm" | tee -a "ffs" # or again change the second tee to w/e you need to mute output
Last edited by c00kiemon5ter (2012-06-29 01:27:12)
.:[ git me! ] :.
Offline
Tell me, why is this happening:
http://sprunge.us/aFQP
I changed "key" to "Key" and nothing happens (it seems that it is reading "key"?).
Offline