You are not logged in.

#26 2012-12-17 00:30:47

netfun81
Member
Registered: 2008-07-20
Posts: 32

Re: xoat - x11 obstinate asymmetric tiler

aerosuidae, I noticed you added the ability to maximize horizontally for spot 3 and vertically for spot 2.   Is it possible to add horizontal and vertical maximization for each spot? Currently if I go fullscreen in spot 1 the apps change to fullscreen mode and remove their toolbars. I would prefer to just horizontally maximize spot 1 at times. Thanks, xoat rocks!

Offline

#27 2012-12-17 04:24:38

aerosuidae
Member
Registered: 2009-03-10
Posts: 248

Re: xoat - x11 obstinate asymmetric tiler

@netfun81, it's possible, yes, but I'm not going to add max h/v to all spots just yet... smile

In the spirit of experimentation, I want to see if the current limited (and still asymmetric) design results in a new or interesting workflow. Note how windows keep the max h/v settings all the time and reapply them in the right spots, allowing one to kind of  bounce windows between states just using normal movement controls. This new possibility would break if max h/v applied everywhere, and end up like every other tiling WM out there, which isn't necessarily the point of xoat.

Spot 2 maximize-vertical is aimed at rapidly expanding a list-style app (say, file manager).

Spot 3 maximize-horizontal is aimed at rapidly accessing to a full width terminal (in my case: a mysql prompt).

I'll have a think how to solve the full-screen-apps-losing-toolbars problem. I expect you mean apps like Chromium which hides it's tabs and address bar, which is annoying, I agree. But, on the other hand, *mplayer hiding controls and borders is good...

Offline

#28 2012-12-24 11:08:59

sime
Member
Registered: 2007-12-14
Posts: 96

Re: xoat - x11 obstinate asymmetric tiler

xoat FTW!

Would really prefer if title bars/tabs are just a single config.h comment on/off option, I have no use for it and feel like they are forced on me.

Offline

#29 2012-12-27 00:53:42

aerosuidae
Member
Registered: 2009-03-10
Posts: 248

Re: xoat - x11 obstinate asymmetric tiler

@sime, disabling title bars turned out to be relatively painless, so you get your wish. In config.h:

#define TITLE NULL

Offline

#30 2013-02-09 03:57:21

netfun81
Member
Registered: 2008-07-20
Posts: 32

Re: xoat - x11 obstinate asymmetric tiler

thanks aerosuidae, I went to ur github today and noticed you added a maximize function that works for spot 1.  That was the only feature I was missing in xoat.  It now is my perfect wm.. Thanks again!

Offline

#31 2013-02-11 05:13:54

aerosuidae
Member
Registered: 2009-03-10
Posts: 248

Re: xoat - x11 obstinate asymmetric tiler

Glad to hear it.

I'm enjoying the spot-specific maximization toggles immensely. No plans to make them more flexible. Obstinacy and Asymmetry all the way!

Offline

#32 2013-05-09 11:40:30

aerosuidae
Member
Registered: 2009-03-10
Posts: 248

Re: xoat - x11 obstinate asymmetric tiler

Still trundling along with this.  Few minor bug fixes are pushed to github since my last post.

Have also added a small usability concession:  xoat now displays the root window name in the spot1 title bar, like dwm and friends smile  yay!  Can finally script something...

Offline

#33 2013-09-26 10:13:29

sime
Member
Registered: 2007-12-14
Posts: 96

Re: xoat - x11 obstinate asymmetric tiler

Almost forgot about xoat! After many months of xmonad usage (almost stock config, just with golden ratio), I've tried xoat again for the kicks. After just a few config.h tweaks (1px border, no titlebars, 62% approx. golden ratios and a few bindings) it feels great running it again!

Thanks for the bug fixes and additional config.h tweaks/flexibility!

It appears I can only tolerate xmonad, xoat and cwm (OpenBSD version, when I don't have other options). Everything else just gets me crazy.

Offline

#34 2013-09-26 12:05:20

aerosuidae
Member
Registered: 2009-03-10
Posts: 248

Re: xoat - x11 obstinate asymmetric tiler

:-)

I'm using xoat pretty much 100% of the time now too and consider it feature complete.

Poor goomwwm gets only the odd bug fix. Should have tried static tiling years ago...

Offline

#35 2013-09-26 12:29:43

jdarnold
Member
From: Medford MA USA
Registered: 2009-12-15
Posts: 459
Website

Re: xoat - x11 obstinate asymmetric tiler

Why isn't this in the AUR? Thinking of giving it a go (been on a bit of a WM kick lately) but don't like the idea of remembering to update it by hand (I use packer for AUR interfacing).


My tech blog: Daemon Dancing in the Dark
My Arch Linux blog: ArChatter

Offline

#36 2013-09-26 15:07:19

jdarnold
Member
From: Medford MA USA
Registered: 2009-12-15
Posts: 459
Website

Re: xoat - x11 obstinate asymmetric tiler

Ah, never mind. I guess due to its source code time reconfigurability, it's probably best that it isn't in the AUR, although I supposed if you were happy with the defaults, why not?


My tech blog: Daemon Dancing in the Dark
My Arch Linux blog: ArChatter

Offline

#37 2013-09-26 20:16:36

jdarnold
Member
From: Medford MA USA
Registered: 2009-12-15
Posts: 459
Website

Re: xoat - x11 obstinate asymmetric tiler

I just built the latest xoat and noticed a problem - the function keys launch programs, even if the modifier isn't down. So, with the default config.h, hit F3 anywhere brings up pcmanfm, as does Mod4+F3.


My tech blog: Daemon Dancing in the Dark
My Arch Linux blog: ArChatter

Offline

#38 2013-09-26 23:58:04

aerosuidae
Member
Registered: 2009-03-10
Posts: 248

Re: xoat - x11 obstinate asymmetric tiler

Note the subtle difference in that F3 finds or starts pcmanfm, while Mod4+F3 always starts a new window.

The behaviour is intentional. See the last part of keys[] in config.h. If you want your function keys back remove the AnyModifier / action_find_or_start lines.

This is why xoat is not in the AUR. Having git around to branch and customize yourself a config.h is necessary ;-)

Offline

#39 2013-09-27 00:15:34

jdarnold
Member
From: Medford MA USA
Registered: 2009-12-15
Posts: 459
Website

Re: xoat - x11 obstinate asymmetric tiler

Ah. I thought AnyModifier meant a it needed a modifier, but I guess 'None' is a modifier smile


My tech blog: Daemon Dancing in the Dark
My Arch Linux blog: ArChatter

Offline

#40 2013-09-27 00:28:34

aerosuidae
Member
Registered: 2009-03-10
Posts: 248

Re: xoat - x11 obstinate asymmetric tiler

Fair point. I guess the X devs thought it was cool.

Last edited by aerosuidae (2013-09-27 00:28:47)

Offline

#41 2013-09-30 15:28:43

jdarnold
Member
From: Medford MA USA
Registered: 2009-12-15
Posts: 459
Website

Re: xoat - x11 obstinate asymmetric tiler

I think this may be a bug in xoat, but I'm not sure. If I go to this page in Firefox:

http://www.nfl.com/videos/nfl-videos/0a … cap-Week-4

and click on the button in the video that makes the video go full screen, it, I guess not surprisingly, only uses the area it is allowed to (although it is chopped off, and not scaled). But even worse, if I hit Meta-f to make that window go fullscreen, it crashes X. This happens even if  I fullscreen the page first.


My tech blog: Daemon Dancing in the Dark
My Arch Linux blog: ArChatter

Offline

#42 2013-09-30 15:32:20

jdarnold
Member
From: Medford MA USA
Registered: 2009-12-15
Posts: 459
Website

Re: xoat - x11 obstinate asymmetric tiler

Is there any way to tell xoat to put a particular window in a particular spot? This is especially true on startup, where I would just like to get the layout to be the same way without readjusting everything.


My tech blog: Daemon Dancing in the Dark
My Arch Linux blog: ArChatter

Offline

#43 2013-10-01 02:58:21

aerosuidae
Member
Registered: 2009-03-10
Posts: 248

Re: xoat - x11 obstinate asymmetric tiler

A fix has been committed for the full screen flash video crash. Let me know if it works... Thanks for reporting bugs!

Edit: The fix is for the crash. The video window will still need a manual mod+f to actually go full screen. It seems not to use normal hints and needs more work to figure out how to detect the switch. When I have time. Maybe smile

My solution for putting windows in specific spots is to set SMART in config.h and use apps that remember (or can be told) their geometry. Mostly, my windows always reappear in the spot they used last. I have no plans to add complex placement rules, but am happy to consider a patch or merge request.

I don't restart very often so setting up many windows at once is rare. If I'm tweaking a xoat build, I use `xoat restart` instead of cycling X.

Last edited by aerosuidae (2013-10-01 03:09:44)

Offline

#44 2013-10-01 13:10:02

jdarnold
Member
From: Medford MA USA
Registered: 2009-12-15
Posts: 459
Website

Re: xoat - x11 obstinate asymmetric tiler

Yeah, I'm already using Smart. Not sure why some of the apps don't stay smart. And yes, it is only at morning start up that I do the setting up, so it isn't a big deal. If I have time, maybe I'll check the mod+f problem myself.


My tech blog: Daemon Dancing in the Dark
My Arch Linux blog: ArChatter

Offline

#45 2013-10-01 13:23:03

jdarnold
Member
From: Medford MA USA
Registered: 2009-12-15
Posts: 459
Website

Re: xoat - x11 obstinate asymmetric tiler

Yup, that fix worked (once I figured out how to get it).


My tech blog: Daemon Dancing in the Dark
My Arch Linux blog: ArChatter

Offline

#46 2013-10-02 16:13:49

jdarnold
Member
From: Medford MA USA
Registered: 2009-12-15
Posts: 459
Website

Re: xoat - x11 obstinate asymmetric tiler

Found another bug.

If I invoked emacsclient from the command line, which sends a command to an open Emacs instance in another spot, it correctly puts the keyboard focus on Emacs and waits for the answer back. However, the focused window indicator still says the terminal window. And when I close the buffer in Emacs, which returns control back to the terminal, Emacs still has the keyboard focus. And even more annoying, clicking on the terminal window doesn't help, because xoat thinks it already has focus. So I have to click on another window and then click back.

Not sure how to recreate that without Emacsclient and I'm not sure what makes it special. It sends commands via a socket to a running instance of Emacs. bringing it forward.


My tech blog: Daemon Dancing in the Dark
My Arch Linux blog: ArChatter

Offline

#47 2013-10-03 04:45:34

aerosuidae
Member
Registered: 2009-03-10
Posts: 248

Re: xoat - x11 obstinate asymmetric tiler

It sounds like emacsclient might be stealing focus instead of sending an EWMH message, and just expecting the WM to keep up.  xoat defaults to doing nothing in ambiguous situations relying on co-operative apps.  A bit of googling shows other similar reports on various WMs over the years, and even different behaviour depending on the terminal used to run the command.

But I've also never used emacsclient and may be talking crap :-)

If I get a chance I'll have a look.

Last edited by aerosuidae (2013-10-03 04:47:00)

Offline

#48 2013-12-03 22:30:47

Procyon
Member
Registered: 2008-05-07
Posts: 1,819

Re: xoat - x11 obstinate asymmetric tiler

I'm WM hopping and trying xoat out. I have one issue I'm a bit stumped on. Fullscreen programs like mplayer don't become fullscreen until I use Mod4+f. When I xprop mplayer that just started and should be in fullscreen it doesn't have the STATE_FULLSCREEN set either, so it seems xoat unsets it.

window_build_client in client.c is maybe called too often. I noticed it when I put a printf in there to check fullscreen windows and rapid window switching produces a lot of CPU.

If it's helpful to anyone who wants even fewer keys to work with, I made a few really simple hacks for cycling windows in a spot, cycling spots and "toggling" between 2 monitors in action.c:
http://sprunge.us/QVbQ
I noticed that in action_move_direction and void action_focus_direction current_spot wasn't always set correctly, which might be a bug.
With this hack for action_move_direction and void action_focus_direction the .num argument for cycling should be -1 (cycle forward) or -2 (cycle backward), because LEFT/UP/etc are 1,2,3,4.
For action_cycle set .num to -1 to cycle backward.
And with my hack for action_focus_monitor the .num argument gets ignored, so don't install that if you have more than 2.

One request, the window title bar might look a lot cleaner partitioned and boxed (e.g. with 2 windows in one spot, they take up 50/50, with 3 windows each takes up 33% etc).

Offline

#49 2013-12-04 15:06:47

aerosuidae
Member
Registered: 2009-03-10
Posts: 248

Re: xoat - x11 obstinate asymmetric tiler

Procyon wrote:

I have one issue I'm a bit stumped on. Fullscreen programs like mplayer don't become fullscreen until I use Mod4+f. When I xprop mplayer that just started and should be in fullscreen it doesn't have the STATE_FULLSCREEN set either, so it seems xoat unsets it.

I don't think anything is actively unset... could be xoat missing some relevant event, maybe in client_message().

Procyon wrote:

window_build_client in client.c is maybe called too often. I noticed it when I put a printf in there to check fullscreen windows and rapid window switching produces a lot of CPU.

There is very little internal state so window_build_client() is called just about every time something happens. The main offender is likely to be query_windows(). Adding a client cache would help, but I rarely see xoat CPU topping out my process list.

How much CPU? And compared to X itself at the time? And compared to without the printf() making your terminal scroll?

Procyon wrote:

One request, the window title bar might look a lot cleaner partitioned and boxed (e.g. with 2 windows in one spot, they take up 50/50, with 3 windows each takes up 33% etc).

lol.. you don't like my crappy ascii-art tabs? ;-)

Patches welcome!

Last edited by aerosuidae (2013-12-04 15:08:15)

Offline

#50 2014-01-25 20:25:18

marttt
Member
Registered: 2013-05-21
Posts: 22

Re: xoat - x11 obstinate asymmetric tiler

I really like the concept of xoat -- looks like I just ditched both goomwwm and dwm! cool Thanks very much for your work aerosuidae!

However, I'm having an issue with the automatic rotation (for a portrait screen): the layout does get rotated and tile #1 is where it should be, but the bottom pane (tiles #2 and #3) remains empty. So all windows that should be located there are simply left off the screen.

Here's the xrandr command that produces this:

xrandr --output VGA1 --mode 1680x1050 --rotate left --output LVDS1 --off

And in config.h I have only this, so no dual displays defined:

{ .spot_start = SMART, .spot1_align = LEFT,  .spot1_width_pct = 66, .spot2_height_pct = 66 },

Thanks for any help. Cheers!

Last edited by marttt (2014-01-25 20:26:13)

Offline

Board footer

Powered by FluxBB