freenode/#stumpwm - IRC Chatlog
Search
7:25:16
impaktor
As it is now, every time I switch to a floating group, I get an error, from stumpwm::tile-group-current-frame
7:25:22
MichaelRaskin
Is checking whether current-group is of type tile-group inside the hook too hacky?
7:29:52
impaktor
embarrassing basic question: how do I get stump to re-evaluate my config? calling "reload" still gives old behavior of function.
7:34:44
impaktor
just evaluating (typep (current-group) 'tile-group) gives me an error. 'tile-group unknown.
7:37:48
impaktor
scottj: I have slime (not used it in a long time though), and I've not figured out how to make it interact with stump.
7:38:13
impaktor
I've found this, that's on my to-read list: http://www.kaashif.co.uk/2015/06/28/hacking-stumpwm-with-common-lisp/index.html
7:49:30
impaktor
Also, new question: is it possible to unload a module, to undo (load-module "cpu")? It triggers a divide by zero crash for me when running with libvirt.
7:51:55
MichaelRaskin
load-module just makes the code available, it is not easy to undo, I think. Maybe just remove the reference to it from your modeline format?
7:53:00
impaktor
You're saying if my mode line stops using it, the code will/might not be called, thus cpu-module code is still loaded, but not evaluated, so will not (hopefully) trigger more crashes?
7:56:40
MichaelRaskin
I think so. But do you want to lose the entire mode-line instead of only one indicator?
8:04:53
MichaelRaskin
Do you happen to know a setting that would make it open new windows from window.open and target=_ in new windows and not in tabs?
8:06:34
MichaelRaskin
I am currently using a set of hacks around SlimerJS, but it looks like Firefox is breaking too many APIs with every release…
8:10:06
MichaelRaskin
Well, I just bought Uzbl's «managing switching between things is what your WM is for» some time ago
8:11:19
MichaelRaskin
Switched to SlimerJS because WebKit is too close to Blink and I don't like how all this is going towards a monoculture of «IE 6.0» 2.0… And also tied all the scripts to the idea that «current browser window contains exactly one web page»
8:12:04
MichaelRaskin
So I want to migrate to Conkeror, nuke most of its bindings and make it behave as «Gecko in the form that can interact with StumpWM»
8:14:55
MichaelRaskin
I mean, I do have a Firefox window open for stuff which I am too lazy to configure, webcam/file uploads…
8:19:38
scottj
but you can't easily write a conkeror command that say executes some javascript function that's part of the webapp
8:24:41
MichaelRaskin
I guess if Conkeror breaks down, I will try to learn scripting Marionette, but Firefox is also slow to open new instaces
8:26:13
MichaelRaskin
I mean, if I was OK with WebKit/Blink «IE 6.0» 2.0, I would just go back to Uzbl
8:35:08
MichaelRaskin
That bad — I mean that maybe just stripping Firefox down and using Marionette is cpeed-comparable with Conkeror. Looks like not yet, Firefox open-close takes 2.1s vs 1.6s for Conkeror
8:43:57
MichaelRaskin
Nothing is wrong with Uzbl per se, but WebKit + Blink are close enough to create a «IE6.0» 2.0 situation, where some set of quirks is not even documented well enough but becomes a de-facto web standard, and then everything stagnates
8:53:27
MichaelRaskin
Apparently Microsoft messed up patenting the business method to enforce the patent against Google
8:56:01
MichaelRaskin
Although Microsoft never perfected the technology to the Google level: they naively thought that «extinguish» doesn't include _their own_ implementation.
9:04:53
MichaelRaskin
I wonder if this means that FF54 will make the mysteriously crashing fix for window.open in SlimerJS work
9:05:24
scottj
MichaelRaskin: I think OPEN_NEW_WINDOW not working is a different issue, I'm actually seeing it in ff46 just now
9:08:10
MichaelRaskin
It may be that fixing the crash will make updating the code in SlimerJS and Conkeror feasible