On Sat, 25 Mar 2017 07:54:12 -0700 Marc MERLIN <marc_...@merlins.org> said:
> On Fri, Mar 24, 2017 at 08:59:44AM +0900, Carsten Haitzler wrote: > > > Right, but if you're going to try and tell me that I should stop using > > > the way I've been typing for 20 years, you're not going to win :) > > > > just saying... works fine for me... standard french layout doesn't do alt+e > > = é ... > > > > i can only imagine the xkb settings e configures somehow conflict with your > > custom modmap stuff and you then need to do it entirely yourself. check "do > > not apply any keyboard settings ever" and then entirely do everything via > > custom xmodmap files and so on. > > Right, it's just that the E I tried in debian-unstable went into some bad > CPU loop if I used right alt + e. > Reverting to old E made this go away. > It's definitely doing something wrong. > Did you want my Xmodmap again to try it out? t_unix provided a modmap file i tried and there is an issue - but it's on the xserver side. every hangs for about 5-10sec when running xmodmap and e is getting hung in XGrabKey calls when it ungrabs and re-grabs all its bindings when a kbd layout change event happened. the grabs are taking an eternity inside the xserver (xlib is waiting for a reply from x). this is definitely some server-side issue. i can't see any issue with his modmap ant alt+e causing anything with e... :/ but the setup of the modmap is horribly slow. > > > I have a slghtly newer one at work, and it's terrible with google chrome > > > each time I try to rip out a tab from google chrome, it gets stuck to my > > > mouse pointer and it's very hard to "re-deposit" as a standalone window. > > > > the window keeps following your mouse even though you released the mouse > > button? > > Correct. That bug has been around in E for quite a while now, although > chrome might be partially to blame, but with older E it never happens. well it's not e grabbing the mouse... :) and it's not e responding to the events. e will not be grabbing anything. at most it's just hearing the click then either doing an allow events or not. any grab at this point is a grab done by chromium... and it's not releasing... :) > Fix: Ctrl+T forces that sticky chrome window stuck to my mouse pointer, to > open a new tab, and then things fix themselves. as best i can imagine it's chromium holding onto a grab. it's likely relying on or ASSUMING specific behaviour... and of course only tested on a few wm's... i don't know. but what it is doing is definitely getting into territory where you have to be SUPER careful... as you have an implicit mouse grab already on the original window before the tear-off happened. maybe it got some mouse leave event and thus it forgot about its grab? i don't know... but here the ball is in chromium's ballcourt to get it right... it's getting the mouse move events... i find it rather bizarre that it wouldn't get the mouse release... and if it doesn't that's not actually up to E as events go directly to the client... so i am just imagining it's something like an implicit mouse leave or maybe loss of focus or something that causes the chromium code to then just ignore the mouse release ... because if it's getting the moves it most definitely *SHOULD* get the release... > > > Kind of sad that upgrading e has been so painful each time I've tried > > > :-/ > > > > trust me. chromium has also had very painful points. especially the move to > > aurora - lots of stuff broke. input methods didn't work for maybe 1-2 > > years... all of that disappeared... the "gtk theme mimicking" still looks > > like junk ... it was decent until aurora... chromium would literally > > hard-lock/deadlock itself when displaying tooltips - beats me why, both in > > e and in gnome > > Ah yes, I had this too, but again only with more recent E. happened in gnome too. but it was chromium deadlocking ITSELF... it should just not do that. likely someone assuming a specific order of events that wasn't actually guaranteed and some wm's re-ordered them or dropped some as they were allowed to... that's my guess. > Downgrading E made this problem go away too. > Definitely an interaction between the 2. yup. but this is similar to the tearing off issue i bet... just saying that not every bug that happens because you run e is an e bug. there are definitely bugs that e has. there are ones that are "wm and client disagree" and those are tough, and those where "client is just dead wrong but in the end everyone blames the wm so we have to figure out a work around in the end". :) in the past few years chrome has actually been one of the problem clients. most others work just fine. :) > > for aaaages. but it was an internal deadlock and no app should deadlock just > > due to some ordering of events due to how a wm may or may not be managing > > windows. still no idea why. it's stopped... but it should never have > > happened. > > Agreed. > > > welcome to the software world where everyone creates pain and it's worse > > when developers don't test together. > > True :) :) the best i can advise is "stay up to date and help us sort out issues". downgrading doesn't help things get sorted out. :( the altgr+e - i cant see. things are working fine for me... but there is a pause issue and i dig in and e basically sits forever in every XGrabButton xlib call waiting for the xserver to respond... :( > Marc > -- > "A mouse is a device used to point at the xterm you want to type in" - A.S.R. > Microsoft is to operating systems .... > .... what McDonalds is to gourmet > cooking Home page: http://marc.merlins.org/ > > ------------------------------------------------------------------------------ > Check out the vibrant tech community on one of the world's most > engaging tech sites, Slashdot.org! http://sdm.link/slashdot > _______________________________________________ > enlightenment-users mailing list > enlightenment-users@lists.sourceforge.net > https://lists.sourceforge.net/lists/listinfo/enlightenment-users > -- ------------- Codito, ergo sum - "I code, therefore I am" -------------- The Rasterman (Carsten Haitzler) ras...@rasterman.com ------------------------------------------------------------------------------ Check out the vibrant tech community on one of the world's most engaging tech sites, Slashdot.org! http://sdm.link/slashdot _______________________________________________ enlightenment-users mailing list enlightenment-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/enlightenment-users
