Re: [e-users] $QT_QPA_PLATFORMTHEME $QT_STYLE_OVERRIDE both set to gtk2
On 26/09/2018 16:37, Carsten Haitzler (The Rasterman) wrote: maybe we should provide some options in e to set these to specific values for user preferences e should check existence of these variables and do not redefine them with default settings! -- sergio. ___ enlightenment-users mailing list enlightenment-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/enlightenment-users
Re: [e-users] more mouse button bindings feature request
On 26/09/2018 13:38, Carsten Haitzler wrote: P.S. There no List-Id headers in your replies. So my sieve dosn't put them to separate folder and I'm not able to "reply to list". i "Reply to all" I know. so you may be looking at the mail from me directly to. there should also be mails from the list. that last one should have an list id header. the first would be from me and thus not have one... it'll be the cc line in my reply :) But I receive only one mail, without list id. Looks like it's google magic which doesn't send me mail when I'm in cc. I've replied to this letter with reply all. You are in cc. Check it yourself. -- sergio. ___ enlightenment-users mailing list enlightenment-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/enlightenment-users
Re: [e-users] Enlightenment memory leaks?
On Wed, 26 Sep 2018 18:18:23 +0200 leoutat...@gmx.fr said: > On 9/26/18 3:30 PM, Carsten Haitzler (The Rasterman) wrote: > > On Wed, 26 Sep 2018 13:37:42 +0200 leoutat...@gmx.fr said: > > > >> On 9/26/18 12:04 PM, Carsten Haitzler (The Rasterman) wrote: > >> > > you definitely have a leak. you want massif. so -massif option. then use > > massif-visualizer to look at the trace and you'll see it probably going > > "up and to the right". select a time point near the right/big end and > > start digging through the tree of allocations (most of the mem allocs > > are near the top). expand the trees for call traces. that there will > > begin to give an idea of what is going on. you also will have learned a > > new debugging technique :) > > > valgrind -massif log file is available here > https://phab.enlightenment.org/T7410 > >>> > >>> private paste - but we don't want the log. we want the massif output file. > >>> massif.out maybe or something (i have an alias that produces a specific > >>> one for me). even then i'm not sure i can usefully open it on a different > >>> system. i have never tried. thus massif-visualizer. > >>> > >>> > >> My apologize, access paste is public, now. Of course, it's massif output > >> file, 'massif.out', you can get it. > >> 'e-valgrind-massif.log' file has never been created, it's mistake. > >> Massif created a massif.out automatically. > > > > It's still private. :( > > > It's displayed public on my account, I don't understand. > I copied past raw text on phab site. I hope it's ok... hmmm... that leak isn't in normal malloc. paths. it seems pretty ok - climbs a bit over time. this smells like something to do with drm/gbm buffers. something not being released? -- - Codito, ergo sum - "I code, therefore I am" -- Carsten Haitzler - ras...@rasterman.com ___ enlightenment-users mailing list enlightenment-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/enlightenment-users
Re: [e-users] Beta API tutorial GUI demo, untidiness
On Sat, 15 Sep 2018 15:02:58 +0200 (CEST) Peter Koellner said: > On Sat, 15 Sep 2018, Carsten Haitzler wrote: > > > No - it's fine in this regard (the tutorial). It's something inside EFL > > that is hitting an invalid object. Unfortunately the "hex dumps" are > > useless to mail you really need to use eina_btlog as described to turn > > it into usable text. EFL does not go do this by default because it's rather > > costly to do, so doing it every time we dump this to stderr is not a good > > idea. We leave it to you to do when you want to know what that backtrace > > says or share it with others, but I can reproduce it. > > Well, basically I would expect a tutorial example to work without errors... I It's not the tutorial. It's EFL internals. the tutorial is fine. it's a tutorial for a beta API that is absolutely not stable yet and is explicitly marked as such. it is changing and not ready for general use. It's available as a kind of preview as well as for developers to give feedback on the API and design and so on. Things like you see are just part of development and happen because it's not a stable "supported" API set at this point. > did try eina_btlog now, but I really have no idea what it is telling me > there, except that efl did hit some invalid or destroyed object in closing > the program, so it probably attemtps to destroy something twice. I do have The btlog output will be human readable so you have a stack call trace to where it happened with details of the EFL objects the symbol is in, symbol, file and line number if available etc. > some other problems with enlightemnent, but I have to find the time to > produce a useful bugreport. Until now, the symptoms are a bit vague as to the > source. For example, occasionally mouse clicks just stop working, or I have > to use the laptop trackpads right mouse button to perform a left mouse click That's bizarre. how a left and right mouse click would get mixed up. I can't imagine how that might happen inside EFL - it just passes the button # in events up the chain from X. that integer would have to magically change at random for that to happen... > to make the usb mouse work temporarily again, sometimes only the systme tray > keeps working, but no window manager actions until I restart enlightenment. > Since a couple of weeks there seem to be some problem with double buffering, > I get flickering artefacts redisplaying invalidated rectangles with the old > content again. Also, there are a few problems with automated focus. When That's buffer age related there. either driver is mis-reporting buffer age or we're totally losing update regions in the buffer "ring" history. I am unsure about the latter being possible, so the only other option might be rectangle math is wrong and just sometimes loses rects when merging multiple frames of update rects... but that also seems unlikely otherwise we'd get odd results for just any update, buffer age or not and lose rects etc. so it leads me to mostly suspect buffer age, but it's hard to prove for sure. literally would have to store every buffer entirely and then before every render compare it against all buffers in the history list to see which one the content matches up with vs the age reported. it's expensive stuff. > IntelliJ Idea pops up some dialog, it sometimes triggers the mouse problem, > and each and every time some Terminology or firefox window is brought to the > front instead of that dialog box now. Also, when idea emits some notification > during the longish build process, it often steals the input focus away from > the program I am just typing in, so my input ends up in the middle of the > current source file opend in idea, even if I am typing on the other screen. I > am not sure if this all affects gtk-based programs more, but I have the not > really well founded suspicion that the problems often start after opening a > gtk-based application. So the idea behind starting with the tutorial was to there are a few possible options here, but the main would be using the netwm activate request where intellij is asking for an activation and e obliges (that means switching to that desktop, raising that window and focusing it). another option is thatr inellij is just setting focus directly. it could also be a new window being opened up and if e's policy is "focus all new windows" then this might happen - the default is to focus new windows only if a parent of that dialog is focused. it's probably the first though. in the end intellij is probably asking this to happen. can you find a setting to turn this off? in the end we could try workarounds like in remembers/locks allow windows to be on a blacklist to never "activate" if asked by a client or only do it in certain cases (that desktop is active, that screen has mouse on it, etc.) which is a workaround a bad app - we'd have to add features for that, but can be done. > start testing if efl-based programs show the same effects. There may also b
Re: [e-users] Enlightenment memory leaks?
On 9/26/18 3:30 PM, Carsten Haitzler (The Rasterman) wrote: On Wed, 26 Sep 2018 13:37:42 +0200 leoutat...@gmx.fr said: On 9/26/18 12:04 PM, Carsten Haitzler (The Rasterman) wrote: you definitely have a leak. you want massif. so -massif option. then use massif-visualizer to look at the trace and you'll see it probably going "up and to the right". select a time point near the right/big end and start digging through the tree of allocations (most of the mem allocs are near the top). expand the trees for call traces. that there will begin to give an idea of what is going on. you also will have learned a new debugging technique :) valgrind -massif log file is available here https://phab.enlightenment.org/T7410 private paste - but we don't want the log. we want the massif output file. massif.out maybe or something (i have an alias that produces a specific one for me). even then i'm not sure i can usefully open it on a different system. i have never tried. thus massif-visualizer. My apologize, access paste is public, now. Of course, it's massif output file, 'massif.out', you can get it. 'e-valgrind-massif.log' file has never been created, it's mistake. Massif created a massif.out automatically. It's still private. :( It's displayed public on my account, I don't understand. I copied past raw text on phab site. I hope it's ok... ___ enlightenment-users mailing list enlightenment-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/enlightenment-users
Re: [e-users] $QT_QPA_PLATFORMTHEME $QT_STYLE_OVERRIDE both set to gtk2
On Mon, 24 Sep 2018 12:36:53 +0100 Mick said: > On Monday, 24 September 2018 12:07:48 BST Mick wrote: > > On Monday, 24 September 2018 06:34:35 BST Simon Lees wrote: > > > On 18/09/2018 23:59, Mick wrote: > > > > On Tuesday, 18 September 2018 11:27:35 BST sergio wrote: > > > >> On 18/09/2018 11:31, Mick wrote: > > > >>> Where are QT_QPA_PLATFORMTHEME and QT_STYLE_OVERRIDE specified outside > > > >>> the > > > >>> GUI? In ~/.bashrc? Where could they be set for system wide settings? > > > >> > > > >> QT_STYLE_OVERRIDE is not specified. > > > >> > > > >> QT_QPA_PLATFORMTHEME is from /etc/X11/Xsession.d/99qt5ct, which > > > >> rewrites > > > > > > > >> nothing, and just adds: > > > > OK, this file is not installed in gentoo. > > > > > > > >> moreover, there is ~/.xsession for user settings > > > > > > > > I tried unsetting $QT_STYLE_OVERRIDE in ~.xsession, but I still got > > > > gtk2. > > > > Then, set it to a blank value using e's Settings/Advanced/Environment > > > > Variables and this time it took. However, I'm still no closer to having > > > > KDE applications launching with QT themes and icons, unless I launch > > > > them > > > > from within urxvt. > > > > > > Setting it from enlightenment's environment variable section should > > > work, the way its implemented at the moment is the least worst way we > > > could come up with to not have all users Qt apps broken by default. > > > > Yes, setting $QT_STYLE_OVERRIDE to a blank value using e's > > Settings/Advanced/ Environment Variables works in clearing the default gtk2 > > value. > > > > Adding the contents of the /etc/X11/Xsession.d/99qt5ct file Sergio mentioned > > also works in defining $QT_QPA_PLATFORMTHEME as qt5ct. > > > > What doesn't work is starting any KDE application and having it load the Qt > > theme I have selected. The only exception seems to be when I launch a KDE > > application from within urxvt, whereby the Qt theme is respected and loads > > as expected. > > > > The urxvt terminal output where Qt theme works is this: > > > > $ kwrite > > qt5ct: using qt5ct plugin > > QApplication: invalid style override passed, ignoring it. > > qt5ct: D-Bus global menu: no > > QXcbConnection: XCB error: 5 (BadAtom), sequence: 955, resource id: 0, major > > code: 19 (DeleteProperty), minor code: 0 > > > > > > While the terminology/xterm output where Qt theme is ignored is this: > > > > $ kwrite > > QApplication: invalid style override passed, ignoring it. > > Ha! Found the solution: > > Removed the /etc/X11/Xsession.d/99qt5ct file which is not installed by > default in Gentoo (may be a distro specific solution). > > Used e's GUI to set both an empty QT_STYLE_OVERRIDE and QT_QPA_PLATFORMTHEME > as qt5ct. Now Qt themes are recognised and loaded normally. :-) maybe we should provide some options in e to set these to specific values for user preferences - we set it to gtk2 as this was the most common that allows for the best seamless experience years ago. things change. so could someone do a write-up of what env vars need to be set to what values to get specific results? e.g. qt "kde look": QT_XXX_YYY=a QT_AAA_B=xyz qt "upstream qt5 look": QT_XXX_YYY=b QT_AAA_B=efg qt "mimic gtk2 look": QT_XXX_YYY=c QT_AAA_B=hij qt "mimic gtk3 look": QT_XXX_YY=d QT_AA_B=lmn ... you get the idea. so we can present a list of 3,4 or 5 options for qt app looks and then set env vars right. -- - Codito, ergo sum - "I code, therefore I am" -- Carsten Haitzler - ras...@rasterman.com ___ enlightenment-users mailing list enlightenment-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/enlightenment-users
Re: [e-users] Enlightenment memory leaks?
On Wed, 26 Sep 2018 13:37:42 +0200 leoutat...@gmx.fr said: > On 9/26/18 12:04 PM, Carsten Haitzler (The Rasterman) wrote: > > >>> you definitely have a leak. you want massif. so -massif option. then use > >>> massif-visualizer to look at the trace and you'll see it probably going > >>> "up and to the right". select a time point near the right/big end and > >>> start digging through the tree of allocations (most of the mem allocs are > >>> near the top). expand the trees for call traces. that there will begin to > >>> give an idea of what is going on. you also will have learned a new > >>> debugging technique :) > >>> > >> valgrind -massif log file is available here > >> https://phab.enlightenment.org/T7410 > > > > private paste - but we don't want the log. we want the massif output file. > > massif.out maybe or something (i have an alias that produces a specific one > > for me). even then i'm not sure i can usefully open it on a different > > system. i have never tried. thus massif-visualizer. > > > > > My apologize, access paste is public, now. Of course, it's massif output > file, 'massif.out', you can get it. > 'e-valgrind-massif.log' file has never been created, it's mistake. > Massif created a massif.out automatically. It's still private. :( -- - Codito, ergo sum - "I code, therefore I am" -- Carsten Haitzler - ras...@rasterman.com ___ enlightenment-users mailing list enlightenment-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/enlightenment-users
Re: [e-users] Enlightenment memory leaks?
On 9/26/18 12:04 PM, Carsten Haitzler (The Rasterman) wrote: even then i'm not sure i can usefully open it on a different system. i have never tried. thus massif-visualizer. May be you'll need original massif file name : massif.out.20204 ___ enlightenment-users mailing list enlightenment-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/enlightenment-users
Re: [e-users] Enlightenment memory leaks?
On 9/26/18 12:04 PM, Carsten Haitzler (The Rasterman) wrote: you definitely have a leak. you want massif. so -massif option. then use massif-visualizer to look at the trace and you'll see it probably going "up and to the right". select a time point near the right/big end and start digging through the tree of allocations (most of the mem allocs are near the top). expand the trees for call traces. that there will begin to give an idea of what is going on. you also will have learned a new debugging technique :) valgrind -massif log file is available here https://phab.enlightenment.org/T7410 private paste - but we don't want the log. we want the massif output file. massif.out maybe or something (i have an alias that produces a specific one for me). even then i'm not sure i can usefully open it on a different system. i have never tried. thus massif-visualizer. My apologize, access paste is public, now. Of course, it's massif output file, 'massif.out', you can get it. 'e-valgrind-massif.log' file has never been created, it's mistake. Massif created a massif.out automatically. ___ enlightenment-users mailing list enlightenment-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/enlightenment-users
Re: [e-users] more mouse button bindings feature request
On Tue, 18 Sep 2018 14:09:37 +0300 sergio said: > On 18/09/2018 10:12, Carsten Haitzler (The Rasterman) wrote: > > > this is a really nasty thing > > OK, thank you for explanation. > > > > i'm not sure you're using wayland yet... :) > > I've tried, but it's unusable. Unable to rotate screen, unable to use > several monitors. (I have 3 portrait monitors). Some other multiple > glitches and issues. > > > P.S. > There no List-Id headers in your replies. So my sieve dosn't put them to > separate folder and I'm not able to "reply to list". i "Reply to all" so you may be looking at the mail from me directly to. there should also be mails from the list. that last one should have an list id header. the first would be from me and thus not have one... it'll be the cc line in my reply :) -- - Codito, ergo sum - "I code, therefore I am" -- Carsten Haitzler - ras...@rasterman.com ___ enlightenment-users mailing list enlightenment-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/enlightenment-users
Re: [e-users] Enlightenment memory leaks?
On Wed, 26 Sep 2018 12:10:06 +0200 Vincent Torri said: > On Wed, Sep 26, 2018 at 12:05 PM Carsten Haitzler > wrote: > > > > On Tue, 25 Sep 2018 21:43:29 +0200 leoutat...@gmx.fr said: > > > > > On 9/25/18 9:37 AM, Carsten Haitzler (The Rasterman) wrote: > > > > On Mon, 24 Sep 2018 14:31:24 +0200 leoutat...@gmx.fr said: > > > > > > > >> On 9/24/18 7:41 AM, Simon Lees wrote: > > > >>> For a bit more info, whats the memory usage after 10 minutes and after > > > >>> 8-12 hrs, it is likely that as you run more programs e needs to use > > > >>> more memory, so having a couple more data points would be useful. > > > >> > > > >> I only run firefox, thunderbird, terminology and sometimes, pcmanfm. > > > >> Resident memory increases hugely from 2 hours. > > > >> My efl + enlightenment (Raster is the maintainer) and terminology: > > > >> https://aur.archlinux.org/packages/efl-git/ > > > >> https://aur.archlinux.org/packages/enlightenment-git/ > > > >> https://aur.archlinux.org/packages/terminology-git/ > > > >> In the same time, enlightenment-git-debug efl-git-debug > > > >> terminoly-git-debug are installed. > > > >> How can I debug this memory leak? > > > >> With valgrind? Which option? Only '-valgrind=4' check leak or > > > >> '-valgrind=all' ? > > > > > > > > you definitely have a leak. you want massif. so -massif option. then use > > > > massif-visualizer to look at the trace and you'll see it probably going > > > > "up and to the right". select a time point near the right/big end and > > > > start digging through the tree of allocations (most of the mem allocs > > > > are near the top). expand the trees for call traces. that there will > > > > begin to give an idea of what is going on. you also will have learned a > > > > new debugging technique :) > > > > > > > valgrind -massif log file is available here > > > https://phab.enlightenment.org/T7410 > > > > private paste - but we don't want the log. we want the massif output file. > > massif.out maybe or something (i have an alias that produces a specific one > > for me). even then i'm not sure i can usefully open it on a different > > system. i have never tried. thus massif-visualizer. > > it would be nice to have a massif visualizer in EFL :) I believe you just volunteered for it! :) heheheheh -- - Codito, ergo sum - "I code, therefore I am" -- Carsten Haitzler - ras...@rasterman.com ___ enlightenment-users mailing list enlightenment-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/enlightenment-users
Re: [e-users] Enlightenment memory leaks?
On Wed, Sep 26, 2018 at 12:05 PM Carsten Haitzler wrote: > > On Tue, 25 Sep 2018 21:43:29 +0200 leoutat...@gmx.fr said: > > > On 9/25/18 9:37 AM, Carsten Haitzler (The Rasterman) wrote: > > > On Mon, 24 Sep 2018 14:31:24 +0200 leoutat...@gmx.fr said: > > > > > >> On 9/24/18 7:41 AM, Simon Lees wrote: > > >>> For a bit more info, whats the memory usage after 10 minutes and after > > >>> 8-12 hrs, it is likely that as you run more programs e needs to use more > > >>> memory, so having a couple more data points would be useful. > > >> > > >> I only run firefox, thunderbird, terminology and sometimes, pcmanfm. > > >> Resident memory increases hugely from 2 hours. > > >> My efl + enlightenment (Raster is the maintainer) and terminology: > > >> https://aur.archlinux.org/packages/efl-git/ > > >> https://aur.archlinux.org/packages/enlightenment-git/ > > >> https://aur.archlinux.org/packages/terminology-git/ > > >> In the same time, enlightenment-git-debug efl-git-debug > > >> terminoly-git-debug are installed. > > >> How can I debug this memory leak? > > >> With valgrind? Which option? Only '-valgrind=4' check leak or > > >> '-valgrind=all' ? > > > > > > you definitely have a leak. you want massif. so -massif option. then use > > > massif-visualizer to look at the trace and you'll see it probably going > > > "up > > > and to the right". select a time point near the right/big end and start > > > digging through the tree of allocations (most of the mem allocs are near > > > the top). expand the trees for call traces. that there will begin to give > > > an idea of what is going on. you also will have learned a new debugging > > > technique :) > > > > > valgrind -massif log file is available here > > https://phab.enlightenment.org/T7410 > > private paste - but we don't want the log. we want the massif output file. > massif.out maybe or something (i have an alias that produces a specific one > for me). even then i'm not sure i can usefully open it on a different system. > i > have never tried. thus massif-visualizer. it would be nice to have a massif visualizer in EFL :) Vincent ___ enlightenment-users mailing list enlightenment-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/enlightenment-users
Re: [e-users] Enlightenment memory leaks?
On Tue, 25 Sep 2018 21:43:29 +0200 leoutat...@gmx.fr said: > On 9/25/18 9:37 AM, Carsten Haitzler (The Rasterman) wrote: > > On Mon, 24 Sep 2018 14:31:24 +0200 leoutat...@gmx.fr said: > > > >> On 9/24/18 7:41 AM, Simon Lees wrote: > >>> For a bit more info, whats the memory usage after 10 minutes and after > >>> 8-12 hrs, it is likely that as you run more programs e needs to use more > >>> memory, so having a couple more data points would be useful. > >> > >> I only run firefox, thunderbird, terminology and sometimes, pcmanfm. > >> Resident memory increases hugely from 2 hours. > >> My efl + enlightenment (Raster is the maintainer) and terminology: > >> https://aur.archlinux.org/packages/efl-git/ > >> https://aur.archlinux.org/packages/enlightenment-git/ > >> https://aur.archlinux.org/packages/terminology-git/ > >> In the same time, enlightenment-git-debug efl-git-debug > >> terminoly-git-debug are installed. > >> How can I debug this memory leak? > >> With valgrind? Which option? Only '-valgrind=4' check leak or > >> '-valgrind=all' ? > > > > you definitely have a leak. you want massif. so -massif option. then use > > massif-visualizer to look at the trace and you'll see it probably going "up > > and to the right". select a time point near the right/big end and start > > digging through the tree of allocations (most of the mem allocs are near > > the top). expand the trees for call traces. that there will begin to give > > an idea of what is going on. you also will have learned a new debugging > > technique :) > > > valgrind -massif log file is available here > https://phab.enlightenment.org/T7410 private paste - but we don't want the log. we want the massif output file. massif.out maybe or something (i have an alias that produces a specific one for me). even then i'm not sure i can usefully open it on a different system. i have never tried. thus massif-visualizer. -- - Codito, ergo sum - "I code, therefore I am" -- Carsten Haitzler - ras...@rasterman.com ___ enlightenment-users mailing list enlightenment-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/enlightenment-users
