Summary: | app-editors/emacs-23.1 slow to alt-tab away from (with x11-wm/metacity and assistive technologies enabled) | ||
---|---|---|---|
Product: | Gentoo Linux | Reporter: | Alejandro Dubrovsky <alito> |
Component: | Current packages | Assignee: | Emacs project <emacs> |
Status: | RESOLVED UPSTREAM | ||
Severity: | minor | CC: | adlaiff6, gnome |
Priority: | High | ||
Version: | unspecified | ||
Hardware: | All | ||
OS: | Linux | ||
URL: | http://bugzilla.gnome.org/show_bug.cgi?id=392889 | ||
Whiteboard: | |||
Package list: | Runtime testing required: | --- | |
Attachments: |
xorg.log
xorg.conf new xorg.log |
Description
Alejandro Dubrovsky
2008-07-08 13:24:56 UTC
What window manager are you using? metacity 2.22.0 (In reply to comment #2) > metacity 2.22.0 I use that one, too, though I have a stable system and only some parts of Gnome 2.22 installed. No problems here. I cannot reproduce the problem here. Gnome team: adding you to CC, any ideas about this? More likely an X driver issue. Could you tell us what versions you have for : - x11-base/xorg-server - x11-drivers/* and could you attach a copy of your Xorg.O.log and xorg.conf? Thanks xorg-server 1.4.2 (although the log below says 1.4.0.90 is running. I haven't restarted X for a while) nvidia-drivers 173.14.09 xf86-input-evdev 1.99.2-r2 xf86-input-joystick 1.3.2 xf86-input-keyboard 1.3.1 xf86-input-mouse 1.3.0 That said, this has been an issue for a long time, so I doubt it depends on the current versions running. Xorg.log and xorg.conf attached (bugzilla didn't like them pasted in here) Created attachment 160048 [details]
xorg.log
Created attachment 160050 [details]
xorg.conf
Would you mind restarting your X? Does it happen even with a fresh session or only after a while? Thanks Created attachment 160057 [details]
new xorg.log
Here's the new xorg.log. I rebooted the machine to load the new nvidia drivers (previous session wasn't using the same drivers as I reported).
Issue is still there on a new session.
Something that I didn't make specific previously: the problem happens only when switching away from emacs by using alt-tab; if I use the mouse, the switch is instant. I was reminded of this because since restarting, the keyboard behaviour is a bit strange. The repeat goes fast for the first 5 characters or so, and then slows down as if some buffer has been filled.
<ulm> dberkholz: any idea what could be the issue with bug 231169? <Willikins> ulm: https://bugs.gentoo.org/231169 "app-editors/emacs-cvs-23.0.9999 slow to alt-tab away from"; Gentoo Linux, Applications; NEW; alito@organicrobot.com:emacs@g.o <ulm> dberkholz: remi suspected it to be a driver problem <nightmorph> solution: emerge -C emacs && emerge vim :p <dberkholz> ulm: i don't really support binary drivers. but worth checking whether the reporter has compositing enabled in metacity, and whether he can reproduce with x86-video-nv. <ulm> dberkholz: o.k., i'll ask him. and add x11 to cc So: - Can you reproduce the problem with x11-drivers/xf86-video-nv? - Do you have compositing enabled in metacity? Compositing is disabled The problem is also there when using the nv driver (I hope I don't have to test that very often, it does ugly things to my second monitor, and leaves it unusable till I reboot). Something else I noticed: switching with alt-tab while emacs is loading works fast; it is only when emacs is fully loaded that it starts behaving the way described here. Also, if running emacs 22.2.92, doing fast alt-tabs semi-freezes the X client (mouse cursor still moves, but I can't switch to another program), till I do something like ctrl-alt-f2 and kill emacs. This does not happen with emacs-cvs. Is this still an issue? Yes More possibly-useful information: I tried to see what was going on with the slow down by running emacs under xlibtrace (http://www.kev.pulo.com.au/xlibtrace/), which is just a library that uses a LD_PRELOAD to step in between all calls to xlib and just outputs what is going on, but when run under trace, the slow alt-tabbing behaviour disappears. Upgraded to app-editors/emacs-cvs-23.0.9999-r1 last weekend. Fixed! Thanks! (In reply to comment #16) > Upgraded to app-editors/emacs-cvs-23.0.9999-r1 last weekend. Fixed! Thanks! Could you verify that the bug is gone in the pretest version, emacs-cvs-23.0.90 (upstream released it today)? My celebration was premature. I had forgotten I had turned off 'Enable assistive technologies' in the Assistive Technologies Preferences a while back after reading this: https://bugs.launchpad.net/ubuntu/+source/emacs22/+bug/287577 (not exactly that, but some other related emacs bug on the ubuntu tracker which I can't find now) The preference change is only activated after restarting gnome (which is why I forgot about it) Turning on 'Enable assistive technologies' and restarting gnome brings the problem back. I'll live without the assistive technologies, but the bug should probably remain open. Finally I succeeded to reproduce this on a current ~x86 system: 1. Make sure gnome-extra/at-spi (1.26.0), gnome-extra/libgail-gnome (1.20.1), and app-editors/emacs-cvs (23.0.92) are installed 2. Start a Gnome session 3. Make sure that "assistive technologies" are enabled (if not, enable them and logout/login again) 4. Open a terminal window 5. Start Emacs with "emacs-23 -Q --daemon" 6. Open an Emacs frame with "emacsclient-23 -n -c" and focus on it 7. Press Alt-TAB At this point, the session freezes for about 30 seconds. Forgot to mention, no binary Xorg driver here, but x11-drivers/xf86-video-intel. IMHO it's unlikely that this is a driver issue. I had this issue as well, and disabling assistive technologies fixed the problem. If I can help bisect the issue, let me know. Here's relevant package versions for now: app-editors/emacs-23.1 app-emacs/emacs-daemon-0.15 gnome-extra/at-spi-1.26.0 gnome-extra/libgail-gnome-1.20.1 x11-base/xorg-server-1.6.3.901-r1 x11-drivers/xf86-video-intel-2.8.1 x11-wm/metacity-2.26.0-r1 (In reply to comment #21) > I had this issue as well, and disabling assistive technologies fixed the > problem. If I can help bisect the issue, let me know. Any help with this problem is welcome. The upstream bug report is here: <http://bugzilla.gnome.org/show_bug.cgi?id=392889> No progress here. Closing. |