[Bug 1053959] New: scrolling does not work in firefox 55
http://bugzilla.suse.com/show_bug.cgi?id=1053959 Bug ID: 1053959 Summary: scrolling does not work in firefox 55 Classification: openSUSE Product: openSUSE Tumbleweed Version: Current Hardware: Other OS: Other Status: NEW Severity: Normal Priority: P5 - None Component: Firefox Assignee: bnc-team-mozilla@forge.provo.novell.com Reporter: jslaby@suse.com QA Contact: qa-bugs@suse.de Found By: --- Blocker: --- I installed ff 55 from the mozilla repo. Scrolling in inactive window never works. And sometimes it does not work even for active window. It was always working with ff 52 and older. Plasma 5. This is an openSUSE "packaging" bug, apparently. It does not happen when I download the ff 55.0.1 tarball, unpack it and run ff from there. Safe mode does not help. Upstream bug references -- there is another guy with the same problem on openSUSE: https://bugzilla.mozilla.org/show_bug.cgi?id=1359226 https://bugzilla.mozilla.org/show_bug.cgi?id=1390184 # rpm -q `rpmqpack |grep -i mozill|sort` ca-certificates-mozilla-2.11-1.2.noarch MozillaFirefox-55.0.1-1.1.x86_64 MozillaFirefox-branding-upstream-55.0.1-1.1.x86_64 MozillaFirefox-translations-common-55.0.1-1.1.x86_64 mozilla-nspr-4.15-1.2.x86_64 mozilla-nspr-32bit-4.15-1.2.x86_64 mozilla-nss-3.31-1.4.x86_64 mozilla-nss-certs-32bit-3.31-1.4.x86_64 mozilla-nss-32bit-3.31-1.4.x86_64 MozillaThunderbird-52.2.1-1.23.x86_64 MozillaThunderbird-translations-common-52.2.1-1.23.x86_64 -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.suse.com/show_bug.cgi?id=1053959 http://bugzilla.suse.com/show_bug.cgi?id=1053959#c1 Wolfgang Rosenauer <wolfgang@rosenauer.org> changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |wolfgang@rosenauer.org Flags| |needinfo? --- Comment #1 from Wolfgang Rosenauer <wolfgang@rosenauer.org> --- I cannot confirm this on 42.3/Xfce and apparently not 42.3/Gnome. So is it Tumbleweed only or KDE related? I mean it's most likely not caused by any "packaging bug" since we do not fiddle with anything related unless the toolchain on TW broke something or so. Can you try if the 42.3 version runs on TW and if it behaves the same for example? -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.suse.com/show_bug.cgi?id=1053959 http://bugzilla.suse.com/show_bug.cgi?id=1053959#c2 Jiri Slaby <jslaby@suse.com> changed: What |Removed |Added ---------------------------------------------------------------------------- Flags|needinfo? | --- Comment #2 from Jiri Slaby <jslaby@suse.com> --- It behaves the same with the 42.3 version. BTW after I downloaded and unpacked the "upstream" version, I did a trial. I linked everything from the openSUSE package as follows: lrwxrwxrwx 1 xslaby xslaby 34 16. srp 08.59 application.ini -> /usr/lib64/firefox/application.ini lrwxrwxrwx 1 xslaby xslaby 27 16. srp 09.05 browser -> /usr/lib64/firefox/browser/ lrwxrwxrwx 1 xslaby xslaby 32 16. srp 08.59 crashreporter -> /usr/lib64/firefox/crashreporter lrwxrwxrwx 1 xslaby xslaby 36 16. srp 08.59 crashreporter.ini -> /usr/lib64/firefox/crashreporter.ini lrwxrwxrwx 1 xslaby xslaby 28 16. srp 08.58 defaults -> /usr/lib64/firefox/defaults/ lrwxrwxrwx 1 xslaby xslaby 37 16. srp 09.03 dependentlibs.list -> /usr/lib64/firefox/dependentlibs.list lrwxrwxrwx 1 xslaby xslaby 32 16. srp 09.03 dictionaries -> /usr/lib64/firefox/dictionaries/ -rw-r--r-- 1 xslaby xslaby 8187 16. srp 09.10 dist lrwxrwxrwx 1 xslaby xslaby 26 16. srp 08.44 firefox -> /usr/lib64/firefox/firefox lrwxrwxrwx 1 xslaby xslaby 30 16. srp 08.44 firefox-bin -> /usr/lib64/firefox/firefox-bin lrwxrwxrwx 1 xslaby xslaby 25 16. srp 09.02 fonts -> /usr/lib64/firefox/fonts/ lrwxrwxrwx 1 xslaby xslaby 32 16. srp 09.01 gmp-clearkey -> /usr/lib64/firefox/gmp-clearkey/ lrwxrwxrwx 1 xslaby xslaby 24 16. srp 09.01 gtk2 -> /usr/lib64/firefox/gtk2/ lrwxrwxrwx 1 xslaby xslaby 34 16. srp 09.00 chrome.manifest -> /usr/lib64/firefox/chrome.manifest drwxr-xr-x 2 xslaby xslaby 4096 9. srp 17.26 icons lrwxrwxrwx 1 xslaby xslaby 31 16. srp 08.49 icudt58l.dat -> /usr/lib64/firefox/icudt58l.dat lrwxrwxrwx 1 xslaby xslaby 25 16. srp 08.47 libfreeblpriv3.chk -> /lib64/libfreeblpriv3.chk lrwxrwxrwx 1 xslaby xslaby 24 16. srp 08.47 libfreeblpriv3.so -> /lib64/libfreeblpriv3.so lrwxrwxrwx 1 xslaby xslaby 33 16. srp 08.43 liblgpllibs.so -> /usr/lib64/firefox/liblgpllibs.so lrwxrwxrwx 1 xslaby xslaby 35 16. srp 08.43 libmozavcodec.so -> /usr/lib64/firefox/libmozavcodec.so lrwxrwxrwx 1 xslaby xslaby 34 16. srp 08.43 libmozavutil.so -> /usr/lib64/firefox/libmozavutil.so lrwxrwxrwx 1 xslaby xslaby 31 16. srp 08.43 libmozgtk.so -> /usr/lib64/firefox/libmozgtk.so lrwxrwxrwx 1 xslaby xslaby 35 16. srp 08.43 libmozsandbox.so -> /usr/lib64/firefox/libmozsandbox.so lrwxrwxrwx 1 xslaby xslaby 35 16. srp 08.43 libmozsqlite3.so -> /usr/lib64/firefox/libmozsqlite3.so lrwxrwxrwx 1 xslaby xslaby 22 16. srp 08.44 libnspr4.so -> /usr/lib64/libnspr4.so lrwxrwxrwx 1 xslaby xslaby 24 16. srp 08.44 libnssckbi.so -> /usr/lib64/libnssckbi.so lrwxrwxrwx 1 xslaby xslaby 25 16. srp 08.45 libnssdbm3.chk -> /usr/lib64/libnssdbm3.chk lrwxrwxrwx 1 xslaby xslaby 24 16. srp 08.45 libnssdbm3.so -> /usr/lib64/libnssdbm3.so lrwxrwxrwx 1 xslaby xslaby 25 16. srp 08.43 libnssutil3.so -> /usr/lib64/libnssutil3.so lrwxrwxrwx 1 xslaby xslaby 21 16. srp 08.40 libnss3.so -> /usr/lib64/libnss3.so lrwxrwxrwx 1 xslaby xslaby 21 16. srp 08.46 libplc4.so -> /usr/lib64/libplc4.so lrwxrwxrwx 1 xslaby xslaby 22 16. srp 08.46 libplds4.so -> /usr/lib64/libplds4.so lrwxrwxrwx 1 xslaby xslaby 23 16. srp 08.46 libsmime3.so -> /usr/lib64/libsmime3.so lrwxrwxrwx 1 xslaby xslaby 26 16. srp 08.47 libsoftokn3.chk -> /usr/lib64/libsoftokn3.chk lrwxrwxrwx 1 xslaby xslaby 25 16. srp 08.47 libsoftokn3.so -> /usr/lib64/libsoftokn3.so lrwxrwxrwx 1 xslaby xslaby 21 16. srp 08.48 libssl3.so -> /usr/lib64/libssl3.so lrwxrwxrwx 1 xslaby xslaby 28 16. srp 08.43 libxul.so -> /usr/lib64/firefox/libxul.so lrwxrwxrwx 1 xslaby xslaby 36 16. srp 09.00 minidump-analyzer -> /usr/lib64/firefox/minidump-analyzer -rw-r--r-- 1 xslaby xslaby 8344 16. srp 09.08 mozz lrwxrwxrwx 1 xslaby xslaby 26 16. srp 08.50 omni.ja -> /usr/lib64/firefox/omni.ja lrwxrwxrwx 1 xslaby xslaby 29 16. srp 09.00 pingsender -> /usr/lib64/firefox/pingsender lrwxrwxrwx 1 xslaby xslaby 31 16. srp 08.59 platform.ini -> /usr/lib64/firefox/platform.ini lrwxrwxrwx 1 xslaby xslaby 35 16. srp 08.49 plugin-container -> /usr/lib64/firefox/plugin-container -rw-r--r-- 1 xslaby xslaby 2343 9. srp 17.46 precomplete -rw-r--r-- 1 xslaby xslaby 707 9. srp 17.26 removed-files -rwxr-xr-x 1 xslaby xslaby 8915 9. srp 17.03 run-mozilla.sh lrwxrwxrwx 1 xslaby xslaby 37 16. srp 09.00 Throbber-small.gif -> /usr/lib64/firefox/Throbber-small.gif -rwxr-xr-x 1 xslaby xslaby 229536 9. srp 17.45 updater -rw-r--r-- 1 xslaby xslaby 681 9. srp 17.45 updater.ini -rw-r--r-- 1 xslaby xslaby 132 9. srp 17.45 update-settings.ini And it still works (why??). So I checked /proc/<pid>/maps of both and the difference is: --- mozz 2017-08-16 09:08:04.115102953 +0200 +++ dist 2017-08-16 09:10:13.347237950 +0200 @@ -40,7 +40,6 @@ /usr/lib64/gconv/gconv-modules.cache /usr/lib64/gconv/UTF-16.so /usr/lib64/gdk-pixbuf-2.0/2.10.0/loaders/libpixbufloader-png.so -/usr/lib64/gdk-pixbuf-2.0/2.10.0/loaders/libpixbufloader-svg.so /usr/lib64/gio/modules/libgiofam.so /usr/lib64/gio/modules/libgvfsdbus.so /usr/lib64/gtk-3.0/modules/libcanberra-gtk3-module.so @@ -60,7 +59,6 @@ /usr/lib64/libcanberra-0.30/libcanberra-oss.so /usr/lib64/libcanberra-0.30/libcanberra-pulse.so /usr/lib64/libcap.so.2.25 -/usr/lib64/libcroco-0.6.so.3.0.1 /usr/lib64/libdatrie.so.1.3.2 /usr/lib64/libdbus-glib-1.so.2.3.3 /usr/lib64/libdbus-1.so.3.14.12 @@ -114,7 +112,6 @@ /usr/lib64/libproxy-0.4.14/modules/network_networkmanager.so /usr/lib64/libproxy-0.4.14/modules/pacrunner_webkit.so /usr/lib64/libpulse.so.0.20.2 -/usr/lib64/librsvg-2.so.2.40.18 /usr/lib64/libsmime3.so /usr/lib64/libSM.so.6.0.1 /usr/lib64/libsndfile.so.1.0.28 @@ -147,7 +144,6 @@ /usr/lib64/libXinerama.so.1.0.0 /usr/lib64/libXi.so.6.1.0 /usr/lib64/libxkbcommon.so.0.0.0 -/usr/lib64/libxml2.so.2.9.4 /usr/lib64/libXrandr.so.2.2.0 /usr/lib64/libXrender.so.1.3.0 /usr/lib64/libXss.so.1.0.0 I.e. the upstream firefox uses some more libs and that's all. I don't know. -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.suse.com/show_bug.cgi?id=1053959 http://bugzilla.suse.com/show_bug.cgi?id=1053959#c3 --- Comment #3 from Wolfgang Rosenauer <wolfgang@rosenauer.org> --- Those libs seem all SVG related and I'm pretty sure not pulled in directly from Firefox during build. Looks more like a gdk build configuration thing. So still I have no idea. If 42.3 builds fail for you as well it could still be runtime environment on TW. I just have a TW VM. Not sure if I can easily reproduce in lab testing. -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.suse.com/show_bug.cgi?id=1053959 http://bugzilla.suse.com/show_bug.cgi?id=1053959#c4 --- Comment #4 from Jiri Slaby <jslaby@suse.com> --- What I don't understand on the above, why the behaviour and libraries differ when all is linked from the distro package? Anyway, it is easily reproducible also in my VM with 42.2 + ff 55 from the mozilla repo. I can provide you with the VM, if you want. -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.suse.com/show_bug.cgi?id=1053959 http://bugzilla.suse.com/show_bug.cgi?id=1053959#c5 --- Comment #5 from Jiri Slaby <jslaby@suse.com> --- Happens also with 54.0.1-2.2. Does not happen with 52.3.0-1.1 -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.suse.com/show_bug.cgi?id=1053959 http://bugzilla.suse.com/show_bug.cgi?id=1053959#c6 --- Comment #6 from Jiri Slaby <jslaby@suse.com> --- $ osc rdiff mozilla firefox52 mozilla firefox54 shows: --- mozilla.sh.in (revision 16) +++ mozilla.sh.in (revision 5) @@ -102,6 +102,9 @@ # disable Gnome crash dialog (doesn't make sense anyway) export GNOME_DISABLE_CRASH_DIALOG=1 +# enable xinput2 (boo#1032003) +export MOZ_USE_XINPUT2=1 + moz_debug=0 script_args="" pass_arg_count=0 And indeed this fixes it! # enable xinput2 (boo#1032003) #export MOZ_USE_XINPUT2=1 -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.suse.com/show_bug.cgi?id=1053959 http://bugzilla.suse.com/show_bug.cgi?id=1053959#c7 --- Comment #7 from Wolfgang Rosenauer <wolfgang@rosenauer.org> --- I'm just more and more confused ;-) Probably I also don't get how to reproduce it. What is an "inactive" window? If it's fully inactive how would you scroll at all in it? Probably I have in theory the same issue but never noticed. Just to be sure. FF 52.3 from TW works correctly? What about 52.3 from 42.x? Just wondering because the latter is built with Gtk2 while all the others (52.x on TW and FF > 52 are all built with Gtk3). Good finding of the MOZ_USE_XINPUT2. Please report your findings when playing with that. -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.suse.com/show_bug.cgi?id=1053959 http://bugzilla.suse.com/show_bug.cgi?id=1053959#c8 --- Comment #8 from Jiri Slaby <jslaby@suse.com> --- In KDE/plasma, open one firefox and one other window (e.g. xterm). Switch to xterm by alt-tab making firefox an "inactive window" = in background. Even though firefox is inactive, you still can scroll the content by using mouse scroll wheel over the inactive window (while xterm is still in foreground = active). That does not work with MOZ_USE_XINPUT2=1. Works: versions up to 52.3 Does not work: version starting 54 No matter which repository I use (42.3 repo on TW installation, TW repo on 42.3 or on 42.2). So only the ff version matters, exactly the one which added MOZ_USE_XINPUT2=1. Needless to say that I use no touchpad, it's all with mouse. -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.suse.com/show_bug.cgi?id=1053959 http://bugzilla.suse.com/show_bug.cgi?id=1053959#c9 --- Comment #9 from Jiri Slaby <jslaby@suse.com> --- Quoting from https://bugzilla.mozilla.org/show_bug.cgi?id=1390795#c1 We explicitly keep MOZ_USE_XINPUT2 off by default because behaviour is buggy when it is turned on. The code is at http://searchfox.org/mozilla-central/rev/13148faaa91a1c823a7d68563d9995480e7... and references a couple of bugs. From the discussion in those bugs it's still not clear to me whether this needs to be fixed in Firefox or elsewhere. Some of the comments seem to imply it affects other apps as well, not just Firefox, which would make it a GTK issue most likely. But I'm not sure. -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.suse.com/show_bug.cgi?id=1053959 Christoph Feck <cfeck@kde.org> changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |cfeck@kde.org -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.suse.com/show_bug.cgi?id=1053959 http://bugzilla.suse.com/show_bug.cgi?id=1053959#c12 --- Comment #12 from Jiri Slaby <jslaby@suse.com> --- (In reply to Yunhe Guo from comment #10)
Most likely, related to kmozillahelper.
No, it's not. Does commenting out #export MOZ_USE_XINPUT2=1 in /urs/bin/firefox as noted in comment 6 help? -- You are receiving this mail because: You are on the CC list for the bug.
participants (1)
-
bugzilla_noreply＠novell.com