[Bug 1178041] New: rkward 0.7.2 [lp152.34.1] crashes if started as root
http://bugzilla.opensuse.org/show_bug.cgi?id=1178041 Bug ID: 1178041 Summary: rkward 0.7.2 [lp152.34.1] crashes if started as root Classification: openSUSE Product: openSUSE Distribution Version: Leap 15.2 Hardware: Other OS: openSUSE Leap 15.2 Status: NEW Severity: Major Priority: P5 - None Component: Maintenance Assignee: maintenance@opensuse.org Reporter: ndordea@gmail.com QA Contact: qa-bugs@suse.de Found By: --- Blocker: --- Hello, After upgrading rkward from 0.7.1 to 0.7.2 [ lp152-34.1-x86_64 ] I had some issues when started as user [ ~/share/rkward/QtWebEngine/Default can not be created ] . I fixed that by manually creating that directory ] An installation check was done and so far it runs ok . R level R 4.0.3 However, when started as root [ for R maintenance ] rkward crashes .... phase 1 ======= as root .... /usr/bin/rkward got the panel to select a RData, select the usual RData, then got these messages ... QStandardPaths: XDG_RUNTIME_DIR not set, defaulting to '/tmp/runtime-root' QStandardPaths: XDG_RUNTIME_DIR not set, defaulting to '/tmp/runtime-root' [27199:27199:1022/175843.788358:ERROR:zygote_host_impl_linux.cc(89)] Running as root without --no-sandbox is not supported. See https://crbug.com/638180. phase 2 ======== try to start rkward again .... got the panel to select a RData select the usual RData, then got Crash recovery file detected ... select recover .... then it crashes again . sudo journalctl -a |grep rkward ... got 6 identical lines as below .. Oct 22 17:31:26 <localhost> rkward.rbackend[25491]: QIODevice::write (QFile, "/tmp/rkward.rbackend"): device not open Please note that /tmp/runtime-root is empty . Please have a look at this issue . Thank you, Nick -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.opensuse.org/show_bug.cgi?id=1178041 http://bugzilla.opensuse.org/show_bug.cgi?id=1178041#c1 Nick Dordea <ndordea@gmail.com> changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |ndordea@gmail.com --- Comment #1 from Nick Dordea <ndordea@gmail.com> --- Hello, I looked to rkward's yast latest log as well as the page mentioned by the failing message [ provided below ]. 1. The config log says ..... a) * Support using QWebEngine instead of QtWebKit (still supported) b) Build with QtWebEngine (instead of QtWebKit) on architectures and distributions where it is available (in a new enough version) Question : How to start rkward and use QtWebKit and NOT the QWebEngine that appears to be a Google/Chromium back door in our machines ? Note : I do not use Google's Chrome, Chromium because I do not trust them. If opensuse is pushing chromium on the users via rkward I will switch to Rstudio or Rcmndr . 2. the url https crbug.com/638180 is clearly discussing a chromium crush ! Note : I want to stay as far as possible from any Google/Facebook/etc products because I do not trust them. Thanks, ND ============================================== 1) rkward latest change log ========================== Wed 21 Oct 2020 07:00:00 AM CDT Wolfgang Bauer <wbauer@tmo.at> - upstream release 0.7.2 - New features and improvements * Script preview keeps vertical scroll position when updating * Add function rk.home() for retrieving application paths, similar to R.home() * Add menu option to switch application language * Remove direct dependency on libintl * Add "Check installation" wizard to test for several common basic installation issues all in one place * Add rkward internal package location to end of library search path, to make it accessible to help.search() * Add menu action to open any supported file type, directly * Support using QWebEngine instead of QtWebKit (still supported) * <text> elements in plugins may now also contain clickable links, including rkward://-scheme links * The new code hinting features from version 0.7.1 are now also available in the console * On unix-systems, RKWard can now be run without installation * Kate addons are now supported within RKWard. Intially, search-in-files, snippets, and projects are loaded by default * Python development scripts have been ported to python3 - Bug fixes * Fix some problems with plot preview in wizard-type plugins * Fix several small memory leaks * Clean some logged warnings and potential issues during R backend shutdown * Fixed: Avoid shortcut clash with kate part by removing conflicting (default) shortcuts, automatically - Recommend kate-plugins that are usable in rkward now - Build with QtWebEngine (instead of QtWebKit) on architectures and distributions where it is available (in a new enough version) - Remove no longer available NO_R_XML cmake option 2) [757:757:1023/102606.667121:ERROR:zygote_host_impl_linux.cc(89)] Running as root without --no-sandbox is not supported. See https://crbug.com/638180. https://crbug.com/638180 ========================= Issue 638180: Exit instead of crashing when running as root without --no-sandbox. Reported by ajha@chromium.org on Tue, Aug 16, 2016, 5:46 AM CDT Note: Magic siganture is not showing in Fracas hence logging this via go/chromecrash. Product name: Chrome_Linux Magic Signature: content::ZygoteHostImpl::LaunchZygote -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.opensuse.org/show_bug.cgi?id=1178041 Nick Dordea <ndordea@gmail.com> changed: What |Removed |Added ---------------------------------------------------------------------------- Component|Maintenance |KDE Applications Assignee|maintenance@opensuse.org |opensuse-kde-bugs@opensuse. | |org -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.opensuse.org/show_bug.cgi?id=1178041 Nick Dordea <ndordea@gmail.com> changed: What |Removed |Added ---------------------------------------------------------------------------- Priority|P5 - None |P2 - High -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.opensuse.org/show_bug.cgi?id=1178041 http://bugzilla.opensuse.org/show_bug.cgi?id=1178041#c2 --- Comment #2 from Christophe Giboudeaux <christophe@krop.fr> --- (In reply to Nick Dordea from comment #1)
Hello, I looked to rkward's yast latest log as well as the page mentioned by the failing message [ provided below ].
1. The config log says ..... a) * Support using QWebEngine instead of QtWebKit (still supported) b) Build with QtWebEngine (instead of QtWebKit) on architectures and distributions where it is available (in a new enough version)
Question : How to start rkward and use QtWebKit and NOT the QWebEngine that appears to be a Google/Chromium back door in our machines ?
Note : I do not use Google's Chrome, Chromium because I do not trust them. If opensuse is pushing chromium on the users via rkward I will switch to Rstudio or Rcmndr .
So you would prefer trusting a renderer engine that's unmaintained upstream for years? chromium != chrome -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.opensuse.org/show_bug.cgi?id=1178041 http://bugzilla.opensuse.org/show_bug.cgi?id=1178041#c3 --- Comment #3 from Christophe Giboudeaux <christophe@krop.fr> --- (In reply to Christophe Giboudeaux from comment #2)
So you would prefer trusting a renderer engine that's unmaintained upstream for years?
chromium != chrome
PS: Did you notice you use a gmail.com email address? -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.opensuse.org/show_bug.cgi?id=1178041 http://bugzilla.opensuse.org/show_bug.cgi?id=1178041#c4 --- Comment #4 from Nick Dordea <ndordea@gmail.com> --- Hello Christophe. Your question---So you would prefer trusting a renderer engine that's unmaintained upstream for years ?--- deflects software's failure . This is that old typical/refuge .... user error . It is may fault that I dare to use the new rkward and it crashed ! Answer : rkward worked fine before the renderer was changed. so the change/enhancement affected rkward in a negative way. The failure is due to the new renderer, NOT the unmaintained old renderer! I would prefer that the rkward maintenance/QA to assure software's backward compatibility, that the new "enhanced" software does not fail .... in our case the failing message points to an unsolved chromium issue many years old .... QStandardPaths: XDG_RUNTIME_DIR not set, defaulting to '/tmp/runtime-root' QStandardPaths: XDG_RUNTIME_DIR not set, defaulting to '/tmp/runtime-root' [27199:27199:1022/175843.788358:ERROR:zygote_host_impl_linux.cc(89)] Running as root without --no-sandbox is not supported. See https://crbug.com/638180 What about.... that the "maintenance" considers that if root user uses the new enhanced rkward, the new rkward starts/calls the magic new renderer with --no-sandbox option ? If that is done, the new enhanced rkward will be better that the failing actual one. Anyhow, your approach to rkward failure was enough to switch to rstudio ..... thanks for your "user oriented support" ! Please close this ticket mentioning that .... "as long as the new enhanced rkward is not used by root users, it works beautifully !" ND -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.opensuse.org/show_bug.cgi?id=1178041 Nick Dordea <ndordea@gmail.com> changed: What |Removed |Added ---------------------------------------------------------------------------- CC|ndordea@gmail.com | -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.opensuse.org/show_bug.cgi?id=1178041 http://bugzilla.opensuse.org/show_bug.cgi?id=1178041#c5 Nick Dordea <ndordea@gmail.com> changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |qa-bugs@suse.de Flags| |needinfo?(qa-bugs@suse.de) --- Comment #5 from Nick Dordea <ndordea@gmail.com> --- What about.... that the "maintenance" considers that if root user uses the new enhanced rkward, the new rkward starts/calls the magic new renderer with --no-sandbox option ? If that is done, the new enhanced rkward will be better that the failing actual one. -- You are receiving this mail because: You are on the CC list for the bug.
participants (1)
-
bugzilla_noreply＠suse.com