GNOME Bugzilla – Bug 109442
Placeholder functions should be updated when the true function becomes available.
Last modified: 2013-01-11 01:06:44 UTC
Package: Gnumeric Severity: normal Version: 1.1.17 Synopsis: Crashed while exiting after activating all the plugins Bugzilla-Product: Gnumeric Bugzilla-Component: General BugBuddy-GnomeVersion: 2.0 (2.2.0.1) Description: Launched gnumeric myfile.xl got on the console: [chema@localhost expenses]$ gnumeric mexican_office_expenses_2003.xls ** (gnumeric:16159): WARNING **: Configured default font not available, trying f allback... ** (gnumeric:16159): WARNING **: font_name == NULL, Using 9.000000 ** (gnumeric:16159): WARNING **: font_name == NULL, Using 9.000000 mexican_office_expenses_2003.xls Excel 95 (gnumeric:16159): Gtk-WARNING **: Unable to locate theme engine in module_path: "bluecurve", (gnumeric:16159): gnumeric:read-WARNING **: EXCEL: color index (110) is out of r ange (0..56). Defaulting to black (gnumeric:16159): gnumeric:read-WARNING **: EXCEL: color index (110) is out of r ange (0..56). Defaulting to black (gnumeric:16159): gnumeric:read-WARNING **: EXCEL: color index (110) is out of r ange (0..56). Defaulting to black (gnumeric:16159): gnumeric:read-WARNING **: EXCEL: color index (110) is out of r ange (0..56). Defaulting to black ** (gnumeric:16159): WARNING **: Unknown Builtin function : AND ** (gnumeric:16159): WARNING **: Unknown Builtin function : IF ** (gnumeric:16159): WARNING **: Unknown Builtin function : OR Went to plugins and activated all plugins, exited gnumeric to try to repload it, it crashed while exiting Debugging Information: Backtrace was generated from '/home/chema/cvs/gnome2/bin/gnumeric' [New Thread 8192 (LWP 16159)] 0x420ae169 in wait4 () from /lib/i686/libc.so.6
+ Trace 35357
Thread 1 (Thread 8192 (LWP 16159))
------- Bug moved to this database by unknown@bugzilla.gnome.org 2003-03-28 14:11 ------- Reassigning to the default owner of the component, jody@gnome.org.
Also got: ** (gnumeric:16159): WARNING **: (leak) Symbol [and] redefined. ** (gnumeric:16159): WARNING **: (leak) Symbol [or] redefined. ** (gnumeric:16159): WARNING **: (leak) Symbol [if] redefined.
FWIW, this appears to be a unique stack trace, according to the simple-dup-finder. Marking severity->critical (it's a crasher), adding bugsquad keyword, and marking as new.
Confirming leak and crash. (The leaks are unimportant.)
Crash has been fixed ==> downgrading. Basically the symbol gets defined twice: once as a placeholder and once as a function. On exit, the two owners both try to delete the function entry. Boom! It isn't clear to me what we should do here, apart from not crashing. Therefore leaving open. Elijah Newren: this was never really critical since the crash was on exit with no data loss possible.
Tried to reproduce this. Disabled the Logic functions plugin, exited, gnumeric, and started gnumeric with a spreadsheet which used logic functions. Enabled Logic functions. Pressed F9. Got #NAME in the cells which depend on logic functions. Is this correct? Or is enabling the plugin after reading the spreadsheet supposed to work?
Retitling - open bugs with crash in the title should preferably be crashers
What exactly is left in this bug?
What's left? "Doctor, It hurts when I hit my head against the wall."
1. Disable Math. 2. Enter =sin(42) into A2. --> Observe #NAME? 3. Enable Math. --> No change 4. Recalc. --> Still no change. We should somehow upgrade the placeholder from (2) to a proper function.
In view of comment #9 I am retitling.
Created attachment 233134 [details] [review] Tentative patch I believe this fixes the issue by upgrading placeholders to stubs when the plugin get activated. It's a little more involved than I am comfortable with right now.
Created attachment 233198 [details] [review] Updated patch Patch updated to match current HEAD.
This problem has been fixed in our software repository. The fix will go into the next software release. Thank you for your bug report.