Ask Your Question

writer : critical error reported but functionality is ok

asked 2018-08-21 01:48:01 +0100

carnish gravatar image

updated 2018-08-21 07:45:48 +0100

ebot gravatar image


 Build ID: 1:6.0.3-0ubuntu1
  CPU threads: 2; OS: Linux 4.15; UI render: default; VCL: gtk3; 
  Locale: en-CA (en_CA.UTF-8); Calc: group

Starting writer from the command line, I see multiple instances of the error below.. Application behavior seems okay, but I'd like to have a clean start. Is there a config I can set, library install/update?

soffice:17056): Gtk-CRITICAL **: 15:47:37.552: gtk_widget_queue_draw_area: assertion 'height >= 0' failed
edit retag flag offensive close merge delete

1 Answer

Sort by » oldest newest most voted

answered 2018-08-21 08:38:04 +0100

ajlittoz gravatar image

This does not happen only with LO. Message comes from one of the various support libraries used by an application.

An application rarely reinvents the wheel, it prefers to rely on APIs provided by a vast community of developers. This is particularly the case for graphics (shown message above is related to GTK, the Gimp toolkit) as developing a reliable and efficient environment is a daunting task, even more than developing a useful document processor.

The problem here is projects advance at their own pace and the binary components are released independently, meaning the version numbers evolve without coordination. Changes may be externally visible in the API requiring dependent projects to adapt (making the new kernel module incompatible with application current version) or merely internal, such as algorithm optimisation/change or more checks on pre/post conditions, input data validity, fallback behaviour, ...

It looks like this is the case (height is negative). I suppose GTK protects itself by "capping" the value to 0 and everything proceeds normally.

You see this message because you launched LO from the command line where stderr is associated with the console. When you launch LO from the GUI (double-click on an icon), stderr is associated with /dev/null and you see none of the library error messages.

If the error was really fatal, the application would either have been aborted by the OS (e.g. a segment fault) or the application itself would have reported the library error with graphical notification and stop.

edit flag offensive delete link more


Thank you for this explanation. I very much appreciate it. I will add a logging redirect to an alias. I like to see the output of the command because it sometimes points to misconfiguration. But it is not usually so chatty.

carnish gravatar imagecarnish ( 2018-08-21 14:58:30 +0100 )edit

You're welcome.

In the vast majority of cases, the "interesting" message is completely lost in a whirl of "noise". Sometimes applications (and the kernel does it) offer the choice to limit messages to a range of severity levels. But here, we're at a cross-road between numerous cooperating libraries. Difficult to filter a specific one.

ajlittoz gravatar imageajlittoz ( 2018-08-21 15:44:59 +0100 )edit
Login/Signup to Answer

Question Tools

1 follower


Asked: 2018-08-21 01:48:01 +0100

Seen: 45 times

Last updated: Aug 21 '18