Bug 96834 - matlab (segmentation fault)
Summary: matlab (segmentation fault)
Alias: None
Product: XQuartz
Classification: Unclassified
Component: New Bugs (show other bugs)
Version: 2.7.9 (xserver-1.17.4)
Hardware: Other All
: medium normal
Assignee: Jeremy Huddleston Sequoia
QA Contact: Jeremy Huddleston Sequoia
Depends on:
Reported: 2016-07-06 19:39 UTC by bbsx
Modified: 2019-05-23 18:30 UTC (History)
0 users

See Also:
i915 platform:
i915 features:


Description bbsx 2016-07-06 19:39:39 UTC
The new version of the XQuartz caused the matlab segmentation fault (core dumped) when using pcolor or any other 2D plots. This even happened in the beta version (2.7.10).  the version 2.7.8 is working.
Comment 1 Larry 2016-07-21 19:05:37 UTC
I have this bug as well, and wanted to add a couple more relevant details.

1) I'm ssh -X'ing into a linux machine running CentOS 5.11 from an iMac OS10.10.5, xquartz 2.7.9. If I try logging into a different machine running more recent versions of CentOS, (6.X, 7.X) from this iMac, this segmentation fault does not occur. Logging into other remote machines running CentOS 5.11 from this iMac also has this error. 

2) This exact same error also occurs with cygwin running on windows when logging into my remote machine running CentOS 5.11. I've known about this for some time and assumed it was an error in cygwin or windows, but it seems likely that these issues are related. Using cygwin to log into another linux machine running CentOS 6.6 does not produce this error.

3) Logging into this linux machine from other linux machines running CentOS 6.X and 7.X does not produce this error.

4) A minimal reproducible example: ssh -X remote.machine_running_centos5.11 , matlab , pcolor(rand(20,20))

Oddly, doing pcolor(rand(10,10)) does not produce the error. The only difference between pcolor(rand(20,20)) and pcolor(rand(10,10)) is the number of tiles (100 vs 400) that matlab is displaying in the figure. 

5) This error occurs for all versions of matlab that will run on CentOS 5.11 without complaining (up through 2013a)
Thus, it seems like something in the CentOS 5.11 distribution is the culprit which has been accounted for in newer distributions. My solution is to upgrade CentOS on my remote machine since newer CentOS's play nice with xquartz. Because of the configuration of my remote machine, this is a major headache which has left me dead in the water for a week while I backup 70TB raid and pray that the new OS recognizes drives and their contents.

Comment 2 Jeremy Huddleston Sequoia 2016-09-18 22:49:03 UTC
So to be clear, matlab is the process crashing, not XQuartz.  And Matlab isn't even on the same system as XQuartz.  You're sshing into a Linux box to run it?

The crash sounds like possibly a bug in Matlab, but it's certainly possible that it's triggered by a change in XQuartz.  Can you give more details?  Can you give a crash log?
Comment 3 GitLab Migration User 2019-05-23 18:30:58 UTC
-- GitLab Migration Automatic Message --

This bug has been migrated to freedesktop.org's GitLab instance and has been closed from further activity.

You can subscribe and participate further through the new bug through this link to our GitLab instance: https://gitlab.freedesktop.org/xorg/xserver/issues/766.

Use of freedesktop.org services, including Bugzilla, is subject to our Code of Conduct. How we collect and use information is described in our Privacy Policy.