Skip to content

segfaults while editing code scripts on 10.9 mavericks

2 messages · Alexander Guczka, Dan Flynn

#
Dear R-SIG,

I have met some Problems since I updated to OSX Mavericks. From time to time does R crash with a "segfault" if I understand it correctly. I think that some of the crashes are associated with the R-GUI.

One recurring (but not completely reproducible) crash happens, when I edit code-documents/scripts and try to search and replace more than one element at a time. I tried to reduce the amount of Elements to "replace all" but with no effect. Some days ago I switched to another code editor because of that.

What happens? R stops working the second I click "replace all" and shows segfault ?  "memory not mapped"

The other editor (sub-etha-edit) warns me that in R-GUI produced .R-scripts have different/not consistent line ending formats and suggest changing all to linux-style LF. Since changing one of my bigger scripts in sub-etha-edit "replace all" works again in R-GUI. Maybe this information helps. 

I had a lot of crashes with 3.0.1 on Mavericks, changed to 3.0.2 which works really good most of the time. Because of the crash problem mentioned above I tried updating to a patched version of 3.0.2 with no significant effect (pun intended). 

current versions:
R version 3.0.2 Patched (2013-11-11 r64202) -- "Frisbee Sailing"
R.app GUI 1.62 (6596) x86_64-apple-darwin10.8.0  


Best wishes,
Alexander Guczka
7 days later
#
Following hints here for debugged version of R.app is working for me.

http://citizen-statistician.org/2013/10/27/warning-mac-os-10-9-mavericks-and-r-dont-play-nicely/

I had exactly the same problem, as well as more frequent crashes which are 
not obviously replicable.

Dan
On Friday, November 15, 2013 7:12:08 AM UTC-5, Alexander Guczka wrote:
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://stat.ethz.ch/pipermail/r-sig-mac/attachments/20131122/d3e0da07/attachment-0002.html>