[plt-scheme] Mysterious bitrot in MrEd

From: Thomas Chust (chust at web.de)
Date: Thu Apr 16 11:25:17 EDT 2009


in the last few months I have observed the following very strange
behaviour of MrEd versions 4.1.3 through 4.1.5 several times both
under MacOS X and Linux both with PowerPC and Intel processors on
several different machines: After running a MrEd application (for
example DrScheme) and terminating it normally, MrEd sometimes fails to
start up correctly ever again (it simply hangs forever when you start
it or produces some random segmentation fault). I finally figured out
that this happens because the native shared library containing the
MrEd backend code actually changes on disk! Of course the problem only
occurs if the user running MrEd has write permissions to that library
file and it happens only once in a while. However it is needless to
say that it spoils my day if it happens :-/

I suspect that the problem is caused by some memory mapping going
ballistic and garbage being written into a segment of the library ;-)
Unfortunately I haven't been able to trigger the problem
deterministically, yet. And I wonder whether I'm going crazy or if
anybody else has ever observed this behaviour as well.

Please let me know if you had any similar experience or if you know
what I'm doing wrong or what may be causing the trouble.


When C++ is your hammer, every problem looks like your thumb.

Posted on the users mailing list.