[Rd] Reproducible Rterm crash. (PR#7072)

dmurdoch at pair.com dmurdoch at pair.com
Fri Jul 9 21:57:36 CEST 2004


On Fri,  9 Jul 2004 21:28:45 +0200 (CEST), rmh at temple.edu wrote :

>Reproducible Rterm crash.
>
>I drew a complicated graph, then resized the graphics window.
>
>R reported
>> Insufficient memory for resize. Killing device
>and then killed the device.  It also killed itself.

I just tried your code (after loading lattice so it would run), and
didn't crash.  I think you just ran out of memory.  We try to recover
gracefully from situations like that, but it's hard, and especially
hard to track down when we can't reproduce it.

All I can suggest (and I know it's not much of a suggestion) is that
you recompile R with debug information, then run it under gdb and see
if you can figure out where we've missed some checks.

It's possible that Emacs has something to do with this, but hard to
know.  You might find that running Rterm outside of Emacs works fine
just because more memory is available.

Duncan Murdoch



More information about the R-devel mailing list