gdb core dump

Thomas Lumley
Fri, 16 Oct 1998 11:04:44 -0700 (PDT)

On Fri, 16 Oct 1998, Paul Gilbert wrote:

> >gorn% R -d gdb
> ...
> >(gdb) backtrace
> Thanks. Now I guess that gets me to the hard part. Does the following point out
> a problem in R 0.62.3?
<great big long trace>

Segfaults in the internal R code seem to usually be due to an unprotected
pointer being garbage collected (which is why they are unpredictable --
you don't know when garbage collection will be triggered).  This looks
like one of these.

Unfortunately the only reliable way to find out exactly where it is is to
get a reproducible segfault.  The development version can be set up to
garbage collect at every opportunity, which helps make the problem
reproducible, and then it's possible to track down when the offending
pointer gets mangled and see what wasn't protected. As you can imagine,
it's somewhat time consuming. Quite a few of these bugs have been found
and fixed for 0.63.


r-devel mailing list -- Read
Send "info", "help", or "[un]subscribe"
(in the "body", not the subject !)  To: