[R] R crashing with a segmentation fault: how to locate the cause

[Eric Berger]

I have a littler script which is crashing with a segmentation fault.
I tried to find out why by running it through valgrind, which produced the
output below.
I am not sure how to proceed from here (other than binary search with print
statements).
Any help would be appreciated.

Thanks,
Eric

==12589== Invalid read of size 1
==12589==    at 0x4C2F1B1: strcmp (in
/usr/lib/valgrind/vgpreload_memcheck-amd64-linux.so)
==12589==    by 0x4F71AC1: Rf_inherits (in /usr/lib/R/lib/libR.so)
==12589==    by 0x11AFED3A: dplyr::subset_visitor_vector(SEXPREC*)
(subset_visitor_impl.h:51)
==12589==    by 0x11AFF58C: dplyr::subset_visitor(SEXPREC*,
dplyr::SymbolString const&) (subset_visitor_impl.h:21)
==12589==    by 0x11AFEC18:
dplyr::DataFrameSubsetVisitors::DataFrameSubsetVisitors(Rcpp::DataFrame_Impl<Rcpp::PreserveStorage>
const&) (DataFrameSubsetVisitors.h:33)
==12589==    by 0x11B1F773: subset<Rcpp::Vector<10, Rcpp::PreserveStorage>
> (DataFrameSubsetVisitors.h:120)
==12589==    by 0x11B1F773:
filter_ungrouped(Rcpp::DataFrame_Impl<Rcpp::PreserveStorage>,
dplyr::NamedQuosure const&) (filter.cpp:89)
==12589==    by 0x11B1FD89:
filter_impl(Rcpp::DataFrame_Impl<Rcpp::PreserveStorage>,
dplyr::NamedQuosure) (filter.cpp:106)
==12589==    by 0x11AD45CB: _dplyr_filter_impl (RcppExports.cpp:192)
==12589==    by 0x4F0DBAC: ??? (in /usr/lib/R/lib/libR.so)
==12589==    by 0x4F50E50: Rf_eval (in /usr/lib/R/lib/libR.so)
==12589==    by 0x4F534FF: ??? (in /usr/lib/R/lib/libR.so)
==12589==    by 0x4F50C28: Rf_eval (in /usr/lib/R/lib/libR.so)
==12589==  Address 0x2c is not stack'd, malloc'd or (recently) free'd
==12589==
==12589==
==12589== Process terminating with default action of signal 11 (SIGSEGV)
==12589==  Access not within mapped region at address 0x2C
==12589==    at 0x4C2F1B1: strcmp (in
/usr/lib/valgrind/vgpreload_memcheck-amd64-linux.so)
==12589==    by 0x4F71AC1: Rf_inherits (in /usr/lib/R/lib/libR.so)
==12589==    by 0x11AFED3A: dplyr::subset_visitor_vector(SEXPREC*)
(subset_visitor_impl.h:51)
==12589==    by 0x11AFF58C: dplyr::subset_visitor(SEXPREC*,
dplyr::SymbolString const&) (subset_visitor_impl.h:21)
==12589==    by 0x11AFEC18:
dplyr::DataFrameSubsetVisitors::DataFrameSubsetVisitors(Rcpp::DataFrame_Impl<Rcpp::PreserveStorage>
const&) (DataFrameSubsetVisitors.h:33)
==12589==    by 0x11B1F773: subset<Rcpp::Vector<10, Rcpp::PreserveStorage>
> (DataFrameSubsetVisitors.h:120)
==12589==    by 0x11B1F773:
filter_ungrouped(Rcpp::DataFrame_Impl<Rcpp::PreserveStorage>,
dplyr::NamedQuosure const&) (filter.cpp:89)
==12589==    by 0x11B1FD89:
filter_impl(Rcpp::DataFrame_Impl<Rcpp::PreserveStorage>,
dplyr::NamedQuosure) (filter.cpp:106)
==12589==    by 0x11AD45CB: _dplyr_filter_impl (RcppExports.cpp:192)
==12589==    by 0x4F0DBAC: ??? (in /usr/lib/R/lib/libR.so)
==12589==    by 0x4F50E50: Rf_eval (in /usr/lib/R/lib/libR.so)
==12589==    by 0x4F534FF: ??? (in /usr/lib/R/lib/libR.so)
==12589==    by 0x4F50C28: Rf_eval (in /usr/lib/R/lib/libR.so)
==12589==  If you believe this happened as a result of a stack
==12589==  overflow in your program's main thread (unlikely but
==12589==  possible), you can try to increase the size of the
==12589==  main thread stack using the --main-stacksize= flag.
==12589==  The main thread stack size used in this run was 8388608.
==12589==
==12589== HEAP SUMMARY:
==12589==     in use at exit: 224,025,975 bytes in 111,522 blocks
==12589==   total heap usage: 1,104,400 allocs, 992,878 frees, 625,925,991
bytes allocated
==12589==
==12589== LEAK SUMMARY:
==12589==    definitely lost: 0 bytes in 0 blocks
==12589==    indirectly lost: 0 bytes in 0 blocks
==12589==      possibly lost: 18,724 bytes in 47 blocks
==12589==    still reachable: 224,007,251 bytes in 111,475 blocks
==12589==         suppressed: 0 bytes in 0 blocks
==12589== Rerun with --leak-check=full to see details of leaked memory
==12589==
==12589== For counts of detected and suppressed errors, rerun with: -v
==12589== ERROR SUMMARY: 1 errors from 1 contexts (suppressed: 0 from 0)
Segmentation fault (core dumped)

	[[alternative HTML version deleted]]