Mathieu,
I do not understand what you mean in your email.
I am only interested in the cause of the valgrind
message in
> .Call("XXX")
[1] 1
> nd <- 3
> solve(diag(nd) , as.vector(1:nd))
[1] 1 2 3
> .Call("XXX")
==25734== Conditional jump or move depends on uninitialised value(s)
==25734== at 0x401620C: (within /lib/ld-2.8.so)
==25734== by 0x42825C3: (within /lib/libc-2.8.so)
==25734== by 0x4282989: _dl_sym (in /lib/libc-2.8.so)
==25734== by 0x417ADE7: (within /lib/libdl-2.8.so)
==25734== by 0x400DE25: (within /lib/ld-2.8.so)
==25734== by 0x417B0DB: (within /lib/libdl-2.8.so)
==25734== by 0x417AD72: dlsym (in /lib/libdl-2.8.so)
==25734== by 0x811012D: R_dlsym (Rdynload.c:787)
==25734== by 0x8110A98: R_FindSymbol (Rdynload.c:836)
==25734== by 0x816E41C: resolveNativeRoutine (dotcode.c:240)
==25734== by 0x817473A: do_dotcall (dotcode.c:825)
==25734== by 0x8194C12: Rf_eval (eval.c:487)
==25734== by 0x805B163: Rf_ReplIteration (main.c:257)
==25734== by 0x805B3E7: R_ReplConsole (main.c:306)
==25734== by 0x805B6D7: run_Rmainloop (main.c:966)
==25734== by 0x8058F80: main (Rmain.c:33)
==25734==
==25734== Conditional jump or move depends on uninitialised value(s)
==25734== at 0x4016217: (within /lib/ld-2.8.so)
==25734== by 0x42825C3: (within /lib/libc-2.8.so)
==25734== by 0x4282989: _dl_sym (in /lib/libc-2.8.so)
==25734== by 0x417ADE7: (within /lib/libdl-2.8.so)
==25734== by 0x400DE25: (within /lib/ld-2.8.so)
==25734== by 0x417B0DB: (within /lib/libdl-2.8.so)
==25734== by 0x417AD72: dlsym (in /lib/libdl-2.8.so)
==25734== by 0x811012D: R_dlsym (Rdynload.c:787)
==25734== by 0x8110A98: R_FindSymbol (Rdynload.c:836)
==25734== by 0x816E41C: resolveNativeRoutine (dotcode.c:240)
==25734== by 0x817473A: do_dotcall (dotcode.c:825)
==25734== by 0x8194C12: Rf_eval (eval.c:487)
==25734== by 0x805B163: Rf_ReplIteration (main.c:257)
==25734== by 0x805B3E7: R_ReplConsole (main.c:306)
==25734== by 0x805B6D7: run_Rmainloop (main.c:966)
==25734== by 0x8058F80: main (Rmain.c:33)
with the code originally given. Even if it is a problem
of the library ld, there might be interest by R to have
the problem detected.
Martin
______________________________________________
R-devel@r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/r-devel
