On 11/1/24 17:40, Ivan Krylov wrote:
On Fri, 1 Nov 2024 09:56:01 +0100
Tomas Kalibera <tomas.kalib...@gmail.com> wrote:
isn't there a finer-grained way to suppress reports just
of this particular issue, ideally at finer granularity than a
function?
The line saying "function:do_dotCode" in the suppression file will
suppress errors just for mismatched function pointer types inside
do_dotCode(), because "function" is the name of this check [1]. Here's
do_dotCode with a different error added:
bin/R -q -s -e 'library(survival); survdiff(Surv(futime, fustat) ~
rx,data=ovarian)'
dotcode.c:1958:9: runtime error: 1e+100 is outside the range of
representable values of type 'int'
SUMMARY: UndefinedBehaviorSanitizer: undefined-behavior dotcode.c:1958:9
dotcode.c:2012:2: runtime error: call to function survdiff2 through
pointer to incorrect function type
'void (*)(void *, void *, void *, void *, void *, void *, void *, void
*, void *, void *, void *, void
*, void *)'
/tmp/RtmpVPlRnI/R.INSTALL65eb26c0356f/survival/src/survdiff2.c:10:
note: survdiff2 defined here
SUMMARY: UndefinedBehaviorSanitizer: undefined-behavior dotcode.c:2012:2
UBSAN_OPTIONS="suppressions=$(realpath dotcode.supp)" \
bin/R -q -s -e 'library(survival); survdiff(Surv(futime, fustat) ~
rx,data=ovarian)'
dotcode.c:1958:9: runtime error: 1e+100 is outside the range of
representable values of type 'int'
I don't think UBSan can suppress errors in a more precise manner. The
documentation only mentions files, functions ans libraries. Best I
could find was a 2020 comment [2] suggesting to separate the offending
code into a function and let the optimiser inline it.
I feared that might be the case, thanks for looking this up,
Tomas
______________________________________________
R-package-devel@r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/r-package-devel
