Hi Paul, > The testcase has a dependence on libubsan.so, which my LD_LIBRARY_PATH > does not seem to be able to resolve. It therefore fails in the > regression test. When I point to ~/lib64, I get the message: > > /svn/trunk/gcc/testsuite/gfortran.dg/class_allocate_18.f90:8: runtime > error: signed integer overflow: 214358001592 * 214358001560 cannot be > represented in type 'integer(kind=8)'
of course I tested it but did not see any problem on my machine (Ubuntu 14.10, x86_64-unknown-linux-gnu). In fact class_allocate_18.f90 seems to be the only testcase in gfortran.dg/ which uses any -fsanitize options, but there are plenty in other sections (mostly c-c++-common/ubsan/ and g++.dg/ubsan/). Could it be a problem with your local setup or a dejagnu issue that only applears on certain platforms? I have no idea how to fix this. Do you have any suggestion? Cheers, Janus > On 26 January 2015 at 15:57, Janus Weil <ja...@gcc.gnu.org> wrote: >> 2015-01-24 18:18 GMT+01:00 Tobias Burnus <bur...@net-b.de>: >>>> this is a second patch dealing with finalization-related regressions, >>>> [...] >>>> This patch fixes an invalid memory reference inside the finalizer >>>> routine (at runtime), which apparently was caused by dereferencing a >>>> pointer without checking if it's NULL. I simply insert a call to >>>> ASSOCIATED. >>>> [...] >>>> >>>> This one regtests cleanly on x86_64-unknown-linux-gnu. Ok for trunk and >>>> 4.9? >>> >>> >>> OK. Thanks for the patch! >> >> Thanks for the review, Tobias. Committed to trunk as r220125. Will >> backport to 4.9 soon. >> >> Cheers, >> Janus > > > > -- > Outside of a dog, a book is a man's best friend. Inside of a dog it's > too dark to read. > > Groucho Marx
