https://gcc.gnu.org/bugzilla/show_bug.cgi?id=68594
Dominique d'Humieres <dominiq at lps dot ens.fr> changed:
What |Removed |Added
----------------------------------------------------------------------------
Status|UNCONFIRMED |WAITING
Last reconfirmed| |2015-11-28
CC| |jvdelisle at gcc dot gnu.org
Summary|[regression] libfortran |libfortran WRITE() ERR
|WRITE() ERR option causing |option causing segfault
|segfault when used twice |when used twice
Ever confirmed|0 |1
--- Comment #1 from Dominique d'Humieres <dominiq at lps dot ens.fr> ---
All the tests give the expected output at run time with 5.2.0, 5.2.1 (r231024),
and trunk (6.0). With trunk (6.0) the run time errors are followed by a
backtrace:
At line 4 of file pr68594_2.f90 (unit = 6, file = 'stdout')
Fortran runtime error: Expected INTEGER for item 1 in formatted transfer, got
REAL
(I8)
^
Error termination. Backtrace:
...
which disappears if the tests are compiled with -fno-backtrace.
With 4.9.3, the first test gives at run time
Program received signal SIGSEGV: Segmentation fault - invalid memory reference.
Backtrace for this error:
...
while the three others give the expected outputs.
With 4.8.5, the third test gives at run time
Program received signal SIGSEGV: Segmentation fault - invalid memory reference.
Backtrace for this error:
...
while the three others give the expected outputs.
AFAICT this has been fixed between revisions r212778 (2014-07-18, segfault) and
r21296O (2014-07-23, no segfault), likely r212875 (pr61632). I am not sure this
worth the back port and IMO this PR can be closed as FIXED.
