https://gcc.gnu.org/bugzilla/show_bug.cgi?id=77735
Martin Sebor <msebor at gcc dot gnu.org> changed:
What |Removed |Added
----------------------------------------------------------------------------
Keywords| |diagnostic
Status|UNCONFIRMED |ASSIGNED
Last reconfirmed| |2016-09-28
Component|testsuite |middle-end
Assignee|unassigned at gcc dot gnu.org |msebor at gcc dot
gnu.org
Ever confirmed|0 |1
Known to fail| |7.0
--- Comment #1 from Martin Sebor <msebor at gcc dot gnu.org> ---
I have reproduced the test failure with a cross-compiler for
hppa-unknown-linux-gnu. The problem is due to the c_strlen function defined in
builtins.c accepting wide strings but not handling them correctly (in this
case, returning zero for L"1"). The problem doesn't show up in the tests on
little-endian targets because the longest strings used by the wide character
tests are 1 wide character long and the first character is in the least
significant byte. Let me look into enhancing c_strlen to handle wide character
strings if that's doable without too much effort. If not, I'll have to disable
the -Wformat-length warning for wide strings.
