..\..\../boost/test/impl/debug.ipp:211:21: warning: private field 'm_binary_path_buff' is not used [-Wunused-private-field] char m_binary_path_buff[500+1]; // !! ?? ^ I'm not sure what the purpose of this member is, but according to Clang, it's not used anywhere.
Peter Dimov <lists <at> pdimov.com> writes:
..\..\../boost/test/impl/debug.ipp:211:21: warning: private field 'm_binary_path_buff' is not used [-Wunused-private-field] char m_binary_path_buff[500+1]; // !! ?? ^
I'm not sure what the purpose of this member is, but according to Clang, it's not used anywhere.
It is used. Clang is mistaken. Gennadiy
On 15 Jan 2015, at 23:34, Gennadiy Rozental <rogeeff@gmail.com> wrote:
Peter Dimov <lists <at> pdimov.com> writes:
..\..\../boost/test/impl/debug.ipp:211:21: warning: private field 'm_binary_path_buff' is not used [-Wunused-private-field] char m_binary_path_buff[500+1]; // !! ?? ^
I'm not sure what the purpose of this member is, but according to Clang, it's not used anywhere.
It is used. Clang is mistaken.
It is defined but unused if you're Unix but not Linux or Sun, so maybe Cygwin can give you this.
Pete Bartlett <pete <at> pcbartlett.com> writes:
It is defined but unused if you're Unix but not Linux or Sun, so maybe Cygwin can give you this.
I'd rather not overcomplicate the code because of this.
I guess you're concerned about overcomplicating code just to suppress some warnings. What I was thinking of really though is that the process_info class is currently silently doing nothing (i.e. populating parent_pid, binary_name, binary_path with zero or blank) in the Cygwin case, but would probably work (I haven't tested) if what is currently the Linux-only code path was used for Cygwin. Sure the code still isn't getting any less complicated, but at least you'd be adding functionality! And good to know that clang isn't buggy here after all :-) Pete
participants (4)
-
Gennadiy Rozental -
Pete bartlett -
Pete Bartlett
-
Peter Dimov