Resolve compiler warnings emited by clang #103
Conversation
[Why] If compiled with clang and -Wconversion warnings enable, the compiler will complain: > gul14/span.h:218:42: warning: implicit conversion changes signedness: 'long' to 'std::size_t' (aka 'unsigned long') [-Wsign-conversion] > 218 | : storage_(first_elem, last_elem - first_elem) > | ~~~~~~~~ ~~~~~~~~~~^~~~~~~~~~~~ Signed-off-by: Soeren Grunewald <soeren.grunewald@desy.de>
Signed-off-by: Soeren Grunewald <soeren.grunewald@desy.de>
[Why] If compiled with clang, the compiler complains: > include/gul14/date.h:912:41: warning: identifier '_d' preceded by whitespace in a literal operator declaration is deprecated [-Wdeprecated-literal-operator] > 912 | constexpr gul14::date::day operator "" _d(unsigned long long d) noexcept; > | ~~~~~~~~~~~~^~ > | operator""_d > include/gul14/date.h:913:41: warning: identifier '_y' preceded by whitespace in a literal operator declaration is deprecated [-Wdeprecated-literal-operator] > 913 | constexpr gul14::date::year operator "" _y(unsigned long long y) noexcept; > | ~~~~~~~~~~~~^~ > | operator""_y > include/gul14/date.h:1813:13: warning: identifier '_d' preceded by whitespace in a literal operator declaration is deprecated [-Wdeprecated-literal-operator] > 1813 | operator "" _d(unsigned long long d) noexcept > | ~~~~~~~~~~~~^~ > | operator""_d > include/gul14/date.h:1820:13: warning: identifier '_y' preceded by whitespace in a literal operator declaration is deprecated [-Wdeprecated-literal-operator] > 1820 | operator "" _y(unsigned long long y) noexcept > | ~~~~~~~~~~~~^~ > | operator""_y Signed-off-by: Soeren Grunewald <soeren.grunewald@desy.de>
|
My understanding is that GUL14 has been sunsetted and will be removed soon. What is the timeline, @alt-graph |
|
I mean, I see changes in 'date' and 'sliding buffer', maybe the fixes should be applied to GUL17. forward-porting the fixes from GUL14 you did here to GUL17 is kind of ... unneeded work, if you could fix GUL17 directly instead. |
I guess we'll keep GUL14 around for compatibility for a long time, so I would give its maintenance a low (but nonzero) priority. New features should primarily go into GUL17. If you ask for a timeline, we can probably declare GUL14 as unmaintained in a year or two, meaning we leave it on Github but leave a "you're on your own" sign. |
[Why] If compiled with clang, the compiler complains: > include/gul14/string_view.h:576:30: warning: implicit conversion changes signedness: 'std::size_t' (aka 'unsigned long') to 'streamsize' (aka 'long') [-Wsign-conversion] > 576 | os.write(fill_chars, n); > | ~~~~~ ^ Signed-off-by: Soeren Grunewald <soeren.grunewald@desy.de>
[Why] If compiled with clang, which complains about: > include/gul14/string_util.h:154:27: warning: implicit conversion changes signedness: 'typename iterator_traits<unsigned char *>::difference_type' (aka 'long') to 'const std::size_t' (aka 'const unsigned long') [-Wsign-conversion] > 154 | const std::size_t n = std::distance(begin, end); > | ~ ^~~~~~~~~~~~~~~~~~~~~~~~~ Signed-off-by: Soeren Grunewald <soeren.grunewald@desy.de>
Signed-off-by: Soeren Grunewald <soeren.grunewald@desy.de>
Signed-off-by: Soeren Grunewald <soeren.grunewald@desy.de>
soerengrunewald
force-pushed
the
bugfix/resolve-clang-compiler-warnings
branch
from
56588c2 to
7c20d26
Compare
|
@soerengrunewald: I guess there is a higher chance that someone will review this if you break it apart into smaller pieces. Everybody can find five minutes for reviewing, but >15 for a library that we are slowly phasing out is a hard sell. |
Why
There are some really annoying warnings which are spammed if compiling with clang.
Since we use gul includes all over, every include will emit these warnings.