This is the mail archive of the
libstdc++@gcc.gnu.org
mailing list for the libstdc++ project.
Re: [v3] C99 complex math vs. std::complex
Benjamin Kosnik wrote:
While you are at it, before I forget: a few days ago I noticed that a
few 26_numerics/cmath testcases are protected by _GLIBCXX_USE_C99 (and I
added one ;) Should be changed, more correctly, to just
_GLIBCXX_USE_C99_MATH?
Maybe.
I don't know, off hand. I don't know what we should be attempting with
_GLIBCXX_USE_C99 at the moment, and without context. Can you elaborate?
Oh, yes, it's not a big issue, anyway: as I understand the involved
autoconf test,
basically _GLIBCXX_USE_C99 is stronger than _GLIBCXX_USE_C99_MATH,
in that is defined only if various separate C99 features are available
(*). But
those testcases are actually testing only MATH features (things like
fpclassify
and co)
Paolo.
(*) By the way, how _GLIBCXX_USE_C99_COMPLEX_MATH relates to
_GLIBCXX_USE_C99? The new autoconf test should perhaps be ANDed with
the existing ones (in GLIBCXX_ENABLE_C99) thus making
_GLIBCXX_USE_C99 more robust?