Generalize the 80-bit long double in <complex> for Clang #1728
Conversation
The previous check works for the EDG-based Intel compiler ONLY, but Clang and GCC both have the -mlong-double-80 options (see https://godbolt.org/z/h7s4Te). This patch generalizes the 80 bit check to work with any compiler that defines the __LDBL_DIG__ macro. There are a couple of other macro options that are equally as opaque (such as __LDBL_MANT_DIG__ and __LDBL_DECIMAL_DIG__), but this seemed the most general. Note that it ALSO works in ICC, as well as all Clang derivatives (such as Intel's DPCPP).
CaseyCarter
left a comment
CaseyCarter
left a comment
There was a problem hiding this comment.
We don't want to invest much time into evaluating changes to semi-supported machinery like this, but since this doesn't increase complexity nor break the intended use case I suppose the change is fine.
|
Congratulations on your first microsoft/STL commit! 🎉 😸 🚀 |
|
@StephanTLavavej : Do you have any idea which release this fix should be included in? Thanks! |
If you search the https://github.com/microsoft/STL/wiki/Changelog for "1728" (the PR number) you'll see the change listed in the "Expected in VS 2019 16.10 Preview 3" section. |
Perfect, thank you so much for the prompt answer! |
3 participants
The previous check works for the EDG-based Intel compiler ONLY, but
Clang and GCC both have the -mlong-double-80 options (see
https://godbolt.org/z/h7s4Te). This patch generalizes the 80 bit check
to work with any compiler that defines the LDBL_DIG macro.
There are a couple of other macro options that are equally as opaque
(such as LDBL_MANT_DIG and LDBL_DECIMAL_DIG), but this seemed
the most general. Note that it ALSO works in ICC, as well as all Clang
derivatives (such as Intel's DPCPP).