1Code in this sub-directory should follow the GNU Coding Standard, but it is
2not expected to be upstreamed into glibc without modification, so
3glibc-specific conventions need not be followed.
4
5The requirements for portable code apply to non-portable code with the
6following differences:
7
8
91. Worst-case ULP error should be encoded in filenames (e.g. sin_u35.c). There
10   are no specific restrictions on acceptable ULP error, but if functions
11   provide significantly less accuracy than portable equivalents then a clear
12   justification for inclusion should be stated in comments at the top of the
13   source file. Error bounds of the approximation should be clearly documented
14   in comments.
15
162. Functions are assumed to support round-to-nearest mode by default, unless
17   stated; other rounding modes are not required to be provided.
18
193. Handling of special cases may be relaxed for vector functions. Checking
20   whether each vector lane contains special values such as NaN, Inf or
21   denormal numbers can prove too costly for vector functions. This is often
22   not required since vector functions are typically used along with aggressive
23   compiler optimization flags.
24