Eric Blake wrote: > > No, other reasons are allowed as well. > > Finish reading that paragraph: > > "...if and only if all those error conditions can always be treated > identically to the error conditions as described in this volume of > POSIX.1-2008.
This is a chewing-gum sentence: "can always be treated identically" depends on the developer's judgement. Back to the topic: You want this wording, I think. Committed. 2009-01-24 Bruno Haible <[email protected]> * lib/c-strtod.h (c_strtod, c_strtold): Adjust specification. Reported by Eric Blake. --- lib/c-strtod.h.orig 2009-01-24 15:00:01.000000000 +0100 +++ lib/c-strtod.h 2009-01-24 14:59:14.000000000 +0100 @@ -30,8 +30,8 @@ - In case of underflow, return a value very near to 0 and set errno to ERANGE. - If the string does not start with a number at all, return 0 (and recall - that if ENDPTR != NULL, *ENDPTR is set to NPTR). - - In case of other error, return 0 and set errno, for example to EINVAL - or ENOMEM. */ + that if ENDPTR != NULL, *ENDPTR is set to NPTR), and maybe set errno to + EINVAL. + - In case of other error, return 0 and set errno, for example to ENOMEM. */ extern double c_strtod (char const *nptr, char **endptr); extern long double c_strtold (char const *nptr, char **endptr);
