Simon Josefsson wrote: > Any obvious reason we shouldn't create a time.h with the contents of > the current time_r.h, using the same logic we use to replace string.h? > > It would be nice to avoid '#include "time_r.h"' and just do '#include > <time.h>' in my code.
The full TODO list for this topic is below. But I would find it good to wait a month or so, to stabilize the idioms used in <string.h> and string_h.m4. Not everyone has yet tried to compile his project with -DGNULIB_POSIXCHECK and to see whether the warnings and errors are useful. Also Paul has plans to simplify these idioms (although I don't agree that it's possible to simplify them without sacrificing either correctness, quality of error messages, or maintainability). Bruno <inttypes.h> strtoimax strtoumax <limits.h> size_max.h <math.h> mathl.h <netdb.h> getaddrinfo.h <search.h> tsearch.h <signal.h> sigprocmask.h <stdio.h> snprintf.h vsnprintf.h <stdlib.h> exit.h mkdtemp.h mkstemp.h <time.h> time_r.h timespec.h strptime.h <unistd.h> chown dup2 fchdir ftruncate getcwd.h getlogin_r.h getsubopt.h lstat readlink <wchar.h> wcwidth.h <arpa/inet.h> inet_ntop inet_pton
