[RFC PATCH 00/52] Make GLIBC Y2038-proof

[Paul Eggert]

On 09/08/2017 09:43 AM, Joseph Myers wrote:
> (I think this use of long is*less*  of an issue than e.g. printf
> returning int, since you can legitimately print more characters than fit
> in int.)

printf is obviously more-important since it's used all over the place 
and is more likely to run into the arbitrary int limit. That being said, 
in practice GNU Coreutils, GNU Emacs, and other application code that I 
help maintain long ago stopped assuming the POSIX requirement that 
tv_nsec must be of type 'long', precisely because we want to be portable 
to x32. So from my point of view this ship already sailed and the POSIX 
requirement is no longer helpful; on the contrary, it is misleading 
programmers who are trying to write portable software.

If the issue is raised again at the Austin Group level, please let us 
know. I at least would like to put in my two cents.