[PATCH] [BZ #18970]: Reference of pthread_setcancelstate in libc.a

[Roland McGrath]

If this matters it should break some linknamespace test.  If there is no
reference from anything involved in the ISO C implementation, then there is
no actual problem.

For example, wordexp is a POSIX.1 function, so it referring directly to the
public names of other POSIX.1 functions is not a problem.

A change like this is not necessarily worthless even if it is not
necessary.  But you haven't said why you think the change is desireable.
If it makes is easier to understand or maintain the code overall, then that
could be sufficient reason.  But you have to make some sort of positive
case for a change.