readlink() from type `int' to type `size_t'. This isn't an ABI change, since
the calling convention of our only LP64 platform (the Alpha) already promotes
this argument to a `long'.
This may not be the final action on this matter; readlink() still returns
an `int', which may change in a future revision of the standard.
check _STANDALONE as well as _KERNEL. _KERNEL is incorrect for use when
building boot blocks, and it looks like the rest of the code is already
using _STANDALONE for this purpose.
- %C and %y can be used in conjunction properly, though %y can have its
old behaviour if used without %C
- %I formats hous 0..11 with 0 representing "12"; this is correct both
with and without am/pm (%p rule fixed to compensate)
- %j fixed for range [0..365] in struct tm
- %m fixed for range [0..11] in struct tm
- %S fixed for range [0..61] in struct tm
required removing fcloses and/or moving them, nulling the pointers
like they should have been, and adding a terminating null byte and break
in one case (perhaps we never ran into those problems before).
cgetclose() already does a fclose and null of the pointer, so don't
duplicate it when we don't need to.
* $PWD was used if it appeared ``correct''. this `enhancment' was
not documented in getcwd(3), and caused problems for pwd(1) (which
uses getcwd(3)), and in turn <bsd.obj.mk> (which uses pwd) when $PWD
is set to the logical path instead of the real path (e.g, symlink
/usr/src -> /z/NetBSD-current/src).
[$PWD is set to the logical path by tcsh(1) with set symlinks=expand]
* as getcwd() without the $PWD code is just a wrapper to getcwd_physical(),
rename the latter to the latter. hack realpath() to use the getcwd().
yet another case of `it was a good idea, but...'
