Push -Wno-array-bounds down to the cases that depend on it.
Selectively disable warnings for 3rd party software or non-trivial
issues to be reviewed later to get clang -Werror to build most of the
tree.
FORTIFY_SOURCE feature of libssp, thus checking the size of arguments to
various string and memory copy and set functions (as well as a few system
calls and other miscellany) where known at function entry. RedHat has
evidently built all "core system packages" with this option for some time.
This option should be used at the top of Makefiles (or Makefile.inc where
this is used for subdirectories) but after any setting of LIB.
This is only useful for userland code, and cannot be used in libc or in
any code which includes the libc internals, because it overrides certain
libc functions with macros. Some effort has been made to make USE_FORT=yes
work correctly for a full-system build by having the bsd.sys.mk logic
disable the feature where it should not be used (libc, libssp iteself,
the kernel) but no attempt has been made to build the entire system with
USE_FORT and doing so will doubtless expose numerous bugs and misfeatures.
Adjust the system build so that all programs and libraries that are setuid,
directly handle network data (including serial comm data), perform
authentication, or appear likely to have (or have a history of having)
data-driven bugs (e.g. file(1)) are built with USE_FORT=yes by default,
with the exception of libc, which cannot use USE_FORT and thus uses
only USE_SSP by default. Tested on i386 with no ill results; USE_FORT=no
per-directory or in a system build will disable if desired.
* RFC 3542 isn't binary compatible with RFC 2292.
* RFC 2292 support is on by default but can be disabled.
* update ping6, telnet and traceroute6 to the new API.
From the KAME project (www.kame.net).
Reviewed by core.
with MKPIC=no, possibly because the target does not support shared
libraries, these include libraries required to resolve all symbols
which end up referenced from PAM-using applications. The libraries
presently required are -lcrypt, -lrpcsvc and -lutil.
Add use of these variables which are currently set up to use PAM,
so that they compile when MKPIC=no.
Also, in the telnetd case, reorder the order of the libraries, so
that libtelnet.a comes before -ltermcap and -lutil, again to fix
link error when MKPIC=no.
Discussed with thorpej and christos.
This matches the binary interface from before if was de-__Peed (or at least
that of K&R C).
In any case having arguments (and return values) of char and short isn't
especially sensible.
returning a failure code and processing that all the way up the stack.
- deal with a dead peer similarly and do it without setjmp/longjmp() (at
christos' request).
These fix bin/20304 and my own observations when the remote abruptly closes
the connection.
and without Kerberos 4 & 5 (MKKERBEROS=no). Previously checkflist
complained of missing files.
* move kerberos- and kerberos 4-only files into new flists,
distrib/sets/lists/*/krb.*
* make the flist generators grok MKKERBEROS{,4} variables
* fix Makefiles which treat MKKERBEROS=no as MKKERBEROS5=no.
9 out of 10 experts agree that it is ludicrous to build w/
KERBEROS4 and w/o KERBEROS5.
* fix header files, also, which treat MKKERBEROS=no as MKKERBEROS5=no.
* omit some Kerberos-only subdirectories from the build as
MKKERBEROS{,4} indicate
(I acknowledge the sentiment that flists are the wrong way to go,
and that the makefiles should produce the metalog directly. That
sounds to me like the right way to go, but I am not prepared to do
revamp all the makefiles. While my approach is expedient, it fits
painlessly within the current build architecture until we are
delivered from flist purgatory, and it does not postpone our
delivery. Fair enough?)
wiz