[00:12:29] --- Simon Wilkinson has left
[00:12:29] --- Simon Wilkinson has become available
[01:57:10] --- Simon Wilkinson has left
[05:36:25] --- meffie has become available
[06:07:29] --- ktdreyer has left: Replaced by new connection
[06:07:29] --- ktdreyer has become available
[06:19:22] --- mvita has become available
[07:43:55] --- jaltman/FrogsLeap has left: Disconnected
[07:52:24] --- jaltman/FrogsLeap has become available
[07:56:03] --- deason has become available
[08:03:42] <kaduk@jabber.openafs.org/barnowl> Hmm, my clang warns in auth/ktc.c:149 (et seq), basically we have
static struct {
	/* four fields */
} local_tokens[4]  = {{0},{0},{0},{0}};

But without the initializer it would still get set to all zeros.  Is
there any reason to keep the explicit initializer?
[08:20:18] <deason> doesn't seem like it, but what's the warning?
[08:29:59] * ballbery wonders how old that code is; go back far enough and C compilers were not guaranteed to zero static variables... (admittedly most Unix C compilers have had BSS pretty much forever. then again there is cleverness that takes advantage of uninitialized going into BSS and initialized going into DATA, not that I know whether any code here attempts such cleverness)
[08:35:10] <ballbery> (well, *was* such cleverness. not exactly encouraged these days and probably not supported by modern compilers)
[08:44:24] <kaduk@jabber.openafs.org/barnowl> ../../../openafs/src/auth/ktc.c:149:2: warning: missing field 'server'
      initializer [-Wmissing-field-initializers]
0}, {
 ^
[08:46:58] <jaltman/FrogsLeap> the warning is because the initializer is wrong.  It is only initializing one field for each of the four array elements.  It needs to initialize the entire structure if it is going to initialize.
[08:49:10] <kaduk@jabber.openafs.org/barnowl> Seems we are in support of a patch being submitted.
[10:16:37] <kaduk@jabber.openafs.org/barnowl> (I mostly only notice those ktc.c warnings because it is right after
the rxgk_* files in my build log so I see it while paging down.
Of course, there's a -Wsign-compare warning from ktc.c as well, but
maybe I should just turn that warning off, since we're in a pretty bad
state.
[12:24:53] <ballbery> fwiw I just confirmed the OS X 10.9 build error reported in https://lists.openafs.org/pipermail/openafs-devel/2013-December/019717.html
[12:25:58] <kaduk@jabber.openafs.org/barnowl> FWIW, I build master on 10.9.1 just fine (but without -Werror).
[12:26:57] <ballbery> right, this is 1.6.6pre2 and I figured that killing the -Werror was the way to go. I recall seeing some discussion fo this somewhere but can't find it right now
[12:27:33] <ballbery> oh, it was in here, I hadn't scrolled back far enough
[13:19:46] <kaduk@jabber.openafs.org/barnowl> Since I'm on an OS X 10.9.1 box, I might as well try and go fix the
build with the buildbot's flags.
Except, I am seeing errors at
/usr/include/secure/_stdio.h:56:9: note: previous definition is here
#define snprintf(str, len, ...) \
        ^
In file included from
/Users/kaduk/openafs/openafs/src/external/heimdal/roken/ecalloc.c:39:
/Users/kaduk/openafs/build/include/roken.h:470:9: error: 'vsnprintf'
macro
      redefined [-Werror]
not at what Andrew posted.
[13:26:49] <kaduk@jabber.openafs.org/barnowl> Oh, but I was forcing CC=clang, that might do it.
[13:27:01] <kaduk@jabber.openafs.org/barnowl> ... or not.
[13:32:07] <kaduk@jabber.openafs.org/barnowl> I'll just grab the configure invocation wholesale, that will save me
headaches.
[13:38:14] <kaduk@jabber.openafs.org/barnowl> #define RFC3961_NO_ENUMS
in krb5_locl.h just gives:
../../../../openafs/src/crypto/rfc3961/rfc3961.h:53:3: error: unknown
type name
      'CKSUMTYPE'
  CKSUMTYPE cksumtype;

for the curious
[13:42:10] --- Simon Wilkinson has become available
[14:50:13] --- mvita has left
[14:52:56] --- meffie has left
[16:30:18] --- deason has left
[20:27:59] <kaduk@jabber.openafs.org/barnowl> Hmm, sbrk() is deprecated in Mavericks, so all uses are errors to the
buildbot configuration.
Unfortunately, getrusage(2) is a bit more verbose.
[20:28:22] <kaduk@jabber.openafs.org/barnowl> Time for a helper routine...