bug#27736: OSX 10.6.8: Building from master branch fails.

[Charles A. Roelli]

I have no idea why the code goes into an infinite loop, but it seems
somebody else has already had the same problem (on FreeBSD, the source
of Apple's libc):

https://rt.cpan.org/Public/Bug/Display.html?id=40764

They have a freeze on the same value of "time" that we have:

(gdb) frame 4
#4  0x00000001002b9d8b in emacs_localtime_rz (tz=0x102102a60, 
t=0x7fff5fbe8b80, tm=0x7fff5fbe8b48) at editfns.c:103
103      tm = localtime_rz (tz, t, tm);
(gdb) ptype t
type = long *
(gdb) p *t
$2 = -67768038400720896

Somebody in the linked thread also had the same problem on macOS. And
apparently NetBSD had some of the same issues (search for "infinite
loop in localtime" on the linked page):

http://cvsweb.netbsd.org/bsdweb.cgi/src/lib/libc/time/localtime.c?only_with_tag=MAIN

But I'm not sure those fixes would have anything to do with macOS.


On 20/07/2017 00:48, Alan Third wrote:
> On Wed, Jul 19, 2017 at 09:58:03PM +0200, Charles A. Roelli wrote:
> > Err, oops: I guess there will be no patch for this in Lisp, since
> > decode-time is written in C.
> >
> >
> > So the choice is either to make a change in lib/time_rz.c or src/editfns.c.
> Do you know what the actual problem is? If it’s going into an infinite
> loop, why is that happening? That might give us a better idea where
> the fix should go.