Mailing List Archive: exim: users

Exim 4.80 RC1 uploaded

Index | Next | Previous | View Threaded

pdp at exim

May 17, 2012, 8:55 PM

Post #1 of 9 (319 views)
Permalink

Exim 4.80 RC1 uploaded

I have uploaded Exim 4.80 RC1 to:
ftp://ftp.exim.org/pub/exim/exim4/test/

This release contains a number of backwards-incompatible changes, for
both OpenSSL and GnuTLS, in the name of security (about the only reason
we normally accept for being backwards incompatible). Please read over
README.UPDATING carefully! We have jumped from 4.77 to 4.80 for this
reason.

This is the first release of Exim to support OpenSSL 1.0.1+. This
release of Exim abandons maintaining hard-coded lists of ciphers for
GnuTLS in favour of honouring GnuTLS library policy, so MD5-based
certificates will no longer work and are not supported. The GnuTLS
support has been re-written and there is the possibility of bugs.

When building, please do not just recycle your previous Local/Makefile;
there are a number of new possibilities in this release which may make
your life easier, as support for using pkg-config to query CFLAGS/LIBS
for various pieces of software has been added.

Please join me in welcoming Jeremy Harris and Todd Lyons to the Exim
Maintainers team.

The ChangeLog/NewStuff/README.UPDATING can be reviewed at:

http://git.exim.org/exim.git/blob/exim-4_80_RC1:/doc/doc-txt/ChangeLog
http://git.exim.org/exim.git/blob/exim-4_80_RC1:/doc/doc-txt/NewStuff
http://git.exim.org/exim.git/blob/exim-4_80_RC1:/src/README.UPDATING

The files are signed with the PGP key 0x3903637F, which has a uid
"Phil Pennock <pdp [at] exim>". Please use your own discretion in
assessing what trust paths you might have to this uid.

Checksums below. Detached PGP signatures in .asc files are available
alongside the tarballs.

Please report issues in reply to this email, on exim-users.

Thank you for your testing and feedback,
-Phil Pennock, pp The Exim Maintainers.

SHA256(exim-4.80_RC1.tar.bz2)= c2b67b6baea743fc73dcf9d2e298d1eef9a79f2172958fba8fd9b3025c328bba
SHA256(exim-4.80_RC1.tar.gz)= c5a4d1402cf7eff4b4b9ced3fcee8d195f0ff1f335d13c68d58206f4f72215b9
SHA256(exim-html-4.80_RC1.tar.bz2)= fb1189bd1a053e06390d6cee3ad5bd6b508a806c0d143c4a9fc70822605d63ec
SHA256(exim-html-4.80_RC1.tar.gz)= 45fe3b2f3d6f431b010b9c665cd5bdfdd8388f08819d814ecf70b179eccccdc2
SHA256(exim-pdf-4.80_RC1.tar.bz2)= 343599159461b47156d613041cb62afb549f6286f320f9cf62eb856dd95811bc
SHA256(exim-pdf-4.80_RC1.tar.gz)= 1eb6afdfa5fd35d4130e6d305baa285d03007e456e4e4e5baf884c7608ddc287
SHA256(exim-postscript-4.80_RC1.tar.bz2)= 7867c6aba62fbeae46342bc40f384f5abaee8fb2553ceb0a7c25a1ff68b7226b
SHA256(exim-postscript-4.80_RC1.tar.gz)= c7a6a92e4f74e9afad7f702ed703d6b19a285e814d77090a2d02d21fb2b0ca4c

SHA1(exim-4.80_RC1.tar.bz2)= eb92b4b220ec5c0161d6c90503f3d0b0aea9f1f3
SHA1(exim-4.80_RC1.tar.gz)= 53f9add7baf2bf5361695cf2ff6f75e2a24f61c9
SHA1(exim-html-4.80_RC1.tar.bz2)= 0bb5a391fb22d606edf8f16263bdc45426c8ab89
SHA1(exim-html-4.80_RC1.tar.gz)= a0baf96dd903dd4e1a1c86da465b70e633009991
SHA1(exim-pdf-4.80_RC1.tar.bz2)= 961151a4fefc84b6e76ee1fceac54a0c321fe2c4
SHA1(exim-pdf-4.80_RC1.tar.gz)= 858711456ce822d9cc4464b01f419bcbe8daa2e6
SHA1(exim-postscript-4.80_RC1.tar.bz2)= 924f9cece857057e75a23e1e8d556f25ab99a353
SHA1(exim-postscript-4.80_RC1.tar.gz)= 7283d5c762264e00e6102265a8f1eaf65048c0b6

michael at moria

May 19, 2012, 3:20 PM

Post #2 of 9 (305 views)
Permalink

Re: [exim-dev] Exim 4.80 RC1 uploaded [In reply to]

Hello,

I tried RC1 on HP-UX 11.00 with gcc and apart from a few minor issues,
things look fine.

I suggest not to use those functions:

dcc.c: In function 'dcc_process':
dcc.c:129: warning: incompatible implicit declaration of built-in function 'bzero'
dcc.c:196: warning: incompatible implicit declaration of built-in function 'bcopy'

HP-UX 11 needs these additions to OS/os.h-HP-UX:

----------
typedef struct __res_state *res_state;

#define LLONG_MIN LONG_LONG_MIN
#define LLONG_MAX LONG_LONG_MAX

#define strtoll(a,b,c) strtoimax(a,b,c)
----------

The first is needed due to the HP-UX resolver library. I don't know
when res_state was defined, otherwise I would have used __RES to check
if it needs to be defined. 19931104 does not have it yet.

Michael

--
## List details at https://lists.exim.org/mailman/listinfo/exim-users
## Exim details at http://www.exim.org/
## Please use the Wiki with this list - http://wiki.exim.org/

pdp at exim

May 19, 2012, 4:16 PM

Post #3 of 9 (301 views)
Permalink

Re: [exim-dev] Exim 4.80 RC1 uploaded [In reply to]

On 2012-05-20 at 00:20 +0200, Michael Haardt wrote:
> I tried RC1 on HP-UX 11.00 with gcc and apart from a few minor issues,
> things look fine.

Cool, thanks. :)

> I suggest not to use those functions:
>
> dcc.c: In function 'dcc_process':
> dcc.c:129: warning: incompatible implicit declaration of built-in function 'bzero'
> dcc.c:196: warning: incompatible implicit declaration of built-in function 'bcopy'

Those are the EXPERIMENTAL_DCC stuff, which is about to get a little
work from Wolfgang after the release, so I think it best to leave it.

> HP-UX 11 needs these additions to OS/os.h-HP-UX:
>
> ----------
> typedef struct __res_state *res_state;
>
> #define LLONG_MIN LONG_LONG_MIN
> #define LLONG_MAX LONG_LONG_MAX
>
> #define strtoll(a,b,c) strtoimax(a,b,c)
> ----------
>
> The first is needed due to the HP-UX resolver library. I don't know
> when res_state was defined, otherwise I would have used __RES to check
> if it needs to be defined. 19931104 does not have it yet.

*sigh* Thanks, I was worried this assumption would bite me; I'm
surprised directly messing with _res has proved so portable. I think I
saw some slightly struct names and want to avoid the main code depending
upon __doubled_underbar names, so I think that fixing this in the os.h
as you did is the correct path to take.

Only NetBSD has moved away from _res ... mind, their approach is a good
one and arguably correct.

Thanks for the fixes for HP-UX, pushed to master.
-Phil

--
## List details at https://lists.exim.org/mailman/listinfo/exim-users
## Exim details at http://www.exim.org/
## Please use the Wiki with this list - http://wiki.exim.org/

tlyons at ivenue

May 20, 2012, 7:52 AM

Post #4 of 9 (307 views)
Permalink

Re: [exim-dev] Exim 4.80 RC1 uploaded [In reply to]

On Sat, May 19, 2012 at 4:16 PM, Phil Pennock <pdp [at] exim> wrote:
> > HP-UX 11 needs these additions to OS/os.h-HP-UX:
> > ----------
> > typedef struct __res_state *res_state;
> >
> > #define LLONG_MIN LONG_LONG_MIN
> > #define LLONG_MAX LONG_LONG_MAX
> >
> > #define strtoll(a,b,c) strtoimax(a,b,c)
> > ----------
> > The first is needed due to the HP-UX resolver library. �I don't know
> > when res_state was defined, otherwise I would have used __RES to check
> > if it needs to be defined. �19931104 does not have it yet.
>
> *sigh* �Thanks, I was worried this assumption would bite me; I'm
> surprised directly messing with _res has proved so portable. �I think I
> saw some slightly struct names and want to avoid the main code depending
> upon __doubled_underbar names, so I think that fixing this in the os.h
> as you did is the correct path to take.

On CentOS 5.x, I had to add -std=gnu99 to my CFLAGS for the
LLONG_MIN/LLONG_MAX errors to go away. Per googling, if I had not
been specifying CFLAGS in my spec file, I probably wouldn't have
experienced those errors because apparently on CentOS, override CFLAGS
causes it to lose some default set args.

...Todd
--
Always code as if the guy who ends up maintaining your code will be a
violent psychopath who knows where you live. -- Martin Golding

--
## List details at https://lists.exim.org/mailman/listinfo/exim-users
## Exim details at http://www.exim.org/
## Please use the Wiki with this list - http://wiki.exim.org/

michael at moria

May 20, 2012, 10:38 AM

Post #5 of 9 (302 views)
Permalink

Re: [exim-dev] Exim 4.80 RC1 uploaded [In reply to]

> > The first is needed due to the HP-UX resolver library. I don't know
> > when res_state was defined, otherwise I would have used __RES to check
> > if it needs to be defined. 19931104 does not have it yet.
>
> *sigh* Thanks, I was worried this assumption would bite me; I'm
> surprised directly messing with _res has proved so portable. I think I
> saw some slightly struct names and want to avoid the main code depending
> upon __doubled_underbar names, so I think that fixing this in the os.h
> as you did is the correct path to take.

__RES was made specifically to let code work with different resolver
versions. I expect the same issue to come up with other old systems
and checking __RES would allow building without trouble on those old
systems where a current resolver is installed.

It should be ok for the release, though, and admins that run a current
resolver on HP-UX probably know what to edit. :)

Michael

--
## List details at https://lists.exim.org/mailman/listinfo/exim-users
## Exim details at http://www.exim.org/
## Please use the Wiki with this list - http://wiki.exim.org/

D.H.Davis at bath

May 21, 2012, 3:31 AM

Post #6 of 9 (301 views)
Permalink

Re: [exim-dev] Exim 4.80 RC1 uploaded [In reply to]

On Sun, 20 May 2012, Todd Lyons wrote:

> From: Todd Lyons <tlyons [at] ivenue>
> To: Michael Haardt <michael [at] moria>, exim-users [at] exim,
> Wolfgang Breyha <wbreyha [at] gmx>
> Date: Sun, 20 May 2012 15:52:28
> Subject: Re: [exim] [exim-dev] Exim 4.80 RC1 uploaded
> X-Spam-Score: 0.0 (/)
>
> On Sat, May 19, 2012 at 4:16 PM, Phil Pennock <pdp [at] exim> wrote:
> > > HP-UX 11 needs these additions to OS/os.h-HP-UX:
> > > ----------
> > > typedef struct __res_state *res_state;

...

OpenBSD needs the above change applied to OS/os.h-OpenBSD:

diff -u os.h-OpenBSD.orig os.h-OpenBSD
--- os.h-OpenBSD.orig Mon May 21 05:32:11 2012
+++ os.h-OpenBSD Mon May 21 09:55:27 2012
@@ -11,4 +11,6 @@
#define os_strsignal strsignal
#define OS_STRSIGNAL

+typedef struct __res_state *res_state;
+

Now have exim-4.80_RC4 running on OpenBSD5.1. Nothing special, it's
just relaying mail to a smarthost.
--
Dennis Davis, BUCS, University of Bath, Bath, BA2 7AY, UK
D.H.Davis [at] bath Phone: +44 1225 386101

--
## List details at https://lists.exim.org/mailman/listinfo/exim-users
## Exim details at http://www.exim.org/
## Please use the Wiki with this list - http://wiki.exim.org/

exim-users at spodhuis

May 21, 2012, 3:51 AM

Post #7 of 9 (298 views)
Permalink

Re: [exim-dev] Exim 4.80 RC1 uploaded [In reply to]

On 2012-05-21 at 11:31 +0100, Dennis Davis wrote:
> > > > typedef struct __res_state *res_state;

> OpenBSD needs the above change applied to OS/os.h-OpenBSD:

Done, thanks.

Any idea if they're moving towards the NetBSD resolver changes? It's
the sort of safety change I'd expect OpenBSD to make but I don't see the
changes in <http://www.openbsd.org/cgi-bin/man.cgi?query=resolver>.

-Phil

--
## List details at https://lists.exim.org/mailman/listinfo/exim-users
## Exim details at http://www.exim.org/
## Please use the Wiki with this list - http://wiki.exim.org/

D.H.Davis at bath

May 21, 2012, 6:58 AM

Post #8 of 9 (294 views)
Permalink

Re: [exim-dev] Exim 4.80 RC1 uploaded [In reply to]

On Mon, 21 May 2012, Phil Pennock wrote:

> From: Phil Pennock <exim-users [at] spodhuis>
> To: Dennis Davis <D.H.Davis [at] bath>
> Cc: exim-users [at] exim
> Date: Mon, 21 May 2012 11:51:24
> Subject: Re: [exim] [exim-dev] Exim 4.80 RC1 uploaded
> Reply-To: exim-users [at] exim
>
> On 2012-05-21 at 11:31 +0100, Dennis Davis wrote:
> > > > > typedef struct __res_state *res_state;
>
> > OpenBSD needs the above change applied to OS/os.h-OpenBSD:
>
> Done, thanks.

Forgot to mention that this builds fine on Solaris5.10, i386
platform. I've got RC4 built against openssl-1.0.1c and in
production use on a server. This handles little email, so it's
hardly going to be a stress test.

> Any idea if they're moving towards the NetBSD resolver changes?

Not sure what these changes are.

> It's the sort of safety change I'd expect
> OpenBSD to make but I don't see the changes in
> <http://www.openbsd.org/cgi-bin/man.cgi?query=resolver>.

They seem to be moving towards ditching BIND and replacing it with
unbound:

http://www.unbound.net/

and ldns:

http://www.nlnetlabs.nl/projects/ldns/

Looks to be in the post-OpenBSD5.1 source tree but not yet fully
integrated. See the discussion stating at:

http://marc.info/?l=openbsd-misc&m=133745965814866&w=2
--
Dennis Davis, BUCS, University of Bath, Bath, BA2 7AY, UK
D.H.Davis [at] bath Phone: +44 1225 386101

--
## List details at https://lists.exim.org/mailman/listinfo/exim-users
## Exim details at http://www.exim.org/
## Please use the Wiki with this list - http://wiki.exim.org/

exim-users at spodhuis

May 21, 2012, 7:18 PM

Post #9 of 9 (294 views)
Permalink

Re: [exim-dev] Exim 4.80 RC1 uploaded [In reply to]

On 2012-05-21 at 14:58 +0100, Dennis Davis wrote:
> On Mon, 21 May 2012, Phil Pennock wrote:
> > Any idea if they're moving towards the NetBSD resolver changes?
>
> Not sure what these changes are.

Thread-safety with a different set of resolver routines avoiding _res.

http://netbsd.gw.com/cgi-bin/man-cgi?resolver++NetBSD-6.0

> They seem to be moving towards ditching BIND and replacing it with
> unbound:

I have unbound as the system resolver on my dev box, for validating
DNSSEC; it doesn't change the libc interfaces. It's just something
else to listen on ports 53.

It includes a libunbound, which can be used *instead* of the libc
interfaces. It doesn't include variants for getaddrinfo() et al.

-Phil

--
## List details at https://lists.exim.org/mailman/listinfo/exim-users
## Exim details at http://www.exim.org/
## Please use the Wiki with this list - http://wiki.exim.org/

Index | Next | Previous | View Threaded

Interested in having your list archived? Contact Gossamer Threads