Milter-greylist

Here is milter-greylist-4.3.5

http://ftp.espci.fr/pub/milter-greylist/milter-greylist-4.3.5.tgz
MD5 (milter-greylist-4.3.5.tgz) = 66fc347ddda423d0047dc96f6e26db7c

Here are the changes since 4.3.4. I just recovered from a cvs server crash, so
there might be some patches I missed. Please tell me if you sent something
that went unintegrated.

        Update .spec file for libcurl, GeoIP and p0f support (Chris Bennett)
        Really support dkim self and dkim none clauses
        Better configure test to avoid resolver memory leaks (Hajimu UMEMOTO)
        Fix p0f related crashes caused by thread unsafety (Enrico Scholz)
        Fix GeoIP related crashes caused by thread unsafety (Enrico Scholz)
        Start as root and drop priv later on RedHat (Ole Hansen)
        Fix dumpfile argument on command line (Ole Hansen)
        Fix missing reason for whitelisting (Attila Bruncsak)
        Wait 5 seconds before stopping, for database dump (Attila Bruncsak)
        Add mx ACL clause (Rudy Eschauzier)
        Do not use socket mode for the PID file (Kouhei Sutou)
        Fix tarpit deadlock (Kouhei Sutou)

-- 
Emmanuel Dreyfus
http://hcpnet.free.fr/pubz
manu@...

manu-S783fYmB3Ccdnm+yROfE0A@... writes:

>         Fix p0f related crashes caused by thread unsafety (Enrico Scholz)

fwiw; the thread unsafety did not caused crashes but only wrong/missing
results for p0f queries.



Enrico

On 03/13/2010 08:29 AM, manu@... wrote:
>
>
> Here is milter-greylist-4.3.5
>
> http://ftp.espci.fr/pub/milter-greylist/milter-greylist-4.3.5.tgz

...does not compily under Solaris 10 when configured using --enable-mx

gcc -o milter-greylist milter-greylist.o pending.o sync.o dnsrbl.o 
list.o macro.o conf_yacc.o dump_yacc.o conf.o store.o dump.o spf.o acl.o 
urlcheck.o stat.o clock.o geoip.o fd_pool.o prop.o ldapcheck.o 
dkimcheck.o p0f.o spamd.o mx.o -L/usr/local/lib -R/usr/local/lib 
-L/usr/local/BerkeleyDB.4.4/lib -Wl,-R/usr/local/BerkeleyDB.4.4/lib 
-lpthread -lsocket -lresolv -lnsl -lmilter
Undefined                       first referenced
  symbol                             in file
__ns_name_uncompress                mx.o
ld: fatal: Symbol referencing errors. No output written to milter-greylist
collect2: ld returned 1 exit status
gmake: *** [milter-greylist] Error 1

Solaris 10 prototypes ns_name_uncompress() in arpa/nameser.h, but # 
doesn't export it from libresolv.so

# grep ns_name_uncompress nameser.h 
 

#define ns_name_uncompress      __ns_name_uncompress
int             ns_name_uncompress(const uchar_t *, const uchar_t *,


# nm /usr/lib/libresolv.so.2 | grep ns_name_uncompress
[79]    |    133456|      99|FUNC |LOCL |2    |11     |__ns_name_uncompress

Didi

Dietmar Rieder <adrieder@...> wrote:

> Solaris 10 prototypes ns_name_uncompress() in arpa/nameser.h, but # 
> doesn't export it from libresolv.so

Will you contriute a patch fixing this issue?

-- 
Emmanuel Dreyfus
http://hcpnet.free.fr/pubz
manu@...

Hi,

>>>>> On Sat, 13 Mar 2010 08:29:05 +0100
>>>>> manu@... said:

manu> Here is milter-greylist-4.3.5

	snip

manu>         Better configure test to avoid resolver memory leaks (Hajimu UMEMOTO)

Unfortunately, there is same issue in mx.c.  We need to fix it as
well.  Please apply the attached patch.

Sincerely,

Hi,

>>>>> On Sat, 13 Mar 2010 12:27:27 +0100
>>>>> Dietmar Rieder <adrieder@...> said:

adrieder> Solaris 10 prototypes ns_name_uncompress() in arpa/nameser.h, but # 
adrieder> doesn't export it from libresolv.so

Perhaps, we can just replace use of ns_name_uncompress() with
dn_expand(), here.  dn_expand() is a wrapper of ns_name_uncompress(),
and it does bit more than ns_name_uncompress() does.  However, I think
we can ignore the difference in our usage.
Could you try the attached patch?

Sincerely,

Hi,

>>> Mon, 15 Mar 2010 19:03:41 +0900,
>>> Hajimu UMEMOTO <ume@...> said:

>>>>> On Sat, 13 Mar 2010 12:27:27 +0100
>>>>> Dietmar Rieder <adrieder@...> said:

adrieder> Solaris 10 prototypes ns_name_uncompress() in arpa/nameser.h, but # 
adrieder> doesn't export it from libresolv.so

ume> Perhaps, we can just replace use of ns_name_uncompress() with
ume> dn_expand(), here.  dn_expand() is a wrapper of ns_name_uncompress(),
ume> and it does bit more than ns_name_uncompress() does.  However, I think
ume> we can ignore the difference in our usage.
ume> Could you try the attached patch?

Oops, there is one more ns_name_uncompress in the output message.
Please try the attached patch instead.

Sincerely,

Hajimu UMEMOTO <ume@...> wrote:

> Unfortunately, there is same issue in mx.c.  We need to fix it as
> well.  Please apply the attached patch.

Got it.

-- 
Emmanuel Dreyfus
http://hcpnet.free.fr/pubz
manu@...

On 03/15/2010 11:23 AM, Hajimu UMEMOTO wrote:
>
> [Attachment(s) <#TopText> from Hajimu UMEMOTO included below]
>
> Hi,
>
>  >>> Mon, 15 Mar 2010 19:03:41 +0900,
>  >>> Hajimu UMEMOTO <ume@... <mailto:ume%40mahoroba.org>> said:
>
>  >>>>> On Sat, 13 Mar 2010 12:27:27 +0100
>  >>>>> Dietmar Rieder <adrieder@...
> <mailto:adrieder%40sbox.tugraz.at>> said:
>
> adrieder> Solaris 10 prototypes ns_name_uncompress() in arpa/nameser.h,
> but #
> adrieder> doesn't export it from libresolv.so
>
> ume> Perhaps, we can just replace use of ns_name_uncompress() with
> ume> dn_expand(), here. dn_expand() is a wrapper of ns_name_uncompress(),
> ume> and it does bit more than ns_name_uncompress() does. However, I think
> ume> we can ignore the difference in our usage.
> ume> Could you try the attached patch?
>
> Oops, there is one more ns_name_uncompress in the output message.
> Please try the attached patch instead.

Sorry for the the delayed response, I was out of office....
Thank, you so much, it compiles now.

Didi

On 03/13/2010 01:28 PM, manu@... wrote:
>
>
> Dietmar Rieder <adrieder@...
> <mailto:adrieder%40sbox.tugraz.at>> wrote:
>
>  > Solaris 10 prototypes ns_name_uncompress() in arpa/nameser.h, but #
>  > doesn't export it from libresolv.so
>
> Will you contriute a patch fixing this issue?

Sorry for the the delayed response, I was out of office....
Anyway, Hajimu UMEMOTO already contributed a patch for this, so thanks 
to him I don't need to digg deeper into this :-)

Didi

On Tue, Mar 16, 2010 at 11:55:58AM +0100, Dietmar Rieder wrote:
> Sorry for the the delayed response, I was out of office....
> Anyway, Hajimu UMEMOTO already contributed a patch for this, so thanks 
> to him I don't need to digg deeper into this :-)

I understood he was looking for soemone to test the patch...

-- 
Emmanuel Dreyfus
manu@...

On 03/16/2010 12:43 PM, Emmanuel Dreyfus wrote:
>
>
> On Tue, Mar 16, 2010 at 11:55:58AM +0100, Dietmar Rieder wrote:
>  > Sorry for the the delayed response, I was out of office....
>  > Anyway, Hajimu UMEMOTO already contributed a patch for this, so thanks
>  > to him I don't need to digg deeper into this :-)
>
> I understood he was looking for soemone to test the patch...

Well so far it applies and compiles.
Is there any documentation on the "mx" feature? Howto use it?
Sorry but it seems that I'm blind...

Didi

Hi,

> Here is milter-greylist-4.3.5
> 
> http://ftp.espci.fr/pub/milter-greylist/milter-greylist-4.3.5.tgz
> MD5 (milter-greylist-4.3.5.tgz) = 66fc347ddda423d0047dc96f6e26db7c
> 
> Here are the changes since 4.3.4. I just recovered from a cvs server 
> crash, so there might be some patches I missed. Please tell me if 
> you sent something that went unintegrated.
> 
>         Update .spec file for libcurl, GeoIP and p0f support (Chris 
> Bennett)        

A few queries here:

1. How does one get changes applied to the spec?

For dkim support, I added:
%define p0f 1
%{?build_p0f:%define p0f 1}

%define libdkim 1
%{?build_libdkim:%define libdkim 1}

to the top and then:

and then:

%if %{libGeoIP}
        --with-libGeoIP \
%endif
%if %{libdkim}
        --with-libdkim \
%endif

so it will compile with dkim support.

2. For this spec, I also had to change:

##%define user root
%define user smmsp
%{?build_user:%define user %{build_user}}

so it could run as "smmsp". I didn't have to do this with the milter-greylist
4.3.4 spec file but I do have to with this one?

My guess is it's the added "-u user" paramater on the start up (maybe the init
script has changed?). My milter-greylist 4.3.4 servers show:

# ps aux|grep greylist
smmsp    10807  0.0  0.3 186592 19536 ?        Ssl  Mar08   2:43
/usr/bin/milter-greylist -P /var/milter-greylist/milter-greylist.pid -p
/var/milter-greylist/milter-greylist.sock

while the newer milter-greylist 4.3.5 shows:

# ps aux|grep greylist
smmsp    29898  0.0  0.4 157412 10296 ?        Ssl  01:47   0:00
/usr/bin/milter-greylist -u smmsp -P /var/milter-greylist/milter-greylist.pid
-p /var/milter-greylist/milter-greylist.sock

3. the milter-greylist.sock file is "root:root" for 4.3.5, while it's
smmsp:smmsp for 4.3.4 as this shows, for 4.3.4:

# ll /usr/bin/milter-greylist /var/milter-greylist/milter-greylist.pid
/var/milter-greylist/milter-greylist.sock
-rwxr-xr-x 1 root  root  283252 Feb 28 21:24 /usr/bin/milter-greylist
-rw-r--r-- 1 smmsp smmsp      6 Mar  8 23:05
/var/milter-greylist/milter-greylist.pid
srwxr-xr-x 1 smmsp smmsp      0 Mar  8 23:05
/var/milter-greylist/milter-greylist.sock

while 4.3.5:

# ll /usr/bin/milter-greylist /var/milter-greylist/milter-greylist.pid
/var/milter-greylist/milter-greylist.sock
-rwxr-xr-x 1 root  root  281248 Mar 17 01:45 /usr/bin/milter-greylist
-rw-r--r-- 1 smmsp smmsp      6 Mar 17 01:47
/var/milter-greylist/milter-greylist.pid
srwxr-xr-x 1 root  root       0 Mar 17 01:47
/var/milter-greylist/milter-greylist.sock

I've only compiled the 4.3.5 spec on one mail server and am running mgl 4.3.5
only on that server. It still seems to be running fine but I'm concerned that
the milter-greylist.sock file is owned by root:root.

Is there something wrong with the init script to make it do this?

4. Addition of smmsp:smmsp chown to /var/milter-greylist

I had to add:

%post
/sbin/chkconfig --add milter-greylist
/bin/chown smmsp:smmsp /var/milter-greylist

to the spec file so the directory is owned by smmsp:smmsp otherwise
milter-greylist won't run. I know I can add a --user paramater? to the
rpmbuild line but it the spec really needs to be corrected for this to work
with just a normal:

rpmbuild -bb milter-greylist.spec

command.

Thanks.

Michael.

"Michael Mansour" <mic-Su61qOI1guu6c6uEtOJ/EA@...> writes:

> %define p0f 1
> %{?build_p0f:%define p0f 1}
> ...
> %if %{libGeoIP}
>         --with-libGeoIP \

It would be better to use %bcond stuff (see /usr/lib/rpm/macros); e.g.

%bcond_with    p0f
%bcond_without geoip
...
%configure \
           --with%{!?with_p0f:out}-p0f \
           --with%{!?with_geoip:out}-geoip \


> %post
> ...
> /bin/chown smmsp:smmsp /var/milter-greylist

setting

%files
%attr(0711,smmsp,smmsp) /var/milter-greylist

would be a better alternative (although I am not sure whether it is
really needed)