Re: Debian package of SA-Exim 2.2

Nicolas Marchildon <[email protected]>
Newsgroups gmane.mail.exim.spamassassin
Message-ID <[email protected]>
On Mon, 2003-02-10 at 01:09, Marc MERLIN wrote:
> I am loathe to release sa-exim 2.3 without actually using it on my system,
> and I haven't had the time to work on exim 4.12 and sa-exim 2.3 lately.
> sa-exim 2.3 also has a minor bug with X-SA-Exim-Rcpt-To: if you have
> hundreds of receipients in a mail.
> sa-exim 2.2cvs fixes that, and I think you have that.

Well, I guess I can live with that :)

> The 4.10 package should be fine, and you can compile sa-exim yourself
> and drop it in /usr/lib/exim4/
> 
> > For example, with exim 4.10, after having fixed debian/EDITME.diff, it
> > started building exim, but failed here:
> > 
> > /bin/rm -f auths.a
> > ar cq auths.a b64encode.o b64decode.o call_pam.o call_pwcheck.o
> > call_radius.o xtextencode.o xtextdecode.o get_data.o md5.o cram_md5.o
> > plaintext.o pwcheck.o auth-spa.o spa.o
> > ranlib auths.a
> > /bin/rm -rf ../drtables.o
> > make[3]: Leaving directory
> > `/home/nicolas/download/exim-4.10/build-Linux-i386/auths'
> >  
> > make[2]: *** No rule to make target `lookups/lookups.a', needed by
> 
> Ah.
> That's a problem with the exim makefile
> It actually fails earlier, probably with postgres because debian
> changed/renamed the postgres packages recently.
> If you look up the build log, you'll see the real error.
> 
> If you edit EDITME.diff to remove postgres support, I'm fairly sure the
> build will work.

Ok, I removed LOOKUP_PGSQL and it then failed in the same way with
mysql, so I removed LOOKUP_MYSQL too. Then it failed with ldap, so I
removed LOOKUP_LDAP. Then ld complained about missing -lldap, so I
apt-get installed libldap2-dev (I already had libldap2). Then ld
complained about -lmysqlclient, so I installed libmysqlclient10-dev.
Then -lpq, so I installed postgresql-dev.

Then, ha! something interesting ;)

cd debian/sa-exim; /usr/bin/make
make[1]: Entering directory `/home/nicolas/download/sa-exim'
Building sa-exim-2.2.x-cvs.so
gcc -I../../src -DDLOPEN_LOCAL_SCAN
-DSPAMASSASSIN_CONF=\"/etc/exim/spamassassin.conf\"
-DSPAMC_LOCATION=\"/usr/bin/spamc\" -O2 -Wall -shared -o
sa-exim-2.2.x-cvs.so sa-exim.c
sa-exim.c:31: local_scan.h: No such file or directory
sa-exim.c:361: warning: #warning you should not worry about the "might
be clobbered by longjmp", see source
make[1]: *** [sa-exim-2.2.x-cvs.so] Error 1
make[1]: Leaving directory `/home/nicolas/download/sa-exim'
make: *** [build] Error 2

However:

cortex:/home/nicolas/download/exim-4.10# find -name local_scan.h
./src/local_scan.h
./build-Linux-i386/local_scan.h

I had symlinked debian inside exim-4.10 to ../debian, so instead, I
copied debian inside the exim dir.

And it worked :)

I just had to install spamassassin on my test box, enable it in
/etc/default/spamassassin, removed the -c option (dunno why, but it does
not work otherwise), and successfully injected a big message into my
system:

2003-02-11 00:41:59 18iTB9-0004VV-00 SA: SATruncBodyCond expand
returned: '0'
2003-02-11 00:41:59 18iTB9-0004VV-00 SA: check skipped due to message
size (374356 bytes) and SATruncBodyCond expanded to false

What's planned about that size limitation? Will spamc ever raise its
limit? Will they make the ending-mime-boundary check optional so I can
safely enable SATruncBodyCond?

> I'll try to get my exim4 4.12 package out soon, and I'll probably make the
> sa-exim package separate for the benefit of the exim4debian folks.

That would be great. I'm planning to spread the word about teergrubbing
and general, and sa-exim in particular. I really like the idea of
"playing" with spammers almost directly :)

> There was some discussion on the exim4debian list (people putting out a real
> exim4 package), and sa-exim will be packaged separately, which will
> alleviate some of these problems.

Glad to hear it!

> > I also noticed that the debian tree was not stored in CVS, which was
> > probably just from lack of time. It would be great if that was in CVS if
> > we are going to modify it (update EDITME.diff). It could simply be a
> > separate module in the current sa-exim repository.
> 
> Well, my exim4 debian package works, but it's not a work of art, and story
> its debian tree in the sa-exim CVS tree isn't really appropriate, they
> are technically separate.
> I'm sure than 75% of the sa-exim users don't use debian, a few don't even
> use linux, so an exim4 debian build tree wouldn't have it's place there :)

I was not talking about putting it in the exact same tree, but rather in
a module, so that the regular "cvs checkout sa-exim" would not be
affected, and debian people would have to run something like "cvs
checkout sa-exim-deb".

Anyway, now we have two more up-to-date debian packages: the one
Lawrence made earlier, and mine. Because I removed a few options from
EDITME, I'm not sure mine is the best one for others to play with. I
guess I could put the options back and try to build again, as I probably
installed all the required libraries anyway in the end.

Hmm, I just realized Lawrence did *not* post his link to the
mailing-list for his debian package of exim 4.10 and sa-exim cvs.

So, here are mines, PGP signed (only useful if you trust my key in the
first place ;)

http://marchildon.net/sa-exim/

Nicolas


-- 
OpenPGP public key:         http://nicolas.marchildon.net/pubkey.txt
Key fingerprint:  5E84 1089 0036 BB63 6997  232C 8FFB 777D 39D4 B2D4
Jabber ID: [email protected]              http://www.jabber.org
What have you done for freedom today?             http://www.gnu.org

_______________________________________________
SA-Exim mailing list
[email protected]
http://lists.merlins.org/lists/listinfo/sa-exim
signature.asc (application/pgp-signature, 189 B)
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.0 (GNU/Linux)

iD8DBQA+SJMuj/t3fTnUstQRAuN1AJ99OfXNRwSXen08arYW3FdNCtbEfwCfbycV
aajW6UW+FAUxh1q7cnCYwAY=
=Y6Wo
-----END PGP SIGNATURE-----
lmpx.com only provides a reader for public news (NNTP) servers. It is not affiliated with the servers or forums shown here and is not responsible for the content of articles, which is written by their respective authors.