In my humble opinion, "configure" part of ipsec-tools is broken. Here's why:
1. It doesn't detect the correct version of OpenSSL:
checking openssl version... too old
configure: error: OpenSSL version must be 0.9.6 or higher. Aborting.
[root@... ~]# rpm -q openssl
openssl-0.9.7a-40
note that adding "--with-openssl=/usr" is a tested work-around.
2. On some Linux machines it cannot detect kernel support for NAT. When forced to accept kernel NAT (by hacking "configure" script), it correctly determines that the types present are rfc,00,02.
3. src/include-glibc and rpm directories are empty:
config.status: creating src/include-glibc/Makefile
config.status: error: cannot find input file: src/include-glibc/Makefile.in
config.status: creating rpm/Makefile
config.status: error: cannot find input file: rpm/Makefile.in
4. FWD detection is broken: it says that FWD is not supported:
configure:26416: checking whether we support FWD policy
configure:26441: gcc -c -g -O2 -I/usr/include conftest.c >&5
In file included from conftest.c:66:
/usr/include/linux/ipsec.h:18:22: net/sock.h: No such file or directory
conftest.c: In function `main':
conftest.c:72: error: `IPSEC_DIR_FWD' undeclared (first use in this function)
conftest.c:72: error: (Each undeclared identifier is reported only once
conftest.c:72: error: for each function it appears in.)
while on Linux 2.6.11 "IPSEC_DIR_FWD" is part of enum located in "linux/ipsec.h" file in the kernel includes:
enum {
IPSEC_DIR_ANY = 0,
IPSEC_DIR_INBOUND = 1,
IPSEC_DIR_OUTBOUND = 2,
IPSEC_DIR_FWD = 3, /* It is our own */
IPSEC_DIR_MAX = 4,
IPSEC_DIR_INVALID = 5
};
P.S. I don't know whose "our own" is - but it's not my hack. :-)
In short - tarball still not working.
----- Original Message -----
From: manu@... (Emmanuel Dreyfus)
Date: Tuesday, June 28, 2005 6:00 pm
Subject: Re: [Ipsec-tools-devel] 0.6 doesn't configure
> <urimobile@...> wrote:
>
> > I solved those problems by simply copying stuff from relevant
> 0.6rc1> directories to 0.6. Probably the release tarball should be
> > updated/fixed.
>
> I fixed and updated the tarball. It has the same name, which makes
> sensesince it's the same tag in CVS, but it can be confusing.
> Hint: the
> previous (wrong) file is 443088 bytes long, the good file is 660938
> bytes long.
>
> --
> Emmanuel Dreyfus
> http://hcpnet.free.fr/pubz
> manu@...
>
|