fetchmail-friends
[Top] [All Lists]

Re: [fetchmail]fetchmail 5.7.4: build problems continue

2001-03-13 05:43:22
On Mon, 12 Mar 2001, Steve VanDevender wrote:
 > The tarball should have aclocal.m4 (for dependency purposes), and configure
 > already generated.

The fetchmail 5.7.4 tarball does not contain aclocal.m4.  configure has
already been generated and has a later timestamp than configure.in.

Well, that's something for Eric to fix. Debian source packages do include
aclocal.m4.

I have installed no components of autoconf whatsoever.  Early in the
configure process, these items are reported:

checking for working aclocal... missing
checking for working autoconf... missing
checking for working automake... missing
checking for working autoheader... missing

That looks about right. This should force no calls to aclocal, autoconf and
autoheader to exist in Makefile. Can you verify this, please?

It looks like this may be an interaction between the libc5 headers and
the Linux kernel headers (I have 2.2.19pre16).  This error appears when
compiling ipv6-connect.c:

You really should not do that, you know (never mind the braindamaged HOWTOs
that preach it). The "linux kernel headers" one has in the default include
path should ALWAYS be the ones the libc was built against, NOT the ones
matching the kernel in your system.  Major breakage can happen otherwise.

ipv6-connect.c: In function `sa_len':
Oh, THAT macro. Should be easy enough to fix, I suppose.

Perhaps the #ifdef AF_INET6 checks should be wrapped in #ifdef
INET6_ENABLE?  It actually looks like configure is doing the right thing

Yeah, I think so. It shouldn't break anything to only define those if we're
actually going to need them.

-- 
  "One disk to rule them all, One disk to find them. One disk to bring
  them all and in the darkness grind them. In the Land of Redmond
  where the shadows lie." -- The Silicon Valley Tarot
  Henrique Holschuh