compiling gpg 1.2.0 on HP-UX

David Shaw dshaw@jabberwocky.com
Wed Sep 25 17:49:01 2002


On Wed, Sep 25, 2002 at 07:43:42AM -0700, David Ellement wrote:
> Hello,
> 
> I encountered a couple of problems getting gpg 1.2.0 to compile on HP-UX
> with gcc (3.1.1).
> 
> The compile failed in the mpi directory, due to assembler errors.  Some
> of the assembly code generated had instructions that lacked a leading
> tab character.  Jay Dawes realized the problematic assembler code was in
> mpi/longlong.h.  After adding leading tabs to all the assembler
> instructions in the HP-UX section, the compile succeeded and all the
> checks passed.

Which lines is this in longlong.h?  Is it 318 and 365-380 ?

> The keyserver/Makefile only includes -lldap in gpgkeys_ldap_LDADD, so
> gpgkeys_ldap is unable to resolve host names.  By adding -llber, it is
> able to resolve host names.  (On my linux system at home, I saw the same
> problem, and needed to add -llber -lresolv).

This doesn't make total sense.  The autoconf check should have
detected that.  Did you end up with a gpgkeys_ldap that didn't resolve
host names, or did it not compile at all?

> Has the format for keyserver name changed?  The manual still says:
> 
>     `scheme:[//]keyservername[:port]'
> 
> but it seems the `//' is now required, at least for the hkp scheme.

It's required for any scheme that is host-based (RFC 2396).  For the
schemes that ship with GnuPG, that means both hkp and ldap, but not
mailto.

David

-- 
   David Shaw  |  dshaw@jabberwocky.com  |  WWW http://www.jabberwocky.com/
+---------------------------------------------------------------------------+
   "There are two major products that come out of Berkeley: LSD and UNIX.
      We don't believe this to be a coincidence." - Jeremy S. Anderson