Re: master branch --- build problems

[lfinsto]

On Thu, July 29, 2010 7:35 pm, Simon Josefsson wrote:
> address@hidden writes:
>> On Thu, July 29, 2010 4:07 pm, Simon Josefsson wrote:
>>> address@hidden writes:

>> I'm not actually very satisfied with OpenSuse and I dislike Suse
Enterprise Edition, but I have to use them.  It doesn't surprise me that
>> I'm having problems.
> Perhaps few people have used this system for bootstrapping.

I could be wrong, of course, but I doubt that anyone has, since I would
expect them to have the same problems.  The version of GNUTLS I got with
the Package Manager is 2.4.1-24.4.1 (x86_64) openSUSE 11.1-Update, and I
just checked that 2.4.1 is from June 2008.  However, it works and it works
together with the other outdated versions of packages you get this way.

Since OpenSUSE is fairly widespread, I think it's worthwhile to try to get
this to work.

>> mv lib/build-aux/config.rpath lib/build-aux/config.rpath-
>> test -f ./configure || autoreconf --install
>> configure.ac:66: error: AC_CHECK_SIZEOF: requires literal arguments
../../lib/autoconf/types.m4:765: AC_CHECK_SIZEOF is expanded from...
lib/m4/hooks.m4:23: LIBGNUTLS_HOOKS is expanded from...
>> configure.ac:66: the top level
> That's the problem we should focus on.  Does it help to change:
>   AC_CHECK_SIZEOF(void *)
>   AC_CHECK_SIZEOF(long)
>   AC_CHECK_SIZEOF(int)
> into
>   AC_CHECK_SIZEOF([void *])
>   AC_CHECK_SIZEOF([long])
>   AC_CHECK_SIZEOF([int])

No.  "long" and "int" don't need to be quoted and quoting "void*" doesn't
work.  However, the manual indicates that it should.  I don't whether a)
the behaviour is correct and the manual hasn't been updated or b) it's a
bug in Autoconf.  I'll look through the Autoconf mailing list archives to
see if this has been discussed and if not, I'll ask about it.

> ?  Don't forget to clean the directory before re-attempting the
> bootstrap.

I don't know whether `make clean' works at this point, but I'll try. 
Otherwise, I can just revert my working copy of the repository.

> You don't need to upgrade glibc at all.  And upgrading gcc is only if
> you want to work on fixing these warnings.  I suggest you wait with this
> until you have a stable development environment.

Again, I could be wrong, but I think the most likely explanation is that
the program assumes at these places that pointers and integers are the
same size, which is not true on my system.  This is usually not difficult
to fix, but I found myself tracing back through declarations of data
types.

I don't think these warnings are that important and it would be more
important to document the data types, which is one of things I wanted to
do in the first place.  If I'm right, getting rid of these warnings is
probably just a matter of adding casts appropriately and/or changing a few
declarations, but I don't want to change anything without understanding
what's going on.

Laurence

-------------------------------------------------------------
Laurence Finston
Gesellschaft fuer wissenschaftliche Datenverarbeitung mbH
Am Fassberg 11
37077 Goettingen

Telefon:        +49 551 201-1882
E-Mail:         address@hidden