Should /opt/local/bin/perl be deleted too ?
robert delius royar
apple at frinabulax.org
Fri Dec 21 07:34:00 PST 2007
Fri, 21 Dec 2007 (03:35 +0100 UTC) Vincent Lefevre wrote:
> On 2007-12-20 12:52:48 -0500, Mark Gardner wrote:
>> There are some incompatible changes (see
>> http://search.cpan.org/~rgarcia/perl-5.10.0-RC2/pod/perl5100delta.pod#Incompatible_Changes),
>> so some might want to stick with 5.8.
>
> I'd say that very few modules or scripts should be affected (if any),
> and such problems should be seen as bugs.
Just a note from a simple end user
I compiled perl 5.10.0 with the same configuration that macports 5.8.8
used. [I have the perl5.8 port set as the primary perl interpretor and
have created links in /usr/local/[bin|lib|share] to make it so that
other software finds macports perl before the system one--including
apache.]
When I tried to run a module through 5.10.0 from CPAN (Net::TiVo),
5.10.0 failed because a bundle from macports (p5-digest-sha1) referenced
a symbol not in the 5.10.0 version:
dyld: lazy symbol binding failed: Symbol not found: _Perl_Tstack_sp_ptr
Referenced from: /opt/local/lib/perl5/site_perl/5.10.0/darwin-thread-multi-2level/auto/Digest/SHA1/SHA1.bundle
Expected in: dynamic lookup
dyld: Symbol not found: _Perl_Tstack_sp_ptr
Referenced from: /opt/local/lib/perl5/site_perl/5.10.0/darwin-thread-multi-2level/auto/Digest/SHA1/SHA1.bundle
Expected in: dynamic lookup
Trace/BPT trap
I suspect there may be a number of these. Perl_Tstack_sp_ptr() was in
CORE in 5.8. I believe that it is part of the pre-5.8 legacy code and
in there for compatibility with verions that expected functions for what
are array or incrementable pointer variables.
It is defined in the 5.8.8 source in perlapi.h but not defined anywhere
in the 5.10.0 source.
In a related note
Also, when I watched the return from make test, there were failures in
hints. But it was less than 1% of the overall tests that failed.
I put together a local port file and ran port destroot on the file. The
port file includes the make test. However, beacause the failed tests
caused continues in the make file, the port system did not see any
errors.
--
Dr. Robert Delius Royar Associate Professor of English
Morehead State University Morehead, Kentucky
More information about the macports-dev
mailing list