Ticket #1252 (closed defect: fixed)
2.8.13: cddlib fails to build with gmp in /usr/local/ [OSX 10.4]
| Reported by: | mabshoff | Owned by: | mabshoff |
|---|---|---|---|
| Priority: | major | Milestone: | sage-2.8.14 |
| Component: | packages | Keywords: | |
| Cc: | Author(s): | ||
| Report Upstream: | Reviewer(s): | ||
| Merged in: | Work issues: |
Description (last modified by mabshoff) (diff)
A remark: this ticket is now about cddlib miscompiling if there is a "rogue" gmp in /usr/local. The 10.5.1 issue is larger than gmp and now #1253.
Michael
Justin Walker reported:
I did a full build of 2.8.13 on two systems:
Mac OS X/10.4.11: Dual Quad-Core Xeon
Mac OS X/10.5.1: Core Duo
Problems on both:
10.5.1: blow-up in Flint
ld: duplicate symbol ___gmpz_abs in test-support.o and fmpz_poly-
test.o
10.4.11: blow-up in cddlib ("/usr/local" contamination)
/usr/libexec/gcc/i686-apple-darwin8/4.0.1/ld: Undefined symbols:
___gmpq_init
...
FWIW, I used "-j6" on the 10.4 system and "-j2" on the 10.5 system.
Rerunning the builds without the "j factor" gave me the (more or
less) the same result on 10.4 and 10.5.
The full logs (of the "-j" builds) are in ~justin/logs on
sage.math.washington.edu (tagged with 10.x).
Justin
See http://groups.google.com/group/sage-devel/t/8e446357a1d15a8a
Cheers,
Michael
Change History
Note: See
TracTickets for help on using
tickets.
