Date: Mon, 4 Aug 2003 00:38:45 -0400 (EDT) From: <brian-bugs@bbdab.org> To: FreeBSD-gnats-submit@FreeBSD.org Subject: ports/55225: Python Numeric port build failure Message-ID: <20030804043845.492514030@mail1.bbdab.org> Resent-Message-ID: <200308040440.h744eH6O055226@freefall.freebsd.org>
next in thread | raw e-mail | index | archive | help
>Number: 55225 >Category: ports >Synopsis: Python Numeric port build failure >Confidential: no >Severity: non-critical >Priority: medium >Responsible: freebsd-ports-bugs >State: open >Quarter: >Keywords: >Date-Required: >Class: change-request >Submitter-Id: current-users >Arrival-Date: Sun Aug 03 21:40:17 PDT 2003 >Closed-Date: >Last-Modified: >Originator: Brian Betts <brian-bugs@bbdab.org> >Release: FreeBSD 4.8-RC i386 >Organization: Strudelline Networks >Environment: System: FreeBSD bb.strudelline.net 4.8-RC FreeBSD 4.8-RC #0: Thu Apr 3 06:51:01 EST 2003 notme@bb.strudelline.net:/usr/obj/usr/src/sys/EL_BUBBLY_ONE i386 >Description: When building ports/math/py-numeric, the following condensed errors are displayed: (wrapped for safety) cc -fno-strict-aliasing -DNDEBUG -O -pipe -D_THREAD_SAFE -DTHREAD_STACK_SIZE=0x20000 -O -pipe -fPIC -DHAVE_INVERSE_HYPERBOLIC -IInclude -IPackages/FFT/Include -IPackages/RNG/Include -I/usr/local/include/python2.3 -c Src/umathmodule.c -o build/temp.freebsd-4.8-STABLE-i386-2.3/Src/umathmodule.o Src/umathmodule.c:1952: `acosh' undeclared here (not in a function) Src/umathmodule.c:1952: initializer element is not constant Src/umathmodule.c:1952: (near initialization for `arccosh_data[0]') Src/umathmodule.c:1952: `acosh' undeclared here (not in a function) Src/umathmodule.c:1952: initializer element is not constant Src/umathmodule.c:1952: (near initialization for `arccosh_data[1]') Src/umathmodule.c:1953: `asinh' undeclared here (not in a function) I have not built this module since I needed it for the Python 2.2 series. I am not sure if the new Python 2.3 changes are what may be contributing to this problem regarding reading the distutils setup files. >How-To-Repeat: Try to build the py-numeric port. >Fix: Immediately when I saw the familiar math library functions not being defined, I checked for math.h being included (it is) and the -lm flag passed (which it's not) to the compiler/linker. I looked in the setup.py utility, used to aid in compilation and installation of the software, and it seems to know about -lm as indicated by the line: mathlibs = ['m'] Which, I believe, should add the -lm flag when compiling, though it does not, as shown from the previous cc line. Hope this helps fix the problem. >Release-Note: >Audit-Trail: >Unformatted:
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20030804043845.492514030>