umath import error for Numpy builds on OSX 10.6
in [Python]
|
Prev: C to Python
Next: IndentationError
From: Robert Kern on 10 Dec 2009 23:47 On 2009-12-10 17:41 PM, hardcoreUFO wrote: > On Dec 11, 12:29 pm, Robert Kern<robert.k...(a)gmail.com> wrote: >> Don't specify LDFLAGS if you don't have to. It is obliterating the -l flag for >> the npy_math library that numpy builds internally and tries to link with. > > Unfortunately, that did not eliminate the error. I thought it was > something that was changed in Numpy, because my first go at building > was successful on Snow Leopard; it suddenly stopped working a few > weeks ago. Right, when the -lnpymath stuff got checked in. Looking at the build log you posted to numpy-discusson, it does appear that the $LDFLAGS is obliterating the intended flags. Please post a build log without setting those flags to numpy-discussion. A correct link line should look something like this: gcc -arch i386 -arch ppc -isysroot /Developer/SDKs/MacOSX10.4u.sdk -g -bundle -undefined dynamic_lookup build/temp.macosx-10.3-i386-2.5/numpy/core/src/umath/umathmodule_onefile.o -Lbuild/temp.macosx-10.3-i386-2.5 -lnpymath -o build/lib.macosx-10.3-i386-2.5/numpy/core/umath.so -- Robert Kern "I have come to believe that the whole world is an enigma, a harmless enigma that is made terrible by our own mad attempt to interpret it as though it had an underlying truth." -- Umberto Eco
From: hardcoreUFO on 11 Dec 2009 21:55 On Dec 11, 5:47 pm, Robert Kern <robert.k...(a)gmail.com> wrote: > Right, when the -lnpymath stuff got checked in. Looking at the build log you > posted to numpy-discusson, it does appear that the $LDFLAGS is obliterating the > intended flags. Please post a build log without setting those flags to > numpy-discussion. A correct link line should look something like this: > > gcc -arch i386 -arch ppc -isysroot /Developer/SDKs/MacOSX10.4u.sdk -g -bundle > -undefined dynamic_lookup > build/temp.macosx-10.3-i386-2.5/numpy/core/src/umath/umathmodule_onefile.o > -Lbuild/temp.macosx-10.3-i386-2.5 -lnpymath -o > build/lib.macosx-10.3-i386-2.5/numpy/core/umath.so Thanks Robert, Here is the log from a build without the LDFLAGS set. Having a quick look, all I can see are a few warnings and a CAPI version mismatch message. Any insight most appreciated. The build still gives the same umath error. cf
From: hardcoreUFO on 11 Dec 2009 22:00 On Dec 12, 3:55 pm, hardcoreUFO <fonnesb...(a)gmail.com> wrote: > Here is the log from a build without the LDFLAGS set. Having a quick > look, all I can see are a few warnings and a CAPI version mismatch > message. Any insight most appreciated. The build still gives the same > umath error. Sorry, forgot to post the log: http://files.me.com/fonnesbeck/giaoql
From: Robert Kern on 11 Dec 2009 23:44 On 2009-12-11 20:55 PM, hardcoreUFO wrote: > On Dec 11, 5:47 pm, Robert Kern<robert.k...(a)gmail.com> wrote: > >> Right, when the -lnpymath stuff got checked in. Looking at the build log you >> posted to numpy-discusson, it does appear that the $LDFLAGS is obliterating the >> intended flags. Please post a build log without setting those flags to >> numpy-discussion. A correct link line should look something like this: >> >> gcc -arch i386 -arch ppc -isysroot /Developer/SDKs/MacOSX10.4u.sdk -g -bundle >> -undefined dynamic_lookup >> build/temp.macosx-10.3-i386-2.5/numpy/core/src/umath/umathmodule_onefile.o >> -Lbuild/temp.macosx-10.3-i386-2.5 -lnpymath -o >> build/lib.macosx-10.3-i386-2.5/numpy/core/umath.so > > Thanks Robert, > > Here is the log from a build without the LDFLAGS set. Having a quick > look, all I can see are a few warnings and a CAPI version mismatch > message. Any insight most appreciated. The build still gives the same > umath error. On numpy-discussion, please. But first, start with a clean checkout. I think you have stale files. -- Robert Kern "I have come to believe that the whole world is an enigma, a harmless enigma that is made terrible by our own mad attempt to interpret it as though it had an underlying truth." -- Umberto Eco
From: hardcoreUFO on 12 Dec 2009 17:58
On Dec 12, 5:44 pm, Robert Kern <robert.k...(a)gmail.com> wrote: > On 2009-12-11 20:55 PM, hardcoreUFO wrote: > > > > > > > On Dec 11, 5:47 pm, Robert Kern<robert.k...(a)gmail.com> wrote: > > >> Right, when the -lnpymath stuff got checked in. Looking at the build log you > >> posted to numpy-discusson, it does appear that the $LDFLAGS is obliterating the > >> intended flags. Please post a build log without setting those flags to > >> numpy-discussion. A correct link line should look something like this: > > >> gcc -arch i386 -arch ppc -isysroot /Developer/SDKs/MacOSX10.4u.sdk -g -bundle > >> -undefined dynamic_lookup > >> build/temp.macosx-10.3-i386-2.5/numpy/core/src/umath/umathmodule_onefile.o > >> -Lbuild/temp.macosx-10.3-i386-2.5 -lnpymath -o > >> build/lib.macosx-10.3-i386-2.5/numpy/core/umath.so > > > Thanks Robert, > > > Here is the log from a build without the LDFLAGS set. Having a quick > > look, all I can see are a few warnings and a CAPI version mismatch > > message. Any insight most appreciated. The build still gives the same > > umath error. > > On numpy-discussion, please. But first, start with a clean checkout. I think you > have stale files. Log from build with clean checkout posted to the numpy list. Thanks, cf |