From mbraunm at pe2.kph.tuwien.ac.at Thu Aug 9 09:52:26 2001 From: mbraunm at pe2.kph.tuwien.ac.at (Dr. Moritz Braun) Date: Thu, 09 Aug 2001 15:52:26 +0200 Subject: [SciPy-user] build instructions are not available from scipy home page Message-ID: <3B72959A.70137577@pe2.kph.tuwien.ac.at> Hi there! I have realized, that the build instructions are not available in the moment! are they mirrored in some other place!! regards From travis at scipy.org Thu Aug 9 10:26:04 2001 From: travis at scipy.org (Travis N. Vaught) Date: Thu, 9 Aug 2001 09:26:04 -0500 Subject: [SciPy-user] build instructions are not available from scipy home page In-Reply-To: <3B72959A.70137577@pe2.kph.tuwien.ac.at> Message-ID: Apologies for not having this information on the site yet. Here's a (temporary) link that should get you going until we update the site today or tomorrow(Aug 9, 2001 or Aug 10, 2001). http://scipy.org/bld -----Original Message----- From: scipy-user-admin at scipy.org [mailto:scipy-user-admin at scipy.org]On Behalf Of Dr. Moritz Braun Sent: Thursday, August 09, 2001 8:52 AM To: scipy-user at scipy.net Subject: [SciPy-user] build instructions are not available from scipy home page Hi there! I have realized, that the build instructions are not available in the moment! are they mirrored in some other place!! regards _______________________________________________ SciPy-user mailing list SciPy-user at scipy.net http://www.scipy.net/mailman/listinfo/scipy-user From mbraunm at pe2.kph.tuwien.ac.at Thu Aug 9 10:50:23 2001 From: mbraunm at pe2.kph.tuwien.ac.at (Dr. Moritz Braun) Date: Thu, 09 Aug 2001 16:50:23 +0200 Subject: [SciPy-user] build instructions are not available from scipy home page References: Message-ID: <3B72A32F.D3C9D73B@pe2.kph.tuwien.ac.at> "Travis N. Vaught" wrote: > > Apologies for not having this information on the site yet. Here's a > (temporary) link that should get you going until we update the site today or > tomorrow(Aug 9, 2001 or Aug 10, 2001). > > http://scipy.org/bld > > -----Original Message----- > From: scipy-user-admin at scipy.org [mailto:scipy-user-admin at scipy.org]On > Behalf Of Dr. Moritz Braun > Sent: Thursday, August 09, 2001 8:52 AM > To: scipy-user at scipy.net > Subject: [SciPy-user] build instructions are not available from scipy > home page > > Hi there! > > I have realized, that the build instructions are not available in the > moment! > > are they mirrored in some other place!! > > regards > _______________________________________________ > SciPy-user mailing list > SciPy-user at scipy.net > http://www.scipy.net/mailman/listinfo/scipy-user > > _______________________________________________ > SciPy-user mailing list > SciPy-user at scipy.net > http://www.scipy.net/mailman/listinfo/scipy-user Hi Travis thanks a lot! However. this are only the instructions for Windows I tried it on linux and got stuck first with a problem of find Numerics source code which I fixed by copying the source distribution into the same directory and making a soft link from Numerical to Numeric-xxxxx now the required compilations bombed out with trying to link to the libfftw-thread library I am also unsure whether any existing NumPy installation will be touched or not. Thanks Moritz From oliphant.travis at ieee.org Thu Aug 9 05:44:18 2001 From: oliphant.travis at ieee.org (Travis Oliphant) Date: Thu, 9 Aug 2001 09:44:18 +0000 Subject: [SciPy-user] build instructions are not available from scipy home page In-Reply-To: <3B72A32F.D3C9D73B@pe2.kph.tuwien.ac.at> References: <3B72A32F.D3C9D73B@pe2.kph.tuwien.ac.at> Message-ID: <01080909441802.02554@travis> > > thanks a lot! > > However. > this are only the instructions for Windows > I tried it on linux and got > stuck first with a problem of find Numerics source code > which I fixed by copying the source distribution into the same directory > and making a soft link > from Numerical to Numeric-xxxxx > now the required compilations bombed out with trying to link to the > libfftw-thread library > > I am also unsure whether any existing NumPy installation will be touched > or not. > > > Thanks > > Moritz > _______________________________________________ > SciPy-user mailing list > SciPy-user at scipy.net > http://www.scipy.net/mailman/listinfo/scipy-user If you want to build the interface to the fftw libraries you need to have the fftw libraries previously installed (or you can change the line in setup.py to build_fftw=0) We were having trouble making Numeric a subpackage of scipy so I think the installation will currently upgrade your copy of Numeric as well. -Travis Oliphant (the other Travis) From travis at vaught.net Thu Aug 9 11:59:13 2001 From: travis at vaught.net (Travis N. Vaught) Date: Thu, 9 Aug 2001 10:59:13 -0500 Subject: [SciPy-user] build instructions are not available from scipy home page In-Reply-To: <01080909441802.02554@travis> Message-ID: Dr. Moritz, Travis Oliphant is right, the current version of SciPy from CVS will overwrite ("upgrade") an installed Numeric with Numeric 20.1.1. We are currently discussing alternative strategies of packaging Numeric with SciPy while still retaining the Numeric that may be already installed. We hope to have a solution to the Numeric problem out and updated on the CVS within the next day or so. If you'd like to compile fftw on linux, the following commands *should* work from the fftw directory... (they haven't really been checked, though) #These first three lines are necessary to build double-precision libraries ./configure --enable-threads make make install #Now single precision libraries... make clean ./configure --enable-threads --enable-float --enable-type-prefix make make install Travis Vaught -------------------------------------------------- If you want to build the interface to the fftw libraries you need to have the fftw libraries previously installed (or you can change the line in setup.py to build_fftw=0) We were having trouble making Numeric a subpackage of scipy so I think the installation will currently upgrade your copy of Numeric as well. -Travis Oliphant (the other Travis) _______________________________________________ SciPy-user mailing list SciPy-user at scipy.net http://www.scipy.net/mailman/listinfo/scipy-user -----Original Message----- From: scipy-user-admin at scipy.org [mailto:scipy-user-admin at scipy.org]On Behalf Of Travis Oliphant Sent: Thursday, August 09, 2001 4:44 AM To: scipy-user at scipy.org Subject: Re: [SciPy-user] build instructions are not available from scipy home page > > thanks a lot! > > However. > this are only the instructions for Windows > I tried it on linux and got > stuck first with a problem of find Numerics source code > which I fixed by copying the source distribution into the same directory > and making a soft link > from Numerical to Numeric-xxxxx > now the required compilations bombed out with trying to link to the > libfftw-thread library > > I am also unsure whether any existing NumPy installation will be touched > or not. > > > Thanks > > Moritz > _______________________________________________ > SciPy-user mailing list > SciPy-user at scipy.net > http://www.scipy.net/mailman/listinfo/scipy-user From mbraunm at pe2.kph.tuwien.ac.at Fri Aug 10 03:37:33 2001 From: mbraunm at pe2.kph.tuwien.ac.at (Dr. Moritz Braun) Date: Fri, 10 Aug 2001 09:37:33 +0200 Subject: [SciPy-user] build instructions are not available from scipy home page References: Message-ID: <3B738F3D.493DDF39@pe2.kph.tuwien.ac.at> "Travis N. Vaught" wrote: > > Dr. Moritz, > > Travis Oliphant is right, the current version of SciPy from CVS will > overwrite ("upgrade") an installed Numeric with Numeric 20.1.1. > > We are currently discussing alternative strategies of packaging Numeric with > SciPy while still retaining the Numeric that may be already installed. > > We hope to have a solution to the Numeric problem out and updated on the CVS > within the next day or so. > > If you'd like to compile fftw on linux, the following commands *should* work > from the fftw directory... > (they haven't really been checked, though) > > #These first three lines are necessary to build double-precision libraries > > ./configure --enable-threads > make > make install > > #Now single precision libraries... > > make clean > ./configure --enable-threads --enable-float --enable-type-prefix > make > make install > > Travis Vaught > > -------------------------------------------------- > > > If you want to build the interface to the fftw libraries you need to have > the > fftw libraries previously installed (or you can change the line in setup.py > to build_fftw=0) > > We were having trouble making Numeric a subpackage of scipy so I think the > installation will currently upgrade your copy of Numeric as well. > > -Travis Oliphant (the other Travis) > _______________________________________________ > SciPy-user mailing list > SciPy-user at scipy.net > http://www.scipy.net/mailman/listinfo/scipy-user > > -----Original Message----- > From: scipy-user-admin at scipy.org [mailto:scipy-user-admin at scipy.org]On > Behalf Of Travis Oliphant > Sent: Thursday, August 09, 2001 4:44 AM > To: scipy-user at scipy.org > Subject: Re: [SciPy-user] build instructions are not available from > scipy home page > > > > > thanks a lot! > > > > However. > > this are only the instructions for Windows > > I tried it on linux and got > > stuck first with a problem of find Numerics source code > > which I fixed by copying the source distribution into the same directory > > and making a soft link > > from Numerical to Numeric-xxxxx > > now the required compilations bombed out with trying to link to the > > libfftw-thread library > > > > I am also unsure whether any existing NumPy installation will be touched > > or not. > > > > > > Thanks > > > > Moritz > > _______________________________________________ > > SciPy-user mailing list > > SciPy-user at scipy.net > > http://www.scipy.net/mailman/listinfo/scipy-user > > _______________________________________________ > SciPy-user mailing list > SciPy-user at scipy.net > http://www.scipy.net/mailman/listinfo/scipy-user Thanks a lot! setup.py build and setup.py install have now worked thanks! regards Moritz Braun From mbraunm at pe2.kph.tuwien.ac.at Fri Aug 10 07:04:38 2001 From: mbraunm at pe2.kph.tuwien.ac.at (Dr. Moritz Braun) Date: Fri, 10 Aug 2001 13:04:38 +0200 Subject: [SciPy-user] build instructions are not available from scipy home page References: Message-ID: <3B73BFC6.4BEE3F6A@pe2.kph.tuwien.ac.at> "Travis N. Vaught" wrote: > > Dr. Moritz, > > Travis Oliphant is right, the current version of SciPy from CVS will > overwrite ("upgrade") an installed Numeric with Numeric 20.1.1. > > We are currently discussing alternative strategies of packaging Numeric with > SciPy while still retaining the Numeric that may be already installed. > > We hope to have a solution to the Numeric problem out and updated on the CVS > within the next day or so. > > If you'd like to compile fftw on linux, the following commands *should* work > from the fftw directory... > (they haven't really been checked, though) > > #These first three lines are necessary to build double-precision libraries > > ./configure --enable-threads > make > make install > > #Now single precision libraries... > > make clean > ./configure --enable-threads --enable-float --enable-type-prefix > make > make install > > Travis Vaught > > -------------------------------------------------- > > > If you want to build the interface to the fftw libraries you need to have > the > fftw libraries previously installed (or you can change the line in setup.py > to build_fftw=0) > > We were having trouble making Numeric a subpackage of scipy so I think the > installation will currently upgrade your copy of Numeric as well. > > -Travis Oliphant (the other Travis) > _______________________________________________ > SciPy-user mailing list > SciPy-user at scipy.net > http://www.scipy.net/mailman/listinfo/scipy-user > > -----Original Message----- > From: scipy-user-admin at scipy.org [mailto:scipy-user-admin at scipy.org]On > Behalf Of Travis Oliphant > Sent: Thursday, August 09, 2001 4:44 AM > To: scipy-user at scipy.org > Subject: Re: [SciPy-user] build instructions are not available from > scipy home page > > > > > thanks a lot! > > > > However. > > this are only the instructions for Windows > > I tried it on linux and got > > stuck first with a problem of find Numerics source code > > which I fixed by copying the source distribution into the same directory > > and making a soft link > > from Numerical to Numeric-xxxxx > > now the required compilations bombed out with trying to link to the > > libfftw-thread library > > > > I am also unsure whether any existing NumPy installation will be touched > > or not. > > > > > > Thanks > > > > Moritz > > _______________________________________________ > > SciPy-user mailing list > > SciPy-user at scipy.net > > http://www.scipy.net/mailman/listinfo/scipy-user > > _______________________________________________ > SciPy-user mailing list > SciPy-user at scipy.net > http://www.scipy.net/mailman/listinfo/scipy-user Hi Travis and Travis I managed to build and install fftw both threaded and single precision I also got the python build and install going However there seem to be two bugs 1. the Numeric.pth file in the site-packages directory ends up being empty! thus Numeric will not be found , easily fixed by somebody with enougth python experience to know about the .pth files. 2. the fftw.so fftw-module is not found by the FFT2.py module in spite of being in the same directory namely site-packages/scipy/fft regards Moritz From oliphant.travis at ieee.org Fri Aug 10 05:20:38 2001 From: oliphant.travis at ieee.org (Travis Oliphant) Date: Fri, 10 Aug 2001 09:20:38 +0000 Subject: [SciPy-user] build instructions are not available from scipy home page In-Reply-To: <3B73BFC6.4BEE3F6A@pe2.kph.tuwien.ac.at> References: <3B73BFC6.4BEE3F6A@pe2.kph.tuwien.ac.at> Message-ID: <01081009203800.10374@travis> > > I managed to build and install fftw both threaded and single precision > I also got the python build and install going > However there seem to be two bugs > 1. the Numeric.pth file in the site-packages directory ends up being > empty! > thus Numeric will not be found , easily fixed by somebody with enougth > python experience > to know about the .pth files. This is a problem that has been fixed in the latest CVS tree. > 2. the fftw.so fftw-module is not found by the FFT2.py module in spite > of being in the > same directory namely site-packages/scipy/fft I don't understand this problem. Is there an __init__.py file in the scipy/fft directory? Try import scipy.fft.fftw and report what this produces. Thanks and sorry for your trouble, -Travis O. From eric at scipy.org Fri Aug 10 10:30:54 2001 From: eric at scipy.org (eric jones) Date: Fri, 10 Aug 2001 10:30:54 -0400 Subject: [SciPy-user] build instructions are not available from scipy home page References: <3B73BFC6.4BEE3F6A@pe2.kph.tuwien.ac.at> Message-ID: <003d01c121a9$130c5710$c200a8c0@ericlaptop> Hey Moritz, > 1. the Numeric.pth file in the site-packages directory ends up being > empty! thus Numeric will not be found , easily fixed by somebody with enougth > python experience to know about the .pth files. We found this yesterday too. Team Travis hacked together something that should fix this. Its in the latest CVS. > 2. the fftw.so fftw-module is not found by the FFT2.py module in spite > of being in the > same directory namely site-packages/scipy/fft We're not seeing that one here on our Windows or Linux machine -- although I have seen similar behavior in the past. The import behavior for packages (and especially sub-packages) is not always easy to figure out. I tried sticking Numeric in as a sub-package of SciPy and failed because of similar issues to what your seeing here (although I think it is still possible, just not in the initial release). I have two suggestions. (1) Try the latest CVS. It had a lot of work added yesterday and cursory tests on Windows and Linux show that it is generally usable. (2) If that fails, change the 'import fftw' statement to 'from scipy.fft import fftw' and see if that helps any. eric ----- Original Message ----- From: "Dr. Moritz Braun" To: Sent: Friday, August 10, 2001 7:04 AM Subject: Re: [SciPy-user] build instructions are not available from scipy home page > "Travis N. Vaught" wrote: > > > > Dr. Moritz, > > > > Travis Oliphant is right, the current version of SciPy from CVS will > > overwrite ("upgrade") an installed Numeric with Numeric 20.1.1. > > > > We are currently discussing alternative strategies of packaging Numeric with > > SciPy while still retaining the Numeric that may be already installed. > > > > We hope to have a solution to the Numeric problem out and updated on the CVS > > within the next day or so. > > > > If you'd like to compile fftw on linux, the following commands *should* work > > from the fftw directory... > > (they haven't really been checked, though) > > > > #These first three lines are necessary to build double-precision libraries > > > > ./configure --enable-threads > > make > > make install > > > > #Now single precision libraries... > > > > make clean > > ./configure --enable-threads --enable-float --enable-type-prefix > > make > > make install > > > > Travis Vaught > > > > -------------------------------------------------- > > > > > > If you want to build the interface to the fftw libraries you need to have > > the > > fftw libraries previously installed (or you can change the line in setup.py > > to build_fftw=0) > > > > We were having trouble making Numeric a subpackage of scipy so I think the > > installation will currently upgrade your copy of Numeric as well. > > > > -Travis Oliphant (the other Travis) > > _______________________________________________ > > SciPy-user mailing list > > SciPy-user at scipy.net > > http://www.scipy.net/mailman/listinfo/scipy-user > > > > -----Original Message----- > > From: scipy-user-admin at scipy.org [mailto:scipy-user-admin at scipy.org]On > > Behalf Of Travis Oliphant > > Sent: Thursday, August 09, 2001 4:44 AM > > To: scipy-user at scipy.org > > Subject: Re: [SciPy-user] build instructions are not available from > > scipy home page > > > > > > > > thanks a lot! > > > > > > However. > > > this are only the instructions for Windows > > > I tried it on linux and got > > > stuck first with a problem of find Numerics source code > > > which I fixed by copying the source distribution into the same directory > > > and making a soft link > > > from Numerical to Numeric-xxxxx > > > now the required compilations bombed out with trying to link to the > > > libfftw-thread library > > > > > > I am also unsure whether any existing NumPy installation will be touched > > > or not. > > > > > > > > > Thanks > > > > > > Moritz > > > _______________________________________________ > > > SciPy-user mailing list > > > SciPy-user at scipy.net > > > http://www.scipy.net/mailman/listinfo/scipy-user > > > > _______________________________________________ > > SciPy-user mailing list > > SciPy-user at scipy.net > > http://www.scipy.net/mailman/listinfo/scipy-user > Hi Travis and Travis > > I managed to build and install fftw both threaded and single precision > I also got the python build and install going > However there seem to be two bugs > 1. the Numeric.pth file in the site-packages directory ends up being > empty! > thus Numeric will not be found , easily fixed by somebody with enougth > python experience > to know about the .pth files. > 2. the fftw.so fftw-module is not found by the FFT2.py module in spite > of being in the > same directory namely site-packages/scipy/fft > > regards > > Moritz > _______________________________________________ > SciPy-user mailing list > SciPy-user at scipy.net > http://www.scipy.net/mailman/listinfo/scipy-user From mbraunm at pe2.kph.tuwien.ac.at Mon Aug 13 05:31:02 2001 From: mbraunm at pe2.kph.tuwien.ac.at (Dr. Moritz Braun) Date: Mon, 13 Aug 2001 11:31:02 +0200 Subject: [SciPy-user] build instructions are not available from scipy home page References: <3B73BFC6.4BEE3F6A@pe2.kph.tuwien.ac.at> <003d01c121a9$130c5710$c200a8c0@ericlaptop> Message-ID: <3B779E56.CAD3E842@pe2.kph.tuwien.ac.at> eric jones wrote: > > Hey Moritz, > > > 1. the Numeric.pth file in the site-packages directory ends up being > > empty! thus Numeric will not be found , easily fixed by somebody with > enougth > > python experience to know about the .pth files. > > We found this yesterday too. Team Travis hacked together something that > should fix this. Its in the latest CVS. > > > 2. the fftw.so fftw-module is not found by the FFT2.py module in spite > > of being in the > > same directory namely site-packages/scipy/fft > > We're not seeing that one here on our Windows or Linux machine -- although I > have seen similar behavior in the past. The import behavior for packages > (and especially sub-packages) is not always easy to figure out. I tried > sticking Numeric in as a sub-package of SciPy and failed because of similar > issues to what your seeing here (although I think it is still possible, just > not > in the initial release). > > I have two suggestions. (1) Try the latest CVS. It had a lot of work added > yesterday and cursory tests on Windows and Linux show that it is generally > usable. (2) If that fails, change the 'import fftw' statement to > 'from scipy.fft import fftw' and see if that helps any. > > eric > > ----- Original Message ----- > From: "Dr. Moritz Braun" > To: > Sent: Friday, August 10, 2001 7:04 AM > Subject: Re: [SciPy-user] build instructions are not available from scipy > home page > > > "Travis N. Vaught" wrote: > > > > > > Dr. Moritz, > > > > > > Travis Oliphant is right, the current version of SciPy from CVS will > > > overwrite ("upgrade") an installed Numeric with Numeric 20.1.1. > > > > > > We are currently discussing alternative strategies of packaging Numeric > with > > > SciPy while still retaining the Numeric that may be already installed. > > > > > > We hope to have a solution to the Numeric problem out and updated on the > CVS > > > within the next day or so. > > > > > > If you'd like to compile fftw on linux, the following commands *should* > work > > > from the fftw directory... > > > (they haven't really been checked, though) > > > > > > #These first three lines are necessary to build double-precision > libraries > > > > > > ./configure --enable-threads > > > make > > > make install > > > > > > #Now single precision libraries... > > > > > > make clean > > > ./configure --enable-threads --enable-float --enable-type-prefix > > > make > > > make install > > > > > > Travis Vaught > > > > > > -------------------------------------------------- > > > > > > > > > If you want to build the interface to the fftw libraries you need to > have > > > the > > > fftw libraries previously installed (or you can change the line in > setup.py > > > to build_fftw=0) > > > > > > We were having trouble making Numeric a subpackage of scipy so I think > the > > > installation will currently upgrade your copy of Numeric as well. > > > > > > -Travis Oliphant (the other Travis) > > > _______________________________________________ > > > SciPy-user mailing list > > > SciPy-user at scipy.net > > > http://www.scipy.net/mailman/listinfo/scipy-user > > > > > > -----Original Message----- > > > From: scipy-user-admin at scipy.org [mailto:scipy-user-admin at scipy.org]On > > > Behalf Of Travis Oliphant > > > Sent: Thursday, August 09, 2001 4:44 AM > > > To: scipy-user at scipy.org > > > Subject: Re: [SciPy-user] build instructions are not available from > > > scipy home page > > > > > > > > > > > thanks a lot! > > > > > > > > However. > > > > this are only the instructions for Windows > > > > I tried it on linux and got > > > > stuck first with a problem of find Numerics source code > > > > which I fixed by copying the source distribution into the same > directory > > > > and making a soft link > > > > from Numerical to Numeric-xxxxx > > > > now the required compilations bombed out with trying to link to the > > > > libfftw-thread library > > > > > > > > I am also unsure whether any existing NumPy installation will be > touched > > > > or not. > > > > > > > > > > > > Thanks > > > > > > > > Moritz > > > > _______________________________________________ > > > > SciPy-user mailing list > > > > SciPy-user at scipy.net > > > > http://www.scipy.net/mailman/listinfo/scipy-user > > > > > > _______________________________________________ > > > SciPy-user mailing list > > > SciPy-user at scipy.net > > > http://www.scipy.net/mailman/listinfo/scipy-user > > Hi Travis and Travis > > > > I managed to build and install fftw both threaded and single precision > > I also got the python build and install going > > However there seem to be two bugs > > 1. the Numeric.pth file in the site-packages directory ends up being > > empty! > > thus Numeric will not be found , easily fixed by somebody with enougth > > python experience > > to know about the .pth files. > > 2. the fftw.so fftw-module is not found by the FFT2.py module in spite > > of being in the > > same directory namely site-packages/scipy/fft > > > > regards > > > > Moritz > > _______________________________________________ > > SciPy-user mailing list > > SciPy-user at scipy.net > > http://www.scipy.net/mailman/listinfo/scipy-user > > _______________________________________________ > SciPy-user mailing list > SciPy-user at scipy.net > http://www.scipy.net/mailman/listinfo/scipy-user Dear Eric Dear Travis thanks a lot for your help I am moving back to South Africa in the next two weeks, where I will look into scipy again regards Moritz From jochen at jochen-kuepper.de Wed Aug 15 22:01:13 2001 From: jochen at jochen-kuepper.de (Jochen =?iso-8859-1?q?K=FCpper?=) Date: 15 Aug 2001 22:01:13 -0400 Subject: [SciPy-user] build error Message-ID: I am trying to install latest cvs on Linux, python 2.1, Numeric from latest cvs. python.setup.py build gives ,---- | Traceback (most recent call last): | File "setup.py", line 349, in ? | execfile(os.path.join(numeric_base,'Lib','numeric_version.py')) | IOError: [Errno 2] No such file or directory `---- Well, the only numeric_version.h I have is at ,---- | /usr/lib/python2.1/site-packages/Numeric/numeric_version.py `---- Greetings, Jochen -- Einigkeit und Recht und Freiheit http://www.Jochen-Kuepper.de Libert?, ?galit?, Fraternit? GnuPG key: 44BCCD8E Sex, drugs and rock-n-roll From eric at scipy.org Wed Aug 15 22:16:26 2001 From: eric at scipy.org (eric jones) Date: Wed, 15 Aug 2001 22:16:26 -0400 Subject: [SciPy-user] build error References: Message-ID: <034701c125f9$74856c30$c100a8c0@ericlaptop> Hey Jochen, You need to be sure that the 'Numerical' directory you got from the Numeric CVS is a sub-directory of scipy: scipy/Numerical This is where our setup.py file looks for it. Later, we'll hopefully include Numerical in our CVS to make this easier for people. We just haven't decided how best to keep in sync with the real Numerical CVS yet. Hope this helps, eric ----- Original Message ----- From: "Jochen K?pper" To: Sent: Wednesday, August 15, 2001 10:01 PM Subject: [SciPy-user] build error > I am trying to install latest cvs on Linux, python 2.1, Numeric from > latest cvs. > > python.setup.py build gives > ,---- > | Traceback (most recent call last): > | File "setup.py", line 349, in ? > | execfile(os.path.join(numeric_base,'Lib','numeric_version.py')) > | IOError: [Errno 2] No such file or directory > `---- > > Well, the only numeric_version.h I have is at > ,---- > | /usr/lib/python2.1/site-packages/Numeric/numeric_version.py > `---- > > Greetings, > Jochen > -- > Einigkeit und Recht und Freiheit http://www.Jochen-Kuepper.de > Libert?, ?galit?, Fraternit? GnuPG key: 44BCCD8E > Sex, drugs and rock-n-roll > > _______________________________________________ > SciPy-user mailing list > SciPy-user at scipy.net > http://www.scipy.net/mailman/listinfo/scipy-user From jochen at jochen-kuepper.de Thu Aug 16 00:44:36 2001 From: jochen at jochen-kuepper.de (Jochen =?iso-8859-1?q?K=FCpper?=) Date: 16 Aug 2001 00:44:36 -0400 Subject: [SciPy-user] build error In-Reply-To: <034701c125f9$74856c30$c100a8c0@ericlaptop> References: <034701c125f9$74856c30$c100a8c0@ericlaptop> Message-ID: >>>>> eric jones wrote on Wed, 15 Aug 2001 22:16:26 -0400: eric> You need to be sure that the 'Numerical' directory you got from eric> the Numeric CVS is a sub-directory of scipy: Ah, sure that works:) Maybe there should be a README stating that? ,---- | Make sure to have a subdirectory Numerical containing the NumPy cvs | code. `---- Thanks for your help, Jochen -- Einigkeit und Recht und Freiheit http://www.Jochen-Kuepper.de Libert?, ?galit?, Fraternit? GnuPG key: 44BCCD8E Sex, drugs and rock-n-roll From oliphant.travis at ieee.org Thu Aug 16 06:50:45 2001 From: oliphant.travis at ieee.org (Travis Oliphant) Date: Thu, 16 Aug 2001 10:50:45 +0000 Subject: [SciPy-user] build error In-Reply-To: References: Message-ID: <01081610504500.17428@travis> On Thu, 16 Aug 2001, you wrote: > I am trying to install latest cvs on Linux, python 2.1, Numeric from > latest cvs. > > python.setup.py build gives > ,---- > > | Traceback (most recent call last): > | File "setup.py", line 349, in ? > | execfile(os.path.join(numeric_base,'Lib','numeric_version.py')) > | IOError: [Errno 2] No such file or directory > > `---- > > Well, the only numeric_version.h I have is at > ,---- > > | /usr/lib/python2.1/site-packages/Numeric/numeric_version.py > To build SciPy you need to get the Numeric sources and place them in a subdirectory named Numerical under the main SciPy tree. -Travis From eric at scipy.org Wed Aug 22 17:13:53 2001 From: eric at scipy.org (eric jones) Date: Wed, 22 Aug 2001 17:13:53 -0400 Subject: [SciPy-user] Fw: ANN: SciPy 0.10 -- Scientific Computing with Python Message-ID: <013401c12b4f$59ab8bf0$ec88fea9@ericlaptop> ----- Original Message ----- From: "I?igo Serna" To: "eric jones" ; Sent: Wednesday, August 22, 2001 11:53 AM Subject: Re: ANN: SciPy 0.10 -- Scientific Computing with Python > Hello, > > I have problems with SciPy trying to run the example in SciPy FAQ. > > Simplifying, this is where the problem comes from: > > Python 2.1.1 (#5, Aug 5 2001, 02:09:22) > [GCC 2.96 20000731 (Red Hat Linux 7.1 2.96-81)] on linux2 > Type "copyright", "credits" or "license" for more information. > >>> 6.2 - 3.1 > 3.1000000000000001 > >>> import gui_thread > >>> > >>> > >>> 6.2 - 3.1 > 3.0 > > It seems the problem is originated in gui_thread/main.py when executing main.start() > perhaps problems with threads. > My machine is a x86 with 2 PIII CPUs running Linux 2.4.9. > I have compiled manually python 2.1.1, SciPy, wxgtk (with threads support on) and > wxpython. > > Someone else with the same problems? > > Thanks in advance, > > I?igo Serna > From eric at scipy.org Thu Aug 23 15:16:19 2001 From: eric at scipy.org (eric jones) Date: Thu, 23 Aug 2001 15:16:19 -0400 Subject: [SciPy-user] Re: ANN: SciPy 0.10 -- Scientific Computing with Python References: <998495598.19097.25.camel@lethe.infernus.org> Message-ID: <021401c12c08$172bee00$ec88fea9@ericlaptop> That's a strange error. I don't have a RH machine in front of me to test it on, but on my windows box, it doesn't occur. PythonWin 2.1 (#15, Apr 23 2001, 18:00:35) [MSC 32 bit (Intel)] on win32. Portions Copyright 1994-2001 Mark Hammond (MarkH at ActiveState.com) - see 'Help/About PythonWin' for further copyright information. >>> 6.2 - 3.1 3.1000000000000001 >>> import gui_thread >>> >>> 6.2 - 3.1 3.1000000000000001 I'll check it again when a RH Linux box is handy. thanks, eric ----- Original Message ----- From: "I?igo Serna" To: "eric jones" ; Sent: Wednesday, August 22, 2001 11:53 AM Subject: Re: ANN: SciPy 0.10 -- Scientific Computing with Python > Hello, > > I have problems with SciPy trying to run the example in SciPy FAQ. > > Simplifying, this is where the problem comes from: > > Python 2.1.1 (#5, Aug 5 2001, 02:09:22) > [GCC 2.96 20000731 (Red Hat Linux 7.1 2.96-81)] on linux2 > Type "copyright", "credits" or "license" for more information. > >>> 6.2 - 3.1 > 3.1000000000000001 > >>> import gui_thread > >>> > >>> > >>> 6.2 - 3.1 > 3.0 > > It seems the problem is originated in gui_thread/main.py when executing main.start() > perhaps problems with threads. > My machine is a x86 with 2 PIII CPUs running Linux 2.4.9. > I have compiled manually python 2.1.1, SciPy, wxgtk (with threads support on) and > wxpython. > > Someone else with the same problems? > > Thanks in advance, > > I?igo Serna > From eric at scipy.org Thu Aug 23 21:07:54 2001 From: eric at scipy.org (eric jones) Date: Thu, 23 Aug 2001 21:07:54 -0400 Subject: [SciPy-user] Re: ANN: SciPy 0.10 -- Scientific Computing with Python References: <998495598.19097.25.camel@lethe.infernus.org> <021401c12c08$172bee00$ec88fea9@ericlaptop> Message-ID: <000d01c12c39$34d4bdc0$c27ba8c0@ericlaptop> This fix was posted to the python newsgroup. Don't know if it works. Is this an international issue? Why does importing gui_thread (or maybe wx?) screw this up. eric > Hello, > > I have problems with SciPy trying to run the example in SciPy FAQ. > > Simplifying, this is where the problem comes from: > > Python 2.1.1 (#5, Aug 5 2001, 02:09:22) > [GCC 2.96 20000731 (Red Hat Linux 7.1 2.96-81)] on linux2 > Type "copyright", "credits" or "license" for more information. > >>> 6.2 - 3.1 > 3.1000000000000001 > >>> import gui_thread > >>> > >>> > >>> 6.2 - 3.1 > 3.0 > > It seems the problem is originated in gui_thread/main.py when executing main.start() > perhaps problems with threads. > My machine is a x86 with 2 PIII CPUs running Linux 2.4.9. > I have compiled manually python 2.1.1, SciPy, wxgtk (with threads support on) and > wxpython. > > Someone else with the same problems? > > Thanks in advance, > > I?igo Serna > > > > try this after the import of gui_thread import locale locale.setlocale(locale.LC_ALL,"") regards uwe ----- Original Message ----- From: "eric jones" To: "I?igo Serna" Cc: Sent: Thursday, August 23, 2001 3:16 PM Subject: [SciPy-user] Re: ANN: SciPy 0.10 -- Scientific Computing with Python > That's a strange error. I don't have a RH machine in front of me to test it > on, but on > my windows box, it doesn't occur. > > PythonWin 2.1 (#15, Apr 23 2001, 18:00:35) [MSC 32 bit (Intel)] on win32. > Portions Copyright 1994-2001 Mark Hammond (MarkH at ActiveState.com) - see > 'Help/About PythonWin' for further copyright information. > >>> 6.2 - 3.1 > 3.1000000000000001 > >>> import gui_thread > >>> > >>> 6.2 - 3.1 > > 3.1000000000000001 > > I'll check it again when a RH Linux box is handy. > > thanks, > eric > > ----- Original Message ----- > From: "I?igo Serna" > To: "eric jones" ; > Sent: Wednesday, August 22, 2001 11:53 AM > Subject: Re: ANN: SciPy 0.10 -- Scientific Computing with Python > > > > Hello, > > > > I have problems with SciPy trying to run the example in SciPy FAQ. > > > > Simplifying, this is where the problem comes from: > > > > Python 2.1.1 (#5, Aug 5 2001, 02:09:22) > > [GCC 2.96 20000731 (Red Hat Linux 7.1 2.96-81)] on linux2 > > Type "copyright", "credits" or "license" for more information. > > >>> 6.2 - 3.1 > > 3.1000000000000001 > > >>> import gui_thread > > >>> > > >>> > > >>> 6.2 - 3.1 > > 3.0 > > > > It seems the problem is originated in gui_thread/main.py when executing > main.start() > > perhaps problems with threads. > > My machine is a x86 with 2 PIII CPUs running Linux 2.4.9. > > I have compiled manually python 2.1.1, SciPy, wxgtk (with threads support > on) and > > wxpython. > > > > Someone else with the same problems? > > > > Thanks in advance, > > > > I?igo Serna > > > > _______________________________________________ > SciPy-user mailing list > SciPy-user at scipy.net > http://www.scipy.net/mailman/listinfo/scipy-user From eric at scipy.org Fri Aug 24 13:03:24 2001 From: eric at scipy.org (eric jones) Date: Fri, 24 Aug 2001 13:03:24 -0400 Subject: [SciPy-user] Re: scipy References: Message-ID: <00bc01c12cbe$b00f0fb0$c27ba8c0@ericlaptop> > -----Original Message----- > From: Larry Whitley [mailto:ldwhitley at home.com] > Sent: Wednesday, August 22, 2001 9:24 PM > To: Travis N. Vaught > Subject: Re: scipy > > > Thanks, it looks like its working now. Pretty neat. > > Two questions: Is there somewhere I can read more about the graphing > capabilities? In addition to the scientific graphing I'm looking for > business graphics; bar charts, stacked bar charts, etc. Currently, not that I know of plt certainly doesn't, and gplt doesn't really either because gnuplot doesn't handle shading regions. It does do "box plots" which are kind like bar charts, but there not very good for bar charts. >>> from scipy import gplt >>> gplt.plot((1,2,3)),'with boxes') scipy.xplt for Unix might have some bar chart capabilities, but I haven't checked. There are a few other options. Biggles might do it, http://biggles.sourceforge.net/. Bar charts will eventually be part of the plt module, but there isn't a time line for there inclusion yet. > Also, have you had any experience with cow among pcs? If so, do they all > need to be NTs of one flavor or another, or can W95/98/Me participate too? cow will work on Windows, it just isn't as easy to work with. The main problem is that the machine_cluster.start() method doesn't work because rsh/ssh servers and background processing in DOS shells are not easy to use, and I never found the magic incantations to get them all working right with cow. The best thing to do would be to set it up as a service to circumvent this issue. So to try it out, you have to start up the servers by hand. Open a DOS Shell 1 and type: c:\> python c:\Python21\scipy\cow\sync_cluster.py server 10000 Open another shell and type: c:\> python c:\Python21\scipy\cow\sync_cluster.py server 10001 Then start up a python interpreter in a third shell. C:\WINDOWS\SYSTEM32>python ActivePython 2.1, build 210 ActiveState) based on Python 2.1 (#15, Apr 23 2001, 18:00:35) [MSC 32 bit (Intel)] on win32 Type "help", "copyright", "credits" or "license" for more information. >>> import scipy.cow >>> machines = (('eric_laptop',10000),('eric_laptop',10001)) >>> c = scipy.cow.machine_cluster(machines) >>> c['a'] = 1 >>> c['a'] (1, 1) >>> c['b'] = 2 >>> c['b'] (2, 2) >>> import os >>> c.apply(os.getpid) (972, 2684) Substitute your machine name or ip in for eric_laptop. I just noticed that 'localhost' and '127.0.0.1' don't work, so don't use these until the issue is resolved. eric From inigoserna at terra.es Sat Aug 25 11:31:00 2001 From: inigoserna at terra.es (=?ISO-8859-1?Q?I=F1igo?= Serna) Date: 25 Aug 2001 17:31:00 +0200 Subject: [SciPy-user] Re: ANN: SciPy 0.10 -- Scientific Computing with Python In-Reply-To: <3B85665E.2050007@nospam.de> References: <3B85665E.2050007@nospam.de> Message-ID: <998753460.12831.5.camel@lethe.infernus.org> Thanks, it works with that. But now I'm trying to discover where the problem comes from... perhaphs from wxpython? Thanks again, I?igo El d?a 23 Aug 2001 22:23:58 +0200, Uwe Hoffmann escribi? > I?igo Serna wrote: > > > Hello, > > > > I have problems with SciPy trying to run the example in SciPy FAQ. > > > > Simplifying, this is where the problem comes from: > > > > Python 2.1.1 (#5, Aug 5 2001, 02:09:22) > > [GCC 2.96 20000731 (Red Hat Linux 7.1 2.96-81)] on linux2 > > Type "copyright", "credits" or "license" for more information. > > >>> 6.2 - 3.1 > > 3.1000000000000001 > > >>> import gui_thread > > >>> > > >>> > > >>> 6.2 - 3.1 > > 3.0 > > > > It seems the problem is originated in gui_thread/main.py when executing main.start() > > perhaps problems with threads. > > My machine is a x86 with 2 PIII CPUs running Linux 2.4.9. > > I have compiled manually python 2.1.1, SciPy, wxgtk (with threads support on) and > > wxpython. > > > > Someone else with the same problems? > > > > Thanks in advance, > > > > I?igo Serna > > try this after the import of gui_thread > > import locale > locale.setlocale(locale.LC_ALL,"") > > > regards > uwe From inigoserna at terra.es Sat Aug 25 12:09:33 2001 From: inigoserna at terra.es (=?ISO-8859-1?Q?I=F1igo?= Serna) Date: 25 Aug 2001 18:09:33 +0200 Subject: [SciPy-user] [Fwd: Postmaster notify: see transcript for details] Message-ID: <998755774.906.2.camel@lethe.infernus.org> Thanks, it works with that. But now I'm trying to discover where the problem comes from... perhaphs from wxpython? Thanks again, I?igo El d?a 23 Aug 2001 22:23:58 +0200, Uwe Hoffmann escribi? > I?igo Serna wrote: > > > Hello, > > > > I have problems with SciPy trying to run the example in SciPy FAQ. > > > > Simplifying, this is where the problem comes from: > > > > Python 2.1.1 (#5, Aug 5 2001, 02:09:22) > > [GCC 2.96 20000731 (Red Hat Linux 7.1 2.96-81)] on linux2 > > Type "copyright", "credits" or "license" for more information. > > >>> 6.2 - 3.1 > > 3.1000000000000001 > > >>> import gui_thread > > >>> > > >>>=20 > > >>> 6.2 - 3.1 > > 3.0 > > > > It seems the problem is originated in gui_thread/main.py when executing main.start() > > perhaps problems with threads. > > My machine is a x86 with 2 PIII CPUs running Linux 2.4.9. > > I have compiled manually python 2.1.1, SciPy, wxgtk (with threads support on) and > > wxpython. > > > > Someone else with the same problems? > > > > Thanks in advance, > > > > I?igo Serna > > try this after the import of gui_thread > > import locale > locale.setlocale(locale.LC_ALL,"") > > regards > uwe From eric at scipy.org Sat Aug 25 12:50:06 2001 From: eric at scipy.org (eric jones) Date: Sat, 25 Aug 2001 12:50:06 -0400 Subject: [SciPy-user] [Fwd: Postmaster notify: see transcript for details] References: <998755774.906.2.camel@lethe.infernus.org> Message-ID: <021a01c12d85$feed4f30$c27ba8c0@ericlaptop> I think so. There was a recent note about it on wxPython list. http://lists.wxwindows.org/pipermail/wxpython-users/2001-August/007047.html http://lists.wxwindows.org/pipermail/wxpython-users/2001-August/007048.html see ya, eric ----- Original Message ----- From: "I?igo Serna" To: ; Sent: Saturday, August 25, 2001 12:09 PM Subject: [SciPy-user] [Fwd: Postmaster notify: see transcript for details] > Thanks, it works with that. > > But now I'm trying to discover where the problem comes from... perhaphs > from wxpython? > > Thanks again, > > I?igo > > El d?a 23 Aug 2001 22:23:58 +0200, Uwe Hoffmann escribi? > > I?igo Serna wrote: > > > > > Hello, > > > > > > I have problems with SciPy trying to run the example in SciPy FAQ. > > > > > > Simplifying, this is where the problem comes from: > > > > > > Python 2.1.1 (#5, Aug 5 2001, 02:09:22) > > > [GCC 2.96 20000731 (Red Hat Linux 7.1 2.96-81)] on linux2 > > > Type "copyright", "credits" or "license" for more information. > > > >>> 6.2 - 3.1 > > > 3.1000000000000001 > > > >>> import gui_thread > > > >>> > > > >>>=20 > > > >>> 6.2 - 3.1 > > > 3.0 > > > > > > It seems the problem is originated in gui_thread/main.py when executing main.start() > > > perhaps problems with threads. > > > My machine is a x86 with 2 PIII CPUs running Linux 2.4.9. > > > I have compiled manually python 2.1.1, SciPy, wxgtk (with threads support on) and > > > wxpython. > > > > > > Someone else with the same problems? > > > > > > Thanks in advance, > > > > > > I?igo Serna > > > > try this after the import of gui_thread > > > > import locale > > locale.setlocale(locale.LC_ALL,"") > > > > regards > > uwe > > > _______________________________________________ > SciPy-user mailing list > SciPy-user at scipy.net > http://www.scipy.net/mailman/listinfo/scipy-user