Please build, test, and report! We'd love to hear about your experiences with this release.
== Notes ==
* This release of Sage contains a GCC (GNU Compiler Collection) package. This includes compilers for C, C++ and Fortran. GCC is not always built, whether or not it is built is determined automatically. You can use the environment variable SAGE_INSTALL_GCC to override the default. Set SAGE_INSTALL_GCC=no if you don't want to build GCC or SAGE_INSTALL_GCC=yes to force building GCC. * With the addition of GCC, the gfortran binaries for OS X have been removed. gfortran is part of GCC and will be built from source. * The moin moin Wiki has been removed from Sage, saving 28MB in the compressed tarball.
== Tickets ==
* We closed 447 tickets in this release. For details, see
#2102: add incoming/outgoing wrappers to HG objects (like hg_sage) [Reviewed by Mike Hansen] #4780: relative number field constructor -- error message when given poly of degree < 1 is bad [Reviewed by Mike Hansen, David Loeffler] #7038: ratpoints 2.1.2.p2 ignores CC and uses gcc whatever [Reviewed by Leif Leonhardy] #8125: problem with "text" in matplotlib [Reviewed by John Palmieri] #11702: interfaces/magma.py test fails [Reviewed by Marco Streng, David Loeffler] #11875: Correct general brokenness of Farey symbols [Reviewed by David Loeffler] #12004: copying a linear program using Coin solver consumes enormous amounts of memory [Reviewed by Nathann Cohen]
Merged in sage-5.0.beta13:
#9563: Mike Hansen: Remove the English-language tutorial's Makefile [Reviewed by Jeroen Demeyer] #12011: Jeroen Demeyer, John Palmieri: cvxopt: fix illegal BLAS call and fix Solaris build [Reviewed by Jeroen Demeyer, John Palmieri] #12112: John Palmieri, Jeroen Demeyer: Update the prereq script [Reviewed by Jeroen Demeyer, David Kirkby] #12220: John Perry, Nathann Cohen: Updated CBC spkg [Reviewed by John Perry, Nathann Cohen] #12369: Jeroen Demeyer: Add a gcc package [Reviewed by Simon King] #12515: Jeroen Demeyer: Upgrade mpc and make it a standard package [Reviewed by Jean-Pierre Flori, Volker Braun] #12576: John Palmieri: OS X Lion: don't require setting SAGE_PORT [Reviewed by Jeroen Demeyer] #12613: John Palmieri: Add option "-c" to sage-spkg to run the test-suite [Reviewed by Jeroen Demeyer] #12631: Jeroen Demeyer: Get rid of spkg/base/dir-0.1-install [Reviewed by John Palmieri] #12655: Alexander Dreyer: Update PolyBoRi to release 0.8.1 [Reviewed by Martin Albrecht, Jeroen Demeyer, Leif Leonhardy] #12668: David Loeffler: Delete sage/rings/coerce_python.py [Reviewed by Jeroen Demeyer] #12713: John Palmieri: Excise MoinMoin [Reviewed by Jeroen Demeyer] #12784: John Palmieri: Add comment to deps explaining dependency of cvxopt on matplotlib [Reviewed by Karl-Dieter Crisman] #12799: Alexander Dreyer: Fix PolyBoRi's dependencies in `module_list.py` [Reviewed by Leif Leonhardy] #12805: John Palmieri: Do not create SAGE_TESTDIR/tmp [Reviewed by Leif Leonhardy] #12814: Jeroen Demeyer: Add prereq-0.9-install to .hgignore [Reviewed by John Palmieri]
Built on 10.6.8 with no errors after export SAGE_PORT=1 (this is normal for me as I guess I have the wrong Xcode but the right gcc). However I got a bunch of doctest failures when ran sage -t -long /devel/sage-main/sage, they are all basically of the following type:
********************************************************************** File "/Users/Starx/sage-dev/devel/sage-main/sage/schemes/elliptic_curves/modular _parametrization.py", line 17: sage: phi = EllipticCurve('11a1').modular_parametrization() Exception raised: Traceback (most recent call last): File "/Users/Starx/sage-dev/local/bin/ncadoctest.py", line 1231, in run_one_test self.run_one_example(test, example, filename, compileflags) File "/Users/Starx/sage-dev/local/bin/sagedoctest.py", line 38, in run_one_example OrigDocTestRunner.run_one_example(self, test, example, filename, compileflags) File "/Users/Starx/sage-dev/local/bin/ncadoctest.py", line 1172, in run_one_example compileflags, 1) in test.globs File "<doctest __main__.example_0[2]>", line 1, in <module> phi = EllipticCurve('11a1').modular_parametrization()###line 17: sage: phi = EllipticCurve('11a1').modular_parametrization() File "/Users/Starx/sage-dev/local/lib/python/site-packages/sage/schemes/elliptic _curves/constructor.py", line 301, in EllipticCurve return ell_rational_field.EllipticCurve_rational_field(x) File "/Users/Starx/sage-dev/local/lib/python/site-packages/sage/schemes/elliptic _curves/ell_rational_field.py", line 198, in __init__ X = sage.databases.cremona.CremonaDatabase()[label] File "/Users/Starx/sage-dev/local/lib/python/site-packages/sage/databases/cremon a.py", line 1363, in CremonaDatabase _db = LargeCremonaDatabase(name) File "/Users/Starx/sage-dev/local/lib/python/site-packages/sage/databases/cremon a.py", line 1119, in __init__ + 'not appear to be a valid SQL Cremona database.') RuntimeError: Database at /Users/Starx/sage-dev/data//cremona/cremona.db does not appear to be a valid SQL Cremona database. **********************************************************************
The file /Users/Starx/sage-dev/data/cremona/cremona.db does exist (not sure why there's a double slash in the error message...) but it is empty.
> Please build, test, and report! We'd love to hear about your > experiences with this release.
> == Notes ==
> * This release of Sage contains a GCC (GNU Compiler Collection) package. > This includes compilers for C, C++ and Fortran. GCC is not always > built, whether or not it is built is determined automatically. You > can use the environment variable SAGE_INSTALL_GCC to override the > default. Set SAGE_INSTALL_GCC=no if you don't want to build GCC or > SAGE_INSTALL_GCC=yes to force building GCC. > * With the addition of GCC, the gfortran binaries for OS X have been > removed. gfortran is part of GCC and will be built from source. > * The moin moin Wiki has been removed from Sage, saving 28MB in the > compressed tarball.
> == Tickets ==
> * We closed 447 tickets in this release. For details, see
> #2102: add incoming/outgoing wrappers to HG objects (like hg_sage) > [Reviewed by Mike Hansen] > #4780: relative number field constructor -- error message when given > poly of degree < 1 is bad [Reviewed by Mike Hansen, David Loeffler] > #7038: ratpoints 2.1.2.p2 ignores CC and uses gcc whatever [Reviewed by > Leif Leonhardy] > #8125: problem with "text" in matplotlib [Reviewed by John Palmieri] > #11702: interfaces/magma.py test fails [Reviewed by Marco Streng, David > Loeffler] > #11875: Correct general brokenness of Farey symbols [Reviewed by David > Loeffler] > #12004: copying a linear program using Coin solver consumes enormous > amounts of memory [Reviewed by Nathann Cohen]
> Merged in sage-5.0.beta13:
> #9563: Mike Hansen: Remove the English-language tutorial's Makefile > [Reviewed by Jeroen Demeyer] > #12011: Jeroen Demeyer, John Palmieri: cvxopt: fix illegal BLAS call and > fix Solaris build [Reviewed by Jeroen Demeyer, John Palmieri] > #12112: John Palmieri, Jeroen Demeyer: Update the prereq script > [Reviewed by Jeroen Demeyer, David Kirkby] > #12220: John Perry, Nathann Cohen: Updated CBC spkg [Reviewed by John > Perry, Nathann Cohen] > #12369: Jeroen Demeyer: Add a gcc package [Reviewed by Simon King] > #12515: Jeroen Demeyer: Upgrade mpc and make it a standard package > [Reviewed by Jean-Pierre Flori, Volker Braun] > #12576: John Palmieri: OS X Lion: don't require setting SAGE_PORT > [Reviewed by Jeroen Demeyer] > #12613: John Palmieri: Add option "-c" to sage-spkg to run the > test-suite [Reviewed by Jeroen Demeyer] > #12631: Jeroen Demeyer: Get rid of spkg/base/dir-0.1-install [Reviewed > by John Palmieri] > #12655: Alexander Dreyer: Update PolyBoRi to release 0.8.1 [Reviewed by > Martin Albrecht, Jeroen Demeyer, Leif Leonhardy] > #12668: David Loeffler: Delete sage/rings/coerce_python.py [Reviewed by > Jeroen Demeyer] > #12713: John Palmieri: Excise MoinMoin [Reviewed by Jeroen Demeyer] > #12784: John Palmieri: Add comment to deps explaining dependency of > cvxopt on matplotlib [Reviewed by Karl-Dieter Crisman] > #12799: Alexander Dreyer: Fix PolyBoRi's dependencies in > `module_list.py` [Reviewed by Leif Leonhardy] > #12805: John Palmieri: Do not create SAGE_TESTDIR/tmp [Reviewed by Leif > Leonhardy] > #12814: Jeroen Demeyer: Add prereq-0.9-install to .hgignore [Reviewed by > John Palmieri]
> -- > You received this message because you are subscribed to the Google Groups "sage-release" group. > To post to this group, send email to sage-release@googlegroups.com. > To unsubscribe from this group, send email to sage-release+unsubscribe@googlegroups.com. > For more options, visit this group at http://groups.google.com/group/sage-release?hl=en.
-- Die Dunkelheit... leitet die Musik. Die Musik... leitet die Seele.
On Mon, Apr 9, 2012 at 13:23, Starx <jst...@gmail.com> wrote: > Built on 10.6.8 with no errors after export SAGE_PORT=1 (this is > normal for me as I guess I have the wrong Xcode but the right gcc). > However I got a bunch of doctest failures when ran sage -t -long > /devel/sage-main/sage, they are all basically of the following type:
> ********************************************************************** > File "/Users/Starx/sage-dev/devel/sage-main/sage/schemes/elliptic_curves/modular _parametrization.py", > line 17: > sage: phi = EllipticCurve('11a1').modular_parametrization() > Exception raised: > Traceback (most recent call last): > File "/Users/Starx/sage-dev/local/bin/ncadoctest.py", line 1231, > in run_one_test > self.run_one_example(test, example, filename, compileflags) > File "/Users/Starx/sage-dev/local/bin/sagedoctest.py", line 38, > in run_one_example > OrigDocTestRunner.run_one_example(self, test, example, > filename, compileflags) > File "/Users/Starx/sage-dev/local/bin/ncadoctest.py", line 1172, > in run_one_example > compileflags, 1) in test.globs > File "<doctest __main__.example_0[2]>", line 1, in <module> > phi = EllipticCurve('11a1').modular_parametrization()###line 17: > sage: phi = EllipticCurve('11a1').modular_parametrization() > File "/Users/Starx/sage-dev/local/lib/python/site-packages/sage/schemes/elliptic _curves/constructor.py", > line 301, in EllipticCurve > return ell_rational_field.EllipticCurve_rational_field(x) > File "/Users/Starx/sage-dev/local/lib/python/site-packages/sage/schemes/elliptic _curves/ell_rational_field.py", > line 198, in __init__ > X = sage.databases.cremona.CremonaDatabase()[label] > File "/Users/Starx/sage-dev/local/lib/python/site-packages/sage/databases/cremon a.py", > line 1363, in CremonaDatabase > _db = LargeCremonaDatabase(name) > File "/Users/Starx/sage-dev/local/lib/python/site-packages/sage/databases/cremon a.py", > line 1119, in __init__ > + 'not appear to be a valid SQL Cremona database.') > RuntimeError: Database at > /Users/Starx/sage-dev/data//cremona/cremona.db does not appear to be a > valid SQL Cremona database. > **********************************************************************
> The file /Users/Starx/sage-dev/data/cremona/cremona.db does exist (not > sure why there's a double slash in the error message...) but it is > empty.
> -Jim
The file should probably not exist (since it should only exist if you have installed the optional cremona database). It was probably created by accident at some point (one of the optional doctests will do this). See http://trac.sagemath.org/sage_trac/ticket/12341.
>> Please build, test, and report! We'd love to hear about your >> experiences with this release.
>> == Notes ==
>> * This release of Sage contains a GCC (GNU Compiler Collection) package. >> This includes compilers for C, C++ and Fortran. GCC is not always >> built, whether or not it is built is determined automatically. You >> can use the environment variable SAGE_INSTALL_GCC to override the >> default. Set SAGE_INSTALL_GCC=no if you don't want to build GCC or >> SAGE_INSTALL_GCC=yes to force building GCC. >> * With the addition of GCC, the gfortran binaries for OS X have been >> removed. gfortran is part of GCC and will be built from source. >> * The moin moin Wiki has been removed from Sage, saving 28MB in the >> compressed tarball.
>> == Tickets ==
>> * We closed 447 tickets in this release. For details, see
>> #2102: add incoming/outgoing wrappers to HG objects (like hg_sage) >> [Reviewed by Mike Hansen] >> #4780: relative number field constructor -- error message when given >> poly of degree < 1 is bad [Reviewed by Mike Hansen, David Loeffler] >> #7038: ratpoints 2.1.2.p2 ignores CC and uses gcc whatever [Reviewed by >> Leif Leonhardy] >> #8125: problem with "text" in matplotlib [Reviewed by John Palmieri] >> #11702: interfaces/magma.py test fails [Reviewed by Marco Streng, David >> Loeffler] >> #11875: Correct general brokenness of Farey symbols [Reviewed by David >> Loeffler] >> #12004: copying a linear program using Coin solver consumes enormous >> amounts of memory [Reviewed by Nathann Cohen]
>> Merged in sage-5.0.beta13:
>> #9563: Mike Hansen: Remove the English-language tutorial's Makefile >> [Reviewed by Jeroen Demeyer] >> #12011: Jeroen Demeyer, John Palmieri: cvxopt: fix illegal BLAS call and >> fix Solaris build [Reviewed by Jeroen Demeyer, John Palmieri] >> #12112: John Palmieri, Jeroen Demeyer: Update the prereq script >> [Reviewed by Jeroen Demeyer, David Kirkby] >> #12220: John Perry, Nathann Cohen: Updated CBC spkg [Reviewed by John >> Perry, Nathann Cohen] >> #12369: Jeroen Demeyer: Add a gcc package [Reviewed by Simon King] >> #12515: Jeroen Demeyer: Upgrade mpc and make it a standard package >> [Reviewed by Jean-Pierre Flori, Volker Braun] >> #12576: John Palmieri: OS X Lion: don't require setting SAGE_PORT >> [Reviewed by Jeroen Demeyer] >> #12613: John Palmieri: Add option "-c" to sage-spkg to run the >> test-suite [Reviewed by Jeroen Demeyer] >> #12631: Jeroen Demeyer: Get rid of spkg/base/dir-0.1-install [Reviewed >> by John Palmieri] >> #12655: Alexander Dreyer: Update PolyBoRi to release 0.8.1 [Reviewed by >> Martin Albrecht, Jeroen Demeyer, Leif Leonhardy] >> #12668: David Loeffler: Delete sage/rings/coerce_python.py [Reviewed by >> Jeroen Demeyer] >> #12713: John Palmieri: Excise MoinMoin [Reviewed by Jeroen Demeyer] >> #12784: John Palmieri: Add comment to deps explaining dependency of >> cvxopt on matplotlib [Reviewed by Karl-Dieter Crisman] >> #12799: Alexander Dreyer: Fix PolyBoRi's dependencies in >> `module_list.py` [Reviewed by Leif Leonhardy] >> #12805: John Palmieri: Do not create SAGE_TESTDIR/tmp [Reviewed by Leif >> Leonhardy] >> #12814: Jeroen Demeyer: Add prereq-0.9-install to .hgignore [Reviewed by >> John Palmieri]
>> -- >> You received this message because you are subscribed to the Google Groups "sage-release" group. >> To post to this group, send email to sage-release@googlegroups.com. >> To unsubscribe from this group, send email to sage-release+unsubscribe@googlegroups.com. >> For more options, visit this group at http://groups.google.com/group/sage-release?hl=en.
> -- > Die Dunkelheit... leitet die Musik. > Die Musik... leitet die Seele.
> -- > You received this message because you are subscribed to the Google Groups "sage-release" group. > To post to this group, send email to sage-release@googlegroups.com. > To unsubscribe from this group, send email to sage-release+unsubscribe@googlegroups.com. > For more options, visit this group at http://groups.google.com/group/sage-release?hl=en.
> Please build, test, and report! We'd love to hear about your
> experiences with this release.
Hi,
I having trouble building gcc on Ubuntu 11.10 (64 bit). This is in a
vmware VM. The message is: fatal error: bits/predefs.h: No such file
or directory and the build stops. I would guess that I am missing some
devel package that is needed or should everything be included in the
spkg?
I am somewhat surprised that gcc is being built at all since I have
gcc version 4.6.1 (Ubuntu/Linaro 4.6.1-9ubuntu3). However the install
log claims: Installing GCC because your 'gcc' is not so recent.
On Mon, 09 Apr 2012 at 06:04PM -0700, Adam Webb wrote: > I having trouble building gcc on Ubuntu 11.10 (64 bit). This is in a > vmware VM. The message is: fatal error: bits/predefs.h: No such file > or directory and the build stops. I would guess that I am missing some > devel package that is needed or should everything be included in the > spkg?
> I am somewhat surprised that gcc is being built at all since I have > gcc version 4.6.1 (Ubuntu/Linaro 4.6.1-9ubuntu3). However the install > log claims: Installing GCC because your 'gcc' is not so recent.
The failure is in the gcc spkg. I found that compiling gcc on recent versions of Ubuntu is a bit tricky, since they are now using this "multiarch" stuff. I forget which package you need to install, but I tried to compile gcc on my own, and eventually gave up.
On my Ubuntu 10.04 machine, gcc didn't get built, even though it has 4.4.3; I wonder why your build decided to make gcc.
Adam Webb wrote: > On Apr 8, 3:25 pm, Jeroen Demeyer<jdeme...@cage.ugent.be> wrote: >> We're releasing Sage 5.0.beta13.
> I having trouble building gcc on Ubuntu 11.10 (64 bit). This is in a > vmware VM. The message is: fatal error: bits/predefs.h: No such file > or directory and the build stops. I would guess that I am missing some > devel package that is needed or should everything be included in the > spkg?
[/usr/include/]bits/predefs.h is part of the GNU C library, i.e., the Ubuntu package libc6-dev.
Wonder how you would be able to compile anything without it, as most headers include parts from bits/*; probably you just have to reinstall it.
> I am somewhat surprised that gcc is being built at all since I have > gcc version 4.6.1 (Ubuntu/Linaro 4.6.1-9ubuntu3). However the install > log claims: Installing GCC because your 'gcc' is not so recent.
No idea. Smells like your Ubuntu / GCC installation is quite broken.
I've successfully built beta13 on Ubuntu 10.04.4 LTS x86_64 with the native 4.4.3 GCC and SAGE_INSTALL_GCC=yes.
On Precise (12.04 beta), I did have some trouble (with SAGE_INSTALL_GCC=yes), partially apparently due to GCC itself, or how it is configured in the spkg, partially due to Ubuntu's multi-arch setup.
The latter can be solved (on x86_64) by e.g. creating symbolic links for /usr/x86_64-linux-gnu/crt?.o in /usr/lib/.
Note that Ubuntu's current [broken, non-FSF] GCC 4.6.3-1ubuntu3 won't build MPFR (a prerequisite for GCC) with '-O3', and won't build Singular with '-O3' (in CPPFLAGS) either. The former issue is already fixed, but the fix hasn't been committed (and hence not released) yet.
-leif
-- () The ASCII Ribbon Campaign /\ Help Cure HTML E-Mail
On Mon, 09 Apr 2012 at 06:04PM -0700, Adam Webb wrote: > I having trouble building gcc on Ubuntu 11.10 (64 bit). This is in a > vmware VM. The message is: fatal error: bits/predefs.h: No such file > or directory and the build stops. I would guess that I am missing some > devel package that is needed or should everything be included in the > spkg?
> I am somewhat surprised that gcc is being built at all since I have > gcc version 4.6.1 (Ubuntu/Linaro 4.6.1-9ubuntu3). However the install > log claims: Installing GCC because your 'gcc' is not so recent.
Building beta13 on my Ubuntu 11.10 machine is also failing. I see that SAGE_ROOT/spkg/install distrusts any gcc version 4.6.1, which is what I have.
I've attached the log. Based on my experiences getting gcc to build on recent versions of Ubuntu, I think we should disable that check in spkg/install on Ubuntu. Comments?
> I've attached the log. Based on my experiences getting gcc to build on > recent versions of Ubuntu, I think we should disable that check in > spkg/install on Ubuntu. Comments?
Please create a file "errno.c" containing just: #include <errno.h>
Then do $ gcc -v -E errno.c -o errno.i
and send me the output of that command as well as the resulting errno.i file.
On Tue, 10 Apr 2012 at 08:44AM +0200, Jeroen Demeyer wrote: > On 2012-04-10 08:15, Dan Drake wrote: > > I've attached the log. Based on my experiences getting gcc to build on > > recent versions of Ubuntu, I think we should disable that check in > > spkg/install on Ubuntu. Comments?
Dan Drake wrote: > On Mon, 09 Apr 2012 at 06:04PM -0700, Adam Webb wrote: >> I am somewhat surprised that gcc is being built at all since I have >> gcc version 4.6.1 (Ubuntu/Linaro 4.6.1-9ubuntu3). However the install >> log claims: Installing GCC because your 'gcc' is not so recent.
> Building beta13 on my Ubuntu 11.10 machine is also failing. I see that > SAGE_ROOT/spkg/install distrusts any gcc version 4.6.1, which is what I > have.
> I've attached the log. Based on my experiences getting gcc to build on > recent versions of Ubuntu, I think we should disable that check in > spkg/install on Ubuntu. Comments?
I wouldn't /generally/ disable it on Ubuntu (maybe all Debians?), but the message is plain wrong, and it would certainly be annoying if Sage did per default build GCC on the [currently] latest stable version of Ubuntu, not to mention it failing while trying to.
Note that Ubuntu's GCCs are quite different to those released by the FSF (and more recent than the version triplet suggests), so at least it doesn't make sense to use the output of `gcc -dumpversion` there alone.
[I have to admit I haven't looked at the test in the prereq script, and OTOH cannot tell what issues we had with GCC 4.6.1 on which platforms.]
-leif
-- () The ASCII Ribbon Campaign /\ Help Cure HTML E-Mail
> Please build, test, and report! We'd love to hear about your > experiences with this release.
> == Notes ==
> * This release of Sage contains a GCC (GNU Compiler Collection) package. > This includes compilers for C, C++ and Fortran. GCC is not always > built, whether or not it is built is determined automatically. You > can use the environment variable SAGE_INSTALL_GCC to override the > default. Set SAGE_INSTALL_GCC=no if you don't want to build GCC or > SAGE_INSTALL_GCC=yes to force building GCC. > * With the addition of GCC, the gfortran binaries for OS X have been > removed. gfortran is part of GCC and will be built from source. > * The moin moin Wiki has been removed from Sage, saving 28MB in the > compressed tarball.
> == Tickets ==
> * We closed 447 tickets in this release. For details, see
> #2102: add incoming/outgoing wrappers to HG objects (like hg_sage) > [Reviewed by Mike Hansen] > #4780: relative number field constructor -- error message when given > poly of degree < 1 is bad [Reviewed by Mike Hansen, David Loeffler] > #7038: ratpoints 2.1.2.p2 ignores CC and uses gcc whatever [Reviewed by > Leif Leonhardy] > #8125: problem with "text" in matplotlib [Reviewed by John Palmieri] > #11702: interfaces/magma.py test fails [Reviewed by Marco Streng, David > Loeffler] > #11875: Correct general brokenness of Farey symbols [Reviewed by David > Loeffler] > #12004: copying a linear program using Coin solver consumes enormous > amounts of memory [Reviewed by Nathann Cohen]
> Merged in sage-5.0.beta13:
> #9563: Mike Hansen: Remove the English-language tutorial's Makefile > [Reviewed by Jeroen Demeyer] > #12011: Jeroen Demeyer, John Palmieri: cvxopt: fix illegal BLAS call and > fix Solaris build [Reviewed by Jeroen Demeyer, John Palmieri] > #12112: John Palmieri, Jeroen Demeyer: Update the prereq script > [Reviewed by Jeroen Demeyer, David Kirkby] > #12220: John Perry, Nathann Cohen: Updated CBC spkg [Reviewed by John > Perry, Nathann Cohen] > #12369: Jeroen Demeyer: Add a gcc package [Reviewed by Simon King] > #12515: Jeroen Demeyer: Upgrade mpc and make it a standard package > [Reviewed by Jean-Pierre Flori, Volker Braun] > #12576: John Palmieri: OS X Lion: don't require setting SAGE_PORT > [Reviewed by Jeroen Demeyer] > #12613: John Palmieri: Add option "-c" to sage-spkg to run the > test-suite [Reviewed by Jeroen Demeyer] > #12631: Jeroen Demeyer: Get rid of spkg/base/dir-0.1-install [Reviewed > by John Palmieri] > #12655: Alexander Dreyer: Update PolyBoRi to release 0.8.1 [Reviewed by > Martin Albrecht, Jeroen Demeyer, Leif Leonhardy] > #12668: David Loeffler: Delete sage/rings/coerce_python.py [Reviewed by > Jeroen Demeyer] > #12713: John Palmieri: Excise MoinMoin [Reviewed by Jeroen Demeyer] > #12784: John Palmieri: Add comment to deps explaining dependency of > cvxopt on matplotlib [Reviewed by Karl-Dieter Crisman] > #12799: Alexander Dreyer: Fix PolyBoRi's dependencies in > `module_list.py` [Reviewed by Leif Leonhardy] > #12805: John Palmieri: Do not create SAGE_TESTDIR/tmp [Reviewed by Leif > Leonhardy] > #12814: Jeroen Demeyer: Add prereq-0.9-install to .hgignore [Reviewed by > John Palmieri]
> On Mon, 09 Apr 2012 at 06:04PM -0700, Adam Webb wrote: >> I having trouble building gcc on Ubuntu 11.10 (64 bit). This is in a >> vmware VM. The message is: fatal error: bits/predefs.h: No such file >> or directory and the build stops. I would guess that I am missing some >> devel package that is needed or should everything be included in the >> spkg?
>> I am somewhat surprised that gcc is being built at all since I have >> gcc version 4.6.1 (Ubuntu/Linaro 4.6.1-9ubuntu3). However the install >> log claims: Installing GCC because your 'gcc' is not so recent.
> Building beta13 on my Ubuntu 11.10 machine is also failing. I see that > SAGE_ROOT/spkg/install distrusts any gcc version 4.6.1, which is what I > have.
> I've attached the log. Based on my experiences getting gcc to build on > recent versions of Ubuntu, I think we should disable that check in > spkg/install on Ubuntu. Comments?
I propose to simply *not* install GCC when this multiarch situation is detected. I think it will be a lot easier than patching GCC to deal with it.
On Tue, 10 Apr 2012 at 04:19PM +0900, Dan Drake wrote: > The problem is gcc isn't looking in the right places in Ubuntu: see > http://gcc.gnu.org/bugzilla/show_bug.cgi?id=48879. I've set > C_INCLUDE_PATH=/usr/include/x86_64-linux-gnu and the build continued, > although it failed when trying to link crti.o, which required setting > LIBRARY_PATH; see > https://bugs.launchpad.net/ubuntu/+source/binutils/+bug/738098. I have > gcc building right now, we'll see if it finishes.
The gcc spkg built successfully with the following variables set:
(as described in the gcc bugzilla link above), and
LIBRARY_PATH==/usr/lib/x86_64-linux-gnu
(as described in the Launchpad link)
On 32-bit systems, it seems like we'll need "i386", of course.
Here's some information on Debian/Ubuntu's multiarch setup, which they acknowledge breaks a lot of traditional assumptions about paths, but seems to solve a lot of problems:
> On Tue, 10 Apr 2012 at 04:19PM +0900, Dan Drake wrote: >> The problem is gcc isn't looking in the right places in Ubuntu: see >> http://gcc.gnu.org/bugzilla/show_bug.cgi?id=48879. I've set >> C_INCLUDE_PATH=/usr/include/x86_64-linux-gnu and the build continued, >> although it failed when trying to link crti.o, which required setting >> LIBRARY_PATH; see >> https://bugs.launchpad.net/ubuntu/+source/binutils/+bug/738098. I have >> gcc building right now, we'll see if it finishes.
> The gcc spkg built successfully with the following variables set:
> I've tried to build beta13 inside an i386 virtualbox with a fresh > install of ubuntu 12.04 beta 2 plus the m4 package (no internet > connection, so I'm lazy to download anything else). > gcc builds correctly, but then MPIR fails because there is no g++. > Is this intended that g++ is not built or?
That's a known problem which should be fixed by #11616.
On Tuesday, April 10, 2012 2:05:40 PM UTC+2, Jeroen Demeyer wrote:
> On 2012-04-10 13:49, Jean-Pierre Flori wrote: > > I've tried to build beta13 inside an i386 virtualbox with a fresh > > install of ubuntu 12.04 beta 2 plus the m4 package (no internet > > connection, so I'm lazy to download anything else). > > gcc builds correctly, but then MPIR fails because there is no g++. > > Is this intended that g++ is not built or?
> That's a known problem which should be fixed by #11616.
Thanks, I located the ticket meanwhile and am currently building Sage again.
Now the MPIR spkg builds fine, but the GCC one fails while looking for bits/predefs.h This file is present on my system (within the libc6-dev ubuntu package which is installed by default), but I guess it is looking for its own file.
On Tuesday, April 10, 2012 2:33:47 PM UTC+2, Jean-Pierre Flori wrote:
> On Tuesday, April 10, 2012 2:05:40 PM UTC+2, Jeroen Demeyer wrote:
>> On 2012-04-10 13:49, Jean-Pierre Flori wrote: >> > I've tried to build beta13 inside an i386 virtualbox with a fresh >> > install of ubuntu 12.04 beta 2 plus the m4 package (no internet >> > connection, so I'm lazy to download anything else). >> > gcc builds correctly, but then MPIR fails because there is no g++. >> > Is this intended that g++ is not built or?
>> That's a known problem which should be fixed by #11616.
> Thanks, I located the ticket meanwhile and am currently building Sage > again.
> Now the MPIR spkg builds fine, but the GCC one fails while looking for > bits/predefs.h > This file is present on my system (within the libc6-dev ubuntu package > which is installed by default), but I guess it is looking for its own file.
This is a known problem with recent versions of Debian/Ubuntu. You will probably need to install g++ and gfortran yourself and install Sage with SAGE_INSTALL_GCC=no.
On Tuesday, April 10, 2012 3:06:14 PM UTC+2, Jeroen Demeyer wrote:
> On 2012-04-10 14:59, Jean-Pierre Flori wrote: > > Now the MPIR spkg builds fine, but the GCC one fails while looking for > > bits/predefs.h > > This file is present on my system (within the libc6-dev ubuntu package > > which is installed by default), but I guess it is looking for its own > file. > This is a known problem with recent versions of Debian/Ubuntu. You will > probably need to install g++ and gfortran yourself and install Sage with > SAGE_INSTALL_GCC=no.
Ok, so I'll have to download the packages and dependencies on another computer. (By the way my error output is similar to what was mentioned here http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=639752, but I guess this is also known.)
Jean-Pierre Flori wrote: > Build and passed make ptest without problem using the system ATLAS lib > on a amd64 Ubuntu 12.04 beta from one or two weeks.
:-) This time took me three attempts to get a segfault-free ATLAS library on a 2nd generation Core i5 (with AVX).
I first got the "usual" three segfaults in three files, all in the same ATLAS function, then I got a different one (in a single file, different ATLAS function); after building it a third time, ptestlong passed all tests.
On the same machine, Ubuntu 12.04 beta x86_64, in a build with SAGE_INSTALL_GCC=yes, I get a doctest error presumably due to numeric noise.
-leif
-- () The ASCII Ribbon Campaign /\ Help Cure HTML E-Mail
There is an "as" process taking up nearly 3 gigabytes of RAM. I don't quite think it's stuck, but I've never seen this behavior when building Sage. I'll leave my system swapping overnight (it's late here) and we'll see if it makes it through.
In any case, the gcc spkg and friends do not seem to be working well on recent versions of Ubuntu.
> There is an "as" process taking up nearly 3 gigabytes of RAM. I don't > quite think it's stuck, but I've never seen this behavior when building > Sage. I'll leave my system swapping overnight (it's late here) and we'll > see if it makes it through.
I had a similar problem with PPL's mipproblem1 IIRC (with SAGE_CHECK=yes, and *without* the GCC spkg), where g++ consumed about 5 GB, and afterwards the assembler about 7(!) GB on an asm file of ~1 GB.
(Don't recall right now which GCC exactly that was, and with which compiler flags; but I think I reported that here on sage-release as well.)
-leif
-- () The ASCII Ribbon Campaign /\ Help Cure HTML E-Mail
On Sun, Apr 8, 2012 at 2:25 PM, Jeroen Demeyer <jdeme...@cage.ugent.be> wrote: > Dear Sage lovers,
> We're releasing Sage 5.0.beta13.
I dropped this on my OS X 10.7.x laptop, didn't type anything special, typed "make ptestlong", and came back later to see "All tests passed!". I.e., it built perfectly.
Note that I've upgraded both the OS and XCode to the latest released version on this laptop. I think this is reasonable to require by anybody building Sage from source on OS X 10.7, given that XCode is free.
> Please build, test, and report! We'd love to hear about your > experiences with this release.
> == Notes ==
> * This release of Sage contains a GCC (GNU Compiler Collection) package. > This includes compilers for C, C++ and Fortran. GCC is not always > built, whether or not it is built is determined automatically. You > can use the environment variable SAGE_INSTALL_GCC to override the > default. Set SAGE_INSTALL_GCC=no if you don't want to build GCC or > SAGE_INSTALL_GCC=yes to force building GCC. > * With the addition of GCC, the gfortran binaries for OS X have been > removed. gfortran is part of GCC and will be built from source. > * The moin moin Wiki has been removed from Sage, saving 28MB in the > compressed tarball.
> == Tickets ==
> * We closed 447 tickets in this release. For details, see
> #2102: add incoming/outgoing wrappers to HG objects (like hg_sage) > [Reviewed by Mike Hansen] > #4780: relative number field constructor -- error message when given > poly of degree < 1 is bad [Reviewed by Mike Hansen, David Loeffler] > #7038: ratpoints 2.1.2.p2 ignores CC and uses gcc whatever [Reviewed by > Leif Leonhardy] > #8125: problem with "text" in matplotlib [Reviewed by John Palmieri] > #11702: interfaces/magma.py test fails [Reviewed by Marco Streng, David > Loeffler] > #11875: Correct general brokenness of Farey symbols [Reviewed by David > Loeffler] > #12004: copying a linear program using Coin solver consumes enormous > amounts of memory [Reviewed by Nathann Cohen]
> Merged in sage-5.0.beta13:
> #9563: Mike Hansen: Remove the English-language tutorial's Makefile > [Reviewed by Jeroen Demeyer] > #12011: Jeroen Demeyer, John Palmieri: cvxopt: fix illegal BLAS call and > fix Solaris build [Reviewed by Jeroen Demeyer, John Palmieri] > #12112: John Palmieri, Jeroen Demeyer: Update the prereq script > [Reviewed by Jeroen Demeyer, David Kirkby] > #12220: John Perry, Nathann Cohen: Updated CBC spkg [Reviewed by John > Perry, Nathann Cohen] > #12369: Jeroen Demeyer: Add a gcc package [Reviewed by Simon King] > #12515: Jeroen Demeyer: Upgrade mpc and make it a standard package > [Reviewed by Jean-Pierre Flori, Volker Braun] > #12576: John Palmieri: OS X Lion: don't require setting SAGE_PORT > [Reviewed by Jeroen Demeyer] > #12613: John Palmieri: Add option "-c" to sage-spkg to run the > test-suite [Reviewed by Jeroen Demeyer] > #12631: Jeroen Demeyer: Get rid of spkg/base/dir-0.1-install [Reviewed > by John Palmieri] > #12655: Alexander Dreyer: Update PolyBoRi to release 0.8.1 [Reviewed by > Martin Albrecht, Jeroen Demeyer, Leif Leonhardy] > #12668: David Loeffler: Delete sage/rings/coerce_python.py [Reviewed by > Jeroen Demeyer] > #12713: John Palmieri: Excise MoinMoin [Reviewed by Jeroen Demeyer] > #12784: John Palmieri: Add comment to deps explaining dependency of > cvxopt on matplotlib [Reviewed by Karl-Dieter Crisman] > #12799: Alexander Dreyer: Fix PolyBoRi's dependencies in > `module_list.py` [Reviewed by Leif Leonhardy] > #12805: John Palmieri: Do not create SAGE_TESTDIR/tmp [Reviewed by Leif > Leonhardy] > #12814: Jeroen Demeyer: Add prereq-0.9-install to .hgignore [Reviewed by > John Palmieri]
> -- > You received this message because you are subscribed to the Google Groups "sage-release" group. > To post to this group, send email to sage-release@googlegroups.com. > To unsubscribe from this group, send email to sage-release+unsubscribe@googlegroups.com. > For more options, visit this group at http://groups.google.com/group/sage-release?hl=en.
-- William Stein Professor of Mathematics University of Washington http://wstein.org
