Update PolyBoRi to release 0.8.0 #11575

alexanderdreyer · 2011-07-05T12:47:09Z

PolyBoRi 0.8 will come with some changes in the interface (in particular the infamous global settings will be removed)

So there was an extensive alpha phase this time:
https://sourceforge.net/projects/polybori/files/polybori/0.8.alphas/

The final release is here:
https://sourceforge.net/projects/polybori/files/polybori/0.8.0/

Current patches/spkg

Install http://boxen.math.washington.edu/home/dreyer/spkg/polybori-0.8.0.p1.spkg
Apply attachment: polybori-0.8.0.p1.2.patch

Depends on #11574
Depends on #9138

Upstream: None of the above - read trac for reasoning.

CC: @alexanderdreyer @sagetrac-PolyBoRi @kiwifb @burcin @malb @simon-king-jena @jdemeyer

Component: packages: standard

Author: Alexander Dreyer

Reviewer: Martin Albrecht

Merged: sage-5.0.beta1

Issue created by migration from https://trac.sagemath.org/ticket/11575

kiwifb · 2011-07-06T00:03:43Z

comment:2

For the record I have a gentoo ebuild that pulls the trunk of hg on bitbucket. That's how I tested things in the issues I have filled for polybori recently.

alexanderdreyer · 2011-07-06T09:43:10Z

comment:3

I prepared a first at http://boxen.math.washington.edu/home/dreyer/spkg/polybori-0.8-alpha.p0.spkg

It builds, but of course sage -br and tests will fail, since the Sage library needs to be adjusted.

kiwifb · 2011-07-09T04:08:43Z

comment:4

Build using my ebuild (you didn't commit anything since you tagged 0.8-alpha so that's what I should have pulled). Used boost-1.42 and m4ri-20100817. Builds and install fine without warnings. Is there a working test suite that can be run? Also I didn't build any docs.

alexanderdreyer · 2011-07-09T20:42:35Z

comment:5

Hi Francois! I gave to access to polybori-testsuite. If you do use it via the cython interface from Sage's Library (not the original PyPolyBoRi,so). I need to update this, but it's not that easy, because I have to c++-ify some classes without defautl constructor.

kiwifb · 2011-07-10T00:30:34Z

comment:6

Hi Alexander, thanks for giving me access. I think it is best to test it separate from sage at first, especially when you consider that installing it will break sage. I hope that it can work that way.

kiwifb · 2011-07-10T04:23:14Z

comment:7

I just had a thought. I should really try this against m4ri-20110715 (#11574). I suppose this version of polybori is ready for it.

alexanderdreyer · 2011-07-10T19:40:35Z

comment:8

Replying to @kiwifb:

I just had a thought. I should really try this against m4ri-20110715 (#11574). I suppose this version of polybori is ready for it.

It should, but I haven't had time for testing the most recent version yet.

kiwifb · 2011-07-10T19:48:19Z

comment:9

Replying to @alexanderdreyer:

Replying to @kiwifb:

I just had a thought. I should really try this against m4ri-20110715 (#11574). I suppose this version of polybori is ready for it.

It should, but I haven't had time for testing the most recent version yet.

I'll have to wait until the sage server are fully back on, m4ri tarball seem to be stored there. That or I clone m4ri tip from bitbucket, which may be an idea since I already to that for polybori.

kiwifb · 2011-07-11T04:22:39Z

comment:10

Pulled m4ri tip. Polybori-0.7.1 compiles against it, no problem. I'll check 0.8.alpha a bit later and then I'll see about that test suite.

alexanderdreyer · 2011-07-15T22:03:26Z

Attachment: polybori-0.8.0alpha.p1.patch.gz

First working patch for supporting PolyBoRi 0.8 by the Sage-lib

alexanderdreyer · 2011-07-15T22:07:20Z

comment:11

Here's now the first patch for making the Sage-lib supporting PolyBoRi 0.8:
https://github.com/sagemath/sage-prod/files/10653268/polybori-0.8.0alpha.p1.patch.gz

I needs to workaround missing default generators of PolyBoRi-functions by reintroducing then (just for now: It should be using the cppclass construct) . Therefore, we do need another patched spkg: http://boxen.math.washington.edu/home/dreyer/spkg/polybori-0.8-alpha.p1.spkg

kiwifb · 2011-07-15T22:21:32Z

comment:13

I have done a build of your latest tip from a couple of days ago and it was fine. I still have to get organized to do the testing. Nice to have a patch for sage but I am afraid it will need to be rebased now that #11377 is in. Please do not use SAGE_ROOT in module_list.py. In your case SAGE_INC will be enough.

alexanderdreyer · 2011-07-15T22:59:08Z

comment:14

Replying to @kiwifb:

I have done a build of your latest tip from a couple of days ago and it was fine. I still have to get organized to do the testing. Nice to have a patch for sage but I am afraid it will need to be rebased now that #11377 is in. Please do not use SAGE_ROOT in module_list.py. In your case SAGE_INC will be enough.

No problem, it's still work to do anyway, since the c++-ification not really done yet. (I just mis-clicked, it should read "NEEDS_WORKS".)

PS: So that crash-issue on gentoo-prefix got away with 0.8-alpha?

kiwifb · 2011-07-15T23:05:16Z

comment:15

Replying to @alexanderdreyer:

Replying to @kiwifb:

I have done a build of your latest tip from a couple of days ago and it was fine. I still have to get organized to do the testing. Nice to have a patch for sage but I am afraid it will need to be rebased now that #11377 is in. Please do not use SAGE_ROOT in module_list.py. In your case SAGE_INC will be enough.

No problem, it's still work to do anyway, since the c++-ification not really done yet. (I just mis-clicked, it should read "NEEDS_WORKS".)

PS: So that crash-issue on gentoo-prefix got away with 0.8-alpha?

Don't know yet. I have to build this on OS X and apply the patch. It may very well not happen until Monday local time (it's Saturday morning here) since this is one of my work machine.

alexanderdreyer · 2011-07-15T23:21:24Z

comment:16

Don't know yet. I have to build this on OS X and apply the patch. It may very well not happen until Monday local time (it's Saturday morning here) since this is one of my work machine.

Here, too (1:00).

In case this still fails: Can you give the plain PolyBoRi source from bitbucket a try? I had to patch the recent spkg to work around some cython vs. c++ issue.

kiwifb · 2011-07-15T23:28:14Z

comment:18

Replying to @alexanderdreyer:

Don't know yet. I have to build this on OS X and apply the patch. It may very well not happen until Monday local time (it's Saturday morning here) since this is one of my work machine.

Here, too (1:00).

In case this still fails: Can you give the plain PolyBoRi source from bitbucket a try? I had to patch the recent spkg to work around some cython vs. c++ issue.

Not quite lunch time here. I'll make a note to try both. The nature of sage-on-gentoo is that the plain sources will be used over spkg unless the spkg is the only thing that exists or works. We still patch stuff if it is not too big (OK sometimes even when it is big).

kiwifb · 2011-07-25T04:24:23Z

comment:19

Pulled the latest mercurial tip. Installed fine on linux. OS X (10.5.8 on x86) didn't like it one bit. I compiled with -ggdb which may have an impact.

i686-apple-darwin9-g++ -o groebner/libpolybori_groebner.dylib -Wl,-dead_strip_dylibs -dynamiclib groebner/src/groebner.os groebner/s
rc/LiteralFactorization.os groebner/src/LiteralFactorizationIterator.os groebner/src/randomset.os groebner/src/pairs.os groebner/src
/groebner_alg.os groebner/src/FGLMStrategy.os groebner/src/polynomial_properties.os groebner/src/LexBucket.os groebner/src/dlex4data
.os groebner/src/dp_asc4data.os groebner/src/lp4data.os groebner/src/nf.os groebner/src/interpolate.os groebner/src/GroebnerStrategy
.os groebner/src/PairManager.os groebner/src/PolyEntry.os groebner/src/ReductionStrategy.os groebner/src/MatrixMonomialOrderTables.o
s -L/Users/frb15/Desktop/Gentoo/usr/lib -L/Users/frb15/Desktop/Gentoo/usr/lib/python2.7/config -Lpolybori -Lgroebner -LCudd -lm -lgd
 -lgd -lm4ri -lgd -lgd -lm4ri
ld: warning: directory 'polybori' following -L not found

That's the critical bit where it fails. I'll compare with what happens on linux shortly.

Francois who is a bit snowed in today (some more falling).

alexanderdreyer · 2011-07-25T21:45:02Z

comment:20

@kiwifb Please pull again. There was indeed an outdated directory name given, which was ignored on Linux.

malb · 2011-11-07T11:41:34Z

comment:138

Something seems to have gone wrong with the rebased patch:

I

took a fresh 4.7.2,
installed polybori-0.8.0.p1.spkg and
applied attachment: polybori-0.8.0.p1.patch

This gave me a few hunks and one conflict in PolynomialConstruct which I fixed. While it now compiles I get these failures:

File "/opt/sage-4.7.2-polybori-0.8/devel/sage/sage/rings/polynomial/pbori.pyx", line 3859:
    sage: loads(dumps(a)) == a
Exception raised:
    Traceback (most recent call last):
    ...
    AttributeError: ('sage.rings.polynomial.pbori.BooleanPolynomialRing' object has no attribute '_element_constructor_', <built-in function unpickle_BooleanPolynomial0>, (Boolean PolynomialRing in a, b, ([[2], [{{a}}]], Boolean PolynomialRing in a, b)))
**********************************************************************
File "/opt/sage-4.7.2-polybori-0.8/devel/sage/sage/rings/polynomial/pbori.pyx", line 4487:
    sage: PolynomialConstruct()(1)
Exception raised:
    Traceback (most recent call last):
    ...
    ValueError: incomplete format
**********************************************************************
File "/opt/sage-4.7.2-polybori-0.8/devel/sage/sage/rings/polynomial/pbori.pyx", line 7445:
    sage: loads(dumps(a+b)) == a+b # indirect doctest
Exception raised:
    Traceback (most recent call last):
    ...
    AttributeError: ('sage.rings.polynomial.pbori.BooleanPolynomialRing' object has no attribute '_element_constructor_', <built-in function unpickle_BooleanPolynomial0>, (Boolean PolynomialRing in a, b, c, d, ([[2], [{{a}, {b}}, {{b}}]], Boolean PolynomialRing in a, b, c, d)))
**********************************************************************
File "/opt/sage-4.7.2-polybori-0.8/devel/sage/sage/rings/polynomial/pbori.pyx", line 7459:
    sage: loads(dumps(a+b)) == a+b # indirect doctest
Exception raised:
    Traceback (most recent call last):
    ...
    AttributeError: ('sage.rings.polynomial.pbori.BooleanPolynomialRing' object has no attribute '_element_constructor_', <built-in function unpickle_BooleanPolynomial0>, (Boolean PolynomialRing in a, b, c, d, ([[2], [{{a}, {b}}, {{b}}]], Boolean PolynomialRing in a, b, c, d)))
**********************************************************************

So something went wrong and Alexander's element_constructor fixes didn't make it into this patch?

malb · 2011-11-07T11:41:34Z

Changed work issues from cython-0.15.1 to none

jdemeyer · 2011-11-07T13:43:47Z

comment:139

Replying to @malb:

Something seems to have gone wrong with the rebased patch:

I

took a fresh 4.7.2,

installed polybori-0.8.0.p1.spkg and

applied attachment: polybori-0.8.0.p1.patch

Did you apply the dependency #9138?

malb · 2011-11-07T14:01:28Z

comment:140

Ha, no! Sorry for the noise then.

malb · 2011-11-07T18:29:09Z

comment:141

Works as advertised, but I didn't try it with the new Cython yet.

malb · 2011-11-09T12:34:22Z

comment:142

Works with Sage 4.8.alpha1.

jdemeyer · 2011-11-09T20:25:12Z

comment:143

Could you re-export the patch with a username added? Maybe you need to create a $HOME/.hgrc file? See instructions at http://sagemath.org/doc/developer/walk_through.html#submitting-a-change

alexanderdreyer · 2011-11-10T00:46:09Z

comment:144

Thank for pointing this out. I wasn't aware, that hg's mq doesn't write the username into the patch header. Now, I regenerated the patch using hg export tip.

alexanderdreyer · 2011-12-09T23:55:24Z

comment:146

@jdemeyer Since #9138 (and its dependencies) is now about to enter Sage 5.0, could we change the milestone to sage-5.0 here? What do you thing?

jdemeyer · 2012-01-17T16:03:52Z

Rebased to latest #9138

jdemeyer · 2012-01-18T08:15:32Z

comment:148

Attachment: polybori-0.8.0.p1.2.patch.gz

jdemeyer · 2012-01-18T08:15:32Z

Merged: sage-5.0.beta1

alexanderdreyer mannequin added this to the sage-4.7.2 milestone Jul 5, 2011

alexanderdreyer mannequin added c: packages: standard labels Jul 5, 2011

This comment has been minimized.

Sign in to view

alexanderdreyer mannequin added the s: needs review label Jul 15, 2011

This comment has been minimized.

Sign in to view

alexanderdreyer mannequin self-assigned this Jul 15, 2011

alexanderdreyer mannequin added s: needs work and removed s: needs review labels Jul 25, 2011

alexanderdreyer mannequin added s: needs review and removed s: needs work labels Nov 5, 2011

malb added s: needs work and removed s: needs review labels Nov 7, 2011

malb added s: needs review and removed s: needs work labels Nov 7, 2011

malb added s: positive review and removed s: needs review labels Nov 9, 2011

This comment has been minimized.

Sign in to view

jdemeyer added this to the sage-5.0 milestone Dec 9, 2011

jdemeyer removed the pending label Dec 9, 2011

jdemeyer removed the s: positive review label Jan 18, 2012

jdemeyer closed this as completed Jan 18, 2012

sagetrac-drkirkby mannequin mentioned this issue Jan 14, 2023

Potential issue in polybori - 0.5rc.p8 and/or 0.5rc.p9 #6582

Closed

This was referenced Jun 12, 2016

Mac Binary distribution contains hardcoded absolute paths #11681

Closed

PolyBoRi 0.7.1 should obey some standard environment variables #11906

Closed

Merging fixes for PolyBoRi 0.7.1 into 0.8 spkg #11909

Closed

sagetrac-cbwolf mannequin mentioned this issue Jul 4, 2012

polybori crash (Linux) #12654

Closed

Update PolyBoRi to release 0.8.0 #11575

Update PolyBoRi to release 0.8.0 #11575

Comments

alexanderdreyer mannequin commented Jul 5, 2011

Current patches/spkg

This comment has been minimized.

kiwifb commented Jul 6, 2011

This comment has been minimized.

alexanderdreyer mannequin commented Jul 6, 2011

kiwifb commented Jul 9, 2011

alexanderdreyer mannequin commented Jul 9, 2011

kiwifb commented Jul 10, 2011

kiwifb commented Jul 10, 2011

alexanderdreyer mannequin commented Jul 10, 2011

kiwifb commented Jul 10, 2011

kiwifb commented Jul 11, 2011

alexanderdreyer mannequin commented Jul 15, 2011

This comment has been minimized.

alexanderdreyer mannequin commented Jul 15, 2011

This comment has been minimized.

kiwifb commented Jul 15, 2011

alexanderdreyer mannequin commented Jul 15, 2011

kiwifb commented Jul 15, 2011

alexanderdreyer mannequin commented Jul 15, 2011

kiwifb commented Jul 15, 2011

kiwifb commented Jul 25, 2011

alexanderdreyer mannequin commented Jul 25, 2011

malb commented Nov 7, 2011

malb commented Nov 7, 2011

jdemeyer commented Nov 7, 2011

malb commented Nov 7, 2011

malb commented Nov 7, 2011

malb commented Nov 9, 2011

jdemeyer commented Nov 9, 2011

alexanderdreyer mannequin commented Nov 10, 2011

This comment has been minimized.

alexanderdreyer mannequin commented Dec 9, 2011

jdemeyer commented Jan 17, 2012

jdemeyer commented Jan 18, 2012

jdemeyer commented Jan 18, 2012