Hi,
I am using CoCoaLib version 0.9933 on Ubuntu 9.04 x86_64 and I have stumbled upon some strange behaviour when invoking intersect on ideals. Consider the following code:

int main()
{
CoCoA::GlobalManager cocoaFoundations;
CoCoA::FractionField Q=CoCoA::RingQ();
CoCoA::SparsePolyRing Qx = CoCoA::NewPolyRing(Q,1);
vector<RingElem> p,q,r;
p.push_back(indets(Qx)[0]);
q.push_back(1+indets(Qx)[0]);
CoCoA::ideal I(Qx,p);
CoCoA::ideal J(Qx,q);
cout<<I<<endl;
cout<<J<<endl;
cout<<intersect(I,J)<<endl;
cout<<TidyGens(I)<<endl;
cout<<TidyGens(J)<<endl;
cout<<TidyGens(intersect(I,J))<<endl;
ComputeIntersection(r,p,q);
cout<<TidyGens(ideal(Qx,r))<<endl;
return 0;
}

The output I get is the following:

ideal(x[0])
ideal(x[0] +1)
ideal(x[0]^2 +x[0])
[x[0]]
[x[0] +1]
[x[0]^2]
[x[0]^2 +x[0]]


So apparently the intersection is computed correctly, but when I invoke TidyGens I obtain an element which is not in the ideal. However, if I call ComputeIntersection and use the resulting PolyList to create an ideal, I obtain the expected result.

Am I doing something wrong?
Thanks in advance
Diego

Hi Diego

Thanks for the error report. We will investigate over the next few days.

John.
PS sorry for being slow to reply -- we're still getting back up to full speed after the break.

Fixed!
The flag IHaveMonomialGensValue was not updated.

The fixed version 0.9935 is now alailable at http://cocoa.dima.unige.it/cocoalib/changelog-0.99.html

Thanks for the feedback!

Anna

PS there were some improvements in the syntax as well, now your test code can be written like this:

Thanks for your reply, but the fix does not appear to work correctly on my system for two reasons.
1) I get errors on compilation (which I didn't get with 0.9933):

Running the CoCoALib tests (43 tests altogether)
------------------------------------------------
test-empty..... OK
test-MachineInteger1..... OK
test-QQ1..... OK
test-RingZ1..... OK
test-RingZmodN1..... OK
test-RingFp1..... OK
test-RingQ1..... OK
test-convert1..... OK
test-symbol1..... OK
test-NumTheory1..... OK
test-RingTwinFloat1..... OK
***** test-RingTwinFloat2 FAILED ***** (non-zero exit status)
***** test-RingTwinFloat3 FAILED ***** (non-zero exit status)
test-RingTwinFloat4..... OK
test-RingTwinFloat5..... OK
test-RingTwinFloat6..... OK
test-SparsePolyRing1..... OK
test-GradedFreeModule1..... OK
test-module1..... OK
test-matrix1..... OK
test-matrix2..... OK
test-PPMonoid1..... OK
test-OpenMath1..... OK
test-GMPAllocator..... OK
test-IsInteger1..... OK
test-RingHom1..... OK
test-RingHom2..... OK
test-RingHom3..... OK
test-bug1..... OK
***** test-bug2 FAILED ***** (wrong output)
test-bug3..... OK
test-bug4..... OK
***** test-ideal1 FAILED ***** (non-zero exit status)
test-MonomialIdeal1..... OK
./RunTests.sh: line 82: 10489 Segmentation fault ./$prog > $prog.cout 2> $prog.cerr
***** test-GOperations1 FAILED ***** (non-zero exit status)
test-GReductor1..... OK
test-factor1..... OK
test-RingWeyl1..... OK
test-hilbert1..... OK
***** test-F5 FAILED ***** (wrong output)
test-DenseUPolyRing1..... OK
test-TmpIsTree..... OK
test-frobby1..... OK
**********************
***** Bad news *****
**********************
***** The following tests failed, please tell us about it.
***** test-RingTwinFloat2 test-RingTwinFloat3 test-bug2 test-ideal1 test-GOperations1 test-F5


2) my test program now gives the following output:

ideal(x[0])
ideal(x[0] +1)
ideal(x[0])
[x[0]]
[x[0]]
[x[0]]
[x[0]^2]

which is also wrong...
Diego

acc....

In my trying to be *very precise* I found a very subtle bug!
Working on it.

Thanks
Anna

The fixed version 0.9935 is now alailable at
http://cocoa.dima.unige.it/cocoalib/changelog-0.99.html
very subtle uncommon little bug in bool3.....

You will/might get errors with some test-RingTwinFloatN, don't worry, we are working on them and our tests are a lot nastier than normal use

Ciao
Anna

Thanks, now it works fine.
Diego