Hi Herbert, hi all,
(included are questions that are addressed to the general audience...)
On 2015-05-20, Herbert Eisenbeis <klassikm...@googlemail.com> wrote:
> looks muuuch better.
Yes, the failures seem mostly harmless now. Nonetheless, I wish Hilbert
driven computation would work (it is much faster than without Hilbert
series).
Since there were test failures, I wonder about our policy regarding
optional packages.
So far, I tried to write code that works the same on different (even
rather old) versions of SageMath. So, I use old import paths as backup
when the new import paths are not available, test the Singular version
and work around bugs in old versions, etc.
Many of the current test failures are in fact not more than a changed
output format in recent SageMath versions.
1) For example, it is now not
set([...])
but
{...}
2) It seems that singular.eval('degbound=0') is now NOT returning the given
input string.
3) A plot is now actually resulting in output, in tests.
Some examples are below. Obviously, if I would change the expected
output so that it works with the new SageMath version, it would fail
with old versions.
QUESTION on our policy:
Is it expected that optional packages build and test fine both on old
and new SageMath versions? We can (I think) not necessarily assume
that all users always upgrade SageMath in a timely manner.
Best regards,
Simon
Here are some details:
> There were doctest failures:
> pGroupCohomology:
> Running doctests with ID 2015-05-20-17-43-51-4c646dc5.
> Git branch: develop
> Doctesting 1 file.
> sage -t --long --warn-long 106.3
> /mnt/nt/he-home-shadow/.sage/temp/he-VirtualBox/20501/dir_TDy7et/file_0.py
> **********************************************************************
> File
> "/mnt/nt/he-home-shadow/.sage/temp/he-VirtualBox/20501/dir_TDy7et/file_0.py",
> line 628, in file_0
> Failed example:
> H0.massey_products(H0.2*H0.1,H0.3*H0.1,H0.2*H0.1,H0.3*H0.1, all=False)
> Expected:
> set([c_2_2^5: 10-Cocycle in H^*(D8; GF(2))])
> Got:
> {c_2_2^5: 10-Cocycle in H^*(D8; GF(2))}
Change in ipython, or what is that? Anyway, it is harmless.
> **********************************************************************
> File
> "/mnt/nt/he-home-shadow/.sage/temp/he-VirtualBox/20501/dir_TDy7et/file_0.py",
> line 1881, in file_0
> Failed example:
> singular.eval('degBound=0')
> Expected:
> 'degBound=0;'
> Got:
> ''
Change in pexpect?
> pGroupCohomology.barcode.BarCode.show:
> Running doctests with ID 2015-05-20-17-39-57-4838b5e9.
> Git branch: develop
> Doctesting 1 file.
> sage -t --long --warn-long 106.9
> /mnt/nt/he-home-shadow/.sage/temp/he-VirtualBox/20501/dir_TDy7et/file_17.py
> **********************************************************************
> File
> "/mnt/nt/he-home-shadow/.sage/temp/he-VirtualBox/20501/dir_TDy7et/file_17.py",
> line 24, in file_17
> Failed example:
> show(B,dmin=3,dmax=15)
> Expected nothing
> Got:
> <html><script
> type="math/tex">\newcommand{\Bold}[1]{\mathbf{#1}}\verb|Persistence|\phantom{\verb!x!}\verb|data|\phantom{\verb!x!}\verb|for|\phantom{\verb!x!}\verb|H^*(SmallGroup(16,3);|\phantom{\verb!x!}\verb|GF(2))|\phantom{\verb!x!}\verb|associated|\phantom{\verb!x!}\verb|with|\phantom{\verb!x!}\verb|LowerCentralSeries|</script></html>
How to test against "show"?
> pGroupCohomology.cohomology.COHO._element_constructor_:
> Running doctests with ID 2015-05-20-17-17-09-35ec203d.
> Git branch: develop
> Doctesting 1 file.
> sage -t --long --warn-long 103.1
> /mnt/nt/he-home-shadow/.sage/temp/he-VirtualBox/20501/dir_TDy7et/file_32.py
> **********************************************************************
> File
> "/mnt/nt/he-home-shadow/.sage/temp/he-VirtualBox/20501/dir_TDy7et/file_32.py",
> line 50, in file_32
> Failed example:
> H.subgps[4,2](H.2+H.3)
> Expected:
> Traceback (most recent call last):
> ...
> TypeError: Cochain belongs to a different cohomology ring, namely H^*(D8;
> GF(2))
> Got:
> ...
> TypeError: Hom() got an unexpected keyword argument 'category'
The "category" keyword has been removed? I rather hoped we would finally
be able to use the keyword in order to create the "additive abelian
group" homomorphisms between two rings...
> pGroupCohomology.cohomology.COHO._singular_:
> Running doctests with ID 2015-05-20-17-21-36-7f1fb4e3.
> Git branch: develop
> Doctesting 1 file.
> sage -t --long --warn-long 104.0
> /mnt/nt/he-home-shadow/.sage/temp/he-VirtualBox/20501/dir_TDy7et/file_109.py
> **********************************************************************
> File
> "/mnt/nt/he-home-shadow/.sage/temp/he-VirtualBox/20501/dir_TDy7et/file_109.py",
> line 17, in file_109
> Failed example:
> R1 = singular(H); R1
> Expected:
> // characteristic : 2
> // number of vars : 2
> // block 1 : ordering M
> // : names b_1_0 b_1_1
> // : weights 1 1
> // : weights -1 0
> // block 2 : ordering C
> // quotient ring from ideal
> _[1]=0
> Got:
> // characteristic : 2
> // number of vars : 2
> // block 1 : ordering M
> // : names b_1_0 b_1_1
> // : weights 1 1
> // : weights -1 0
> // block 2 : ordering C
> **********************************************************************
Has singular become clever and does not form a quotient ring if the
quotient relations are zero? Cool!
--
You received this message because you are subscribed to the Google Groups
"sage-devel" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to sage-devel+unsubscr...@googlegroups.com.
To post to this group, send email to sage-devel@googlegroups.com.
Visit this group at http://groups.google.com/group/sage-devel.
For more options, visit https://groups.google.com/d/optout.
