src/sage/rings/lazy_series_ring.py: test failure for LazyPowerSeriesRing over Zmod(6)

[orlitzky]

I'm not sure why the CI isn't catching this, I can reproduce it at https://sagecell.sagemath.org/

sage: L = LazyPowerSeriesRing(Zmod(6), 's, t')
sage: TestSuite(L).run(skip=['_test_revert'])

Failure in _test_additive_associativity:
Traceback (most recent call last):
  File "/home/sc_serv/sage/src/sage/misc/sage_unittest.py", line 297, in run
    test_method(tester=tester)
  File "/home/sc_serv/sage/src/sage/categories/additive_semigroups.py", line 86, in _test_additive_associativity
    tester.assertEqual((x + y) + z, x + (y + z))
  File "/usr/lib/python3.10/unittest/case.py", line 845, in assertEqual
    assertion_func(first, second, msg=msg)
  File "/usr/lib/python3.10/unittest/case.py", line 838, in _baseAssertEqual
    raise self.failureException(msg)
AssertionError: 3 + s + 4*s^2 + 5*s^3 + 5*s^4 + 2*s^5 + 5*s^6 + O(s,t)^7 != 3 + s + 4*s^2 + 5*s^3 + 5*s^4 + 2*s^5 + 5*s^6 + O(s,t)^7
------------------------------------------------------------
Failure in _test_associativity:
Traceback (most recent call last):
  File "/home/sc_serv/sage/src/sage/misc/sage_unittest.py", line 297, in run
    test_method(tester=tester)
  File "/home/sc_serv/sage/src/sage/categories/semigroups.py", line 121, in _test_associativity
    tester.assertEqual((x * y) * z, x * (y * z))
  File "/usr/lib/python3.10/unittest/case.py", line 845, in assertEqual
    assertion_func(first, second, msg=msg)
  File "/usr/lib/python3.10/unittest/case.py", line 838, in _baseAssertEqual
    raise self.failureException(msg)
AssertionError: s^2 + s^3 + 4*s^4 + 5*s^5 + 5*s^6 + 2*s^7 + 5*s^8 + O(s,t)^9 != s^2 + s^3 + 4*s^4 + 5*s^5 + 5*s^6 + 2*s^7 + 5*s^8 + O(s,t)^9
------------------------------------------------------------
Failure in _test_distributivity:
Traceback (most recent call last):
  File "/home/sc_serv/sage/src/sage/misc/sage_unittest.py", line 297, in run
    test_method(tester=tester)
  File "/home/sc_serv/sage/src/sage/categories/distributive_magmas_and_additive_magmas.py", line 82, in _test_distributivity
    tester.assertEqual((x + y) * z, (x * z) + (y * z))
  File "/usr/lib/python3.10/unittest/case.py", line 845, in assertEqual
    assertion_func(first, second, msg=msg)
  File "/usr/lib/python3.10/unittest/case.py", line 838, in _baseAssertEqual
    raise self.failureException(msg)
AssertionError: 2 + 2*s + 2*s^2 + 4*s^3 + 4*s^4 + 4*s^5 + 4*s^6 + O(s,t)^7 != 2 + 2*s + 2*s^2 + 4*s^3 + 4*s^4 + 4*s^5 + 4*s^6 + O(s,t)^7
------------------------------------------------------------
Failure in _test_div:
Traceback (most recent call last):
  File "/home/sc_serv/sage/src/sage/misc/sage_unittest.py", line 297, in run
    test_method(tester=tester)
  File "/home/sc_serv/sage/src/sage/rings/lazy_series_ring.py", line 1173, in _test_div
    tester.assertEqual(xx, x, "the element (%s) should be the quotient times the divisor (%s)" % (x, y))
  File "/usr/lib/python3.10/unittest/case.py", line 845, in assertEqual
    assertion_func(first, second, msg=msg)
  File "/usr/lib/python3.10/unittest/case.py", line 838, in _baseAssertEqual
    raise self.failureException(msg)
AssertionError: the element (1) should be the quotient times the divisor (1 + s + 4*s^2 + 5*s^3 + 5*s^4 + 2*s^5 + 5*s^6 + O(s,t)^7)
------------------------------------------------------------
The following tests failed: _test_additive_associativity, _test_associativity, _test_distributivity, _test_div

[mantepse]

This is not a bug, you need to set LazyPowerSeriesRing.options.halting_precision, otherwise there is no way the LazyPowerSeriesRing could check equality.

            sage: LazyPowerSeriesRing.options.halting_precision(12)
            sage: L = LazyPowerSeriesRing(Zmod(6), 's, t')
            sage: TestSuite(L).run(skip=['_test_revert'])
            sage: LazyPowerSeriesRing.options._reset()  # reset the options

[orlitzky]

Ok, thank you, that explains why the CI isn't catching it... but not why make ptestlong triggers it on my machine. I'm getting on a plane soon but I can debug it next week.

[mantepse]

Could it be that ptestlong doesn't execute the doc tests in this docstring in one block?

[orlitzky]

Then we're back to why doesn't the CI catch it :)

It's a new machine so possibly some obscure precision issue. I only discounted that explanation at first because I thought I could reproduce it on sagecell.

[orlitzky]

Ok, I'm back. This works:

sage: LazyPowerSeriesRing.options.halting_precision(4)
sage: TestSuite(L).run(skip=['_test_revert'])

but this does not:

sage: LazyPowerSeriesRing.options.halting_precision(5)
sage: TestSuite(L).run(skip=['_test_revert'])
Failure in _test_associativity:
Traceback (most recent call last):
  File "/home/mjo/src/sage.git/local/var/lib/sage/venv-python3.12/lib/python3.12/site-packages/sage/misc/sage_unittest.py", line 298, in run
    test_method(tester=tester)
  File "/home/mjo/src/sage.git/local/var/lib/sage/venv-python3.12/lib/python3.12/site-packages/sage/categories/semigroups.py", line 121, in _test_associativity
    tester.assertEqual((x * y) * z, x * (y * z))
  File "/usr/lib/python3.12/unittest/case.py", line 885, in assertEqual
    assertion_func(first, second, msg=msg)
  File "/usr/lib/python3.12/unittest/case.py", line 878, in _baseAssertEqual
    raise self.failureException(msg)
AssertionError: s + ([61 chars]^4*t+2*s^2*t^3+5*s*t^4) + (s^7+3*s^5*t^2+2*s^4[27 chars],t)^8 != s + ([61 chars]^4*t+4*s^3*t^2+2*s^2*t^3+5*s*t^4) + (s^7+3*s^5[37 chars],t)^8
------------------------------------------------------------
Failure in _test_distributivity:
Traceback (most recent call last):
  File "/home/mjo/src/sage.git/local/var/lib/sage/venv-python3.12/lib/python3.12/site-packages/sage/misc/sage_unittest.py", line 298, in run
    test_method(tester=tester)
  File "/home/mjo/src/sage.git/local/var/lib/sage/venv-python3.12/lib/python3.12/site-packages/sage/categories/distributive_magmas_and_additive_magmas.py", line 83, in _test_distributivity
    tester.assertEqual((x + y) * z, (x * z) + (y * z))
  File "/usr/lib/python3.12/unittest/case.py", line 885, in assertEqual
    assertion_func(first, second, msg=msg)
  File "/usr/lib/python3.12/unittest/case.py", line 878, in _baseAssertEqual
    raise self.failureException(msg)
AssertionError: 2 + ([60 chars]3) + (s^5+5*s^4*t+4*s^3*t^2+4*s^2*t^3+5*s*t^4+[36 chars],t)^7 != 2 + ([60 chars]3) + 4*s^2*t^2 + (s^5+5*s^4*t+4*s^3*t^2+4*s^2*[48 chars],t)^7
------------------------------------------------------------
The following tests failed: _test_associativity, _test_distributivity

I'd love to know what elements cause this to fail, but the POS test runner is hiding them, so now I have N+1 problems.

[mantepse]

Hm, I cannot reproduce this on my computer :-(

[orlitzky]

I added print() statements to the test suites to get x,y,z. This is with LazyPowerSeriesRing.options.halting_precision = 5:

x:  s
y:  1 + (s+t) + (s^2+2*s*t+t^2) + (s^3+3*s^2*t+3*s*t^2+t^3) + (s^4+4*s^3*t+4*s*t^3+t^4) + (s^5+5*s^4*t+4*s^3*t^2+4*s^2*t^3+5*s*t^4+t^5) + (s^6+3*s^4*t^2+2*s^3*t^3+3*s^2*t^4+t^6) + O(s,t)^7
z:  1 + (s+t) + (s^2+2*s*t+t^2) + (s^3+3*s^2*t+3*s*t^2+t^3) + (s^4+4*s^3*t+4*s*t^3+t^4) + (s^5+5*s^4*t+4*s^3*t^2+4*s^2*t^3+5*s*t^4+t^5) + (s^6+3*s^4*t^2+2*s^3*t^3+3*s^2*t^4+t^6) + O(s,t)^7
Failure in _test_associativity:
...
------------------------------------------------------------
x:  1
y:  1 + (s+t) + (s^2+2*s*t+t^2) + (s^3+3*s^2*t+3*s*t^2+t^3) + (s^4+4*s^3*t+4*s*t^3+t^4) + (s^5+5*s^4*t+4*s^3*t^2+4*s^2*t^3+5*s*t^4+t^5) + (s^6+3*s^4*t^2+2*s^3*t^3+3*s^2*t^4+t^6) + O(s,t)^7
z:  1 + (s+t) + (s^2+2*s*t+t^2) + (s^3+3*s^2*t+3*s*t^2+t^3) + (s^4+4*s^3*t+4*s*t^3+t^4) + (s^5+5*s^4*t+4*s^3*t^2+4*s^2*t^3+5*s*t^4+t^5) + (s^6+3*s^4*t^2+2*s^3*t^3+3*s^2*t^4+t^6) + O(s,t)^7
Failure in _test_distributivity:
...

[orlitzky]

This is probably the same issue so I'm just going to focus on the simpler associativity property. I got the left/right computed values too now. With,

x:  s
y:  1 + (s+t) + (s^2+2*s*t+t^2) + (s^3+3*s^2*t+3*s*t^2+t^3) + (s^4+4*s^3*t+4*s*t^3+t^4) + (s^5+5*s^4*t+4*s^3*t^2+4*s^2*t^3+5*s*t^4+t^5) + (s^6+3*s^4*t^2+2*s^3*t^3+3*s^2*t^4+t^6) + O(s,t)^7
z:  1 + (s+t) + (s^2+2*s*t+t^2) + (s^3+3*s^2*t+3*s*t^2+t^3) + (s^4+4*s^3*t+4*s*t^3+t^4) + (s^5+5*s^4*t+4*s^3*t^2+4*s^2*t^3+5*s*t^4+t^5) + (s^6+3*s^4*t^2+2*s^3*t^3+3*s^2*t^4+t^6) + O(s,t)^7

I get:

LHS = s + (2*s^2+2*s*t) + (3*s^3+3*s*t^2) + (4*s^4+4*s*t^3) + (5*s^5+2*s^4*t+2*s^2*t^3+5*s*t^4) + (s^7+3*s^5*t^2+2*s^4*t^3+3*s^3*t^4+s*t^6) + O(s,t)^8
RHS = s + (2*s^2+2*s*t) + (3*s^3+3*s*t^2) + (4*s^4+4*s*t^3) + (5*s^5+2*s^4*t+4*s^3*t^2+2*s^2*t^3+5*s*t^4) + (s^7+3*s^5*t^2+2*s^4*t^3+3*s^3*t^4+s*t^6) + O(s,t)^8

I guess the LHS in this case is (x+y) + z and RHS is x + (y + z)? Or is it testing multiplication?

[mantepse]

could you give me precise instructions on how to reproduce this? That is: exact branch (develop?) and input in a fresh session, and perhaps also a diff that makes it print the above LHS and RHS.

Thank you so much for investigating!

[orlitzky]

I wouldn't worry about it too much, you probably won't be able to reproduce it. It's happening on some new RISC-V hardware, and I would bet that I'm the first person to try to run sage (or at least, the sage test suite) on such a machine.

Mostly things are fine, but I've only had the box for about a month, and I'm still sorting out a lot of other minor issues. Until that is done, I'm working on "develop plus a bunch of random PRs needed to get the tests to pass." Eventually develop will work though, and I may even figure out how to type these elements into sage to get something reproducible.

[orlitzky]

Spoke too soon! I'm testing #38770 on my "normal" amd64 PC at work and I see the same issue.

Unless there are surprise giac conversions in this code, my next guess is that it has something to do with the system packages on Gentoo that are being used right now.

[orlitzky]

Argh, after all that... it's because singular forgot how to do multiplication: #38158 (comment)

Second prize goes to github for not showing any results when I search for lazy_series_ring in issues / pull requests.

Sorry for wasting your time, but thanks for trying to help.