Opened 2 years ago

Closed 2 years ago

## #22364 closed enhancement (fixed)

Reported by: Owned by: rws major sage-7.6 packages: standard Ralf Stephan Volker Braun N/A 8d7801e (Commits) #22421

Pynac-0.7.5 has

• crucial improvements to Gosper algorithm (#22090)

### comment:2 Changed 2 years ago by rws

• Description modified (diff)
• Status changed from new to needs_review

### comment:3 Changed 2 years ago by rws

• Authors set to Ralf Stephan

### comment:4 Changed 2 years ago by rws

• Commit set to 46697e020a94c9ddb9375dcf92597dd5830f8201

New commits:

 ​46697e0 `pkg version/chksum`

### comment:5 Changed 2 years ago by vbraun

• Reviewers set to Volker Braun
• Status changed from needs_review to positive_review

### comment:6 Changed 2 years ago by vbraun

• Status changed from positive_review to needs_work
```sage -t --long --warn-long 79.1 src/sage/functions/orthogonal_polys.py
**********************************************************************
File "src/sage/functions/orthogonal_polys.py", line 1884, in sage.functions.orthogonal_polys.Func_jacobi_P._eval_
Failed example:
jacobi_P(1,n,n,n)
Expected:
(n + 1)*n
Got:
(n + 1)*((n - 1)*gamma(2*n + 2)/gamma(n + 1) + gamma(2*n + 2)/gamma(n + 1))*gamma(n + 1)/gamma(2*n + 2)
**********************************************************************
File "src/sage/functions/orthogonal_polys.py", line 1888, in sage.functions.orthogonal_polys.Func_jacobi_P._eval_
Failed example:
jacobi_P(1,n,n,x)
Expected:
(n + 1)*x
Got:
(n + 1)*((x - 1)*gamma(2*n + 2)/gamma(n + 1) + gamma(2*n + 2)/gamma(n + 1))*gamma(n + 1)/gamma(2*n + 2)
**********************************************************************
2 of  13 in sage.functions.orthogonal_polys.Func_jacobi_P._eval_
[378 tests, 2 failures, 7.87 s]
----------------------------------------------------------------------
sage -t --long --warn-long 79.1 src/sage/functions/orthogonal_polys.py  # 2 doctests failed
----------------------------------------------------------------------
Total time for all tests: 8.1 seconds
cpu time: 5.3 seconds
cumulative wall time: 7.9 seconds
```

### comment:7 Changed 2 years ago by git

• Commit changed from 46697e020a94c9ddb9375dcf92597dd5830f8201 to 32809571474d35f1dda99986d68dde7c1302b5e8

Branch pushed to git repo; I updated commit sha1. New commits:

 ​20f0b48 `Merge branch 'develop' into t/22364/upgrade_to_pynac_0_7_5` ​3280957 `22364: add normalization step to jacobi_P.eval`

### comment:8 Changed 2 years ago by rws

• Status changed from needs_work to needs_review

### comment:9 Changed 2 years ago by vbraun

• Status changed from needs_review to positive_review

### comment:10 Changed 2 years ago by vbraun

I'm getting this on 32-bit

```sage -t --long src/sage/functions/orthogonal_polys.py
**********************************************************************
File "src/sage/functions/orthogonal_polys.py", line 1886, in sage.functions.orthogonal_polys.Func_jacobi_P._eval_
Failed example:
jacobi_P(2,n,n,n)
Expected:
1/4*(2*n - 1)*(n + 2)*(n + 1)^2
Got:
1/4*(2*n^2 + n - 1)*(n + 2)*(n + 1)
**********************************************************************
1 of  13 in sage.functions.orthogonal_polys.Func_jacobi_P._eval_
[378 tests, 1 failure, 8.03 s]
```

### comment:11 Changed 2 years ago by vbraun

There are some more, possibly related to #22057

```sage -t --long src/sage/symbolic/expression.pyx
**********************************************************************
File "src/sage/symbolic/expression.pyx", line 6729, in sage.symbolic.expression.Expression.resultant
Failed example:
(x+y).resultant(x-y, x)
Expected:
-2*y
Got:
2*x
**********************************************************************
File "src/sage/symbolic/expression.pyx", line 6732, in sage.symbolic.expression.Expression.resultant
Failed example:
r.coefficient(a^4)
Expected:
b^4*y^2 - 4*b^3*y^3 + 6*b^2*y^4 - 4*b*y^5 + y^6
Got:
0
**********************************************************************
2 of   7 in sage.symbolic.expression.Expression.resultant
[2670 tests, 2 failures, 35.16 s]
```

### comment:12 Changed 2 years ago by vbraun

• Status changed from positive_review to needs_work

### comment:13 Changed 2 years ago by rws

What's the easiest way to build a 32bit Sage for debugging this? Volker, can you give me an account?

### comment:14 Changed 2 years ago by rws

Hold on, I confirm there is a problem with resultants when I merge #22057 into this.,

### comment:15 Changed 2 years ago by rws

As there were 32bit problems with Giac I think it would be safer if we used Singular on 32bit instead. Fortunately I provided for such build-time switching. Now this should just be a matter of spkg-install...

### comment:16 Changed 2 years ago by jdemeyer

• Dependencies set to #22421

### comment:17 Changed 2 years ago by rws

On the other hand failure on 32 bit indicates underlying bugs so, sadly, I'll switch off Giac support again until #22101 and #22280 have been sorted out and then try again. If the 32 bit problems persist then at least we know it's not Giac in this case.

### comment:18 Changed 2 years ago by git

• Commit changed from 32809571474d35f1dda99986d68dde7c1302b5e8 to 8d7801ec16d3376adc414f241e94bbd4a5af1fb3

Branch pushed to git repo; I updated commit sha1. New commits:

 ​7cdbdb0 `22421: apply commit from (future) Pynac master` ​0a73359 `22421: clean up spkg-install` ​58f10d5 `Merge branch 'u/rws/pynac_s_to_gamma___is_buggy' of git://trac.sagemath.org/sage into t/22364/upgrade_to_pynac_0_7_5` ​8d7801e `22364: turn off Giac usage of Pynac`

### comment:19 Changed 2 years ago by rws

• Status changed from needs_work to needs_review

### comment:20 Changed 2 years ago by vbraun

• Status changed from needs_review to positive_review

### comment:21 Changed 2 years ago by vbraun

• Resolution set to fixed
• Status changed from positive_review to closed

### comment:22 follow-up: ↓ 23 Changed 2 years ago by fbissey

Something weird is happening to me here in sage-on-gentoo. Tests were OK after merging of this and #22057. Now that we have moved past 7.6.beta6 (I work off Volker's tree usually) I have the failures in `symbolic/expression.pyx` and `functions/orthogonal_polys.py` again. For the later it is even stranger, I get the failure only the first time I run the test, subsequent testing of `functions/orthogonal_polys.py` pass but not `symbolic/expression.pyx`.

Could it be caused by ginac, since I still have it enabled in pynac? But why would it re-appear after 7.6.beta6 has been released?