Closed jdemeyer closed 5 years ago
Description changed:
---
+++
@@ -1,5 +1,5 @@
Fix the Cython warning
- +warning: sage/rings/complex_double.pyx:1590:4: Overriding cdef method with def method.
Commit: 61cecb9
Branch pushed to git repo; I updated commit sha1. This was a forced push. New commits:
61cecb9 | Proper powering in CDF |
Patchbot reports some failures:
sage -t --long src/sage/symbolic/random_tests.py # 1 doctest failed
sage -t --long src/sage/modular/modform/numerical.py # 6 doctests failed
There is a strange regression: CDF(-0.26319743704743886) ** -1
now returns a non-real number -3.799429094819641 - 4.652958679558198e-16*I
. The old code didn't seem to do that.
GSL seems to special-case powering by a complex -1 but not by a real -1.
Branch pushed to git repo; I updated commit sha1. This was a forced push. New commits:
7430342 | Proper powering in CDF |
I added a few special cases for powering to an integer.
I agree with the change. Let us wait for a patchbot report.
Failures again reported by the patchbot:
sage -t --long src/sage/symbolic/expression.pyx # 1 doctest failed
sage -t --long src/sage/modular/abvar/torsion_subgroup.py # 1 doctest failed
sage -t --long src/sage/tests/french_book/mpoly.py # 1 doctest failed
although the torsion_subgroup.py
is almost certainly unrelated.
Replying to @tscrim:
the
torsion_subgroup.py
is almost certainly unrelated.
See #27224 for that.
The french book doctest will need to wait for #23572 (let's hope that it will finally get merged now).
Branch pushed to git repo; I updated commit sha1. New commits:
0e0607c | Minor fixes for CDF powering |
So #23572 is going to be a dependency of this? I am guessing you want to wait until that is merged before proceeding.
Dependencies: #23572
Replying to @tscrim:
I am guessing you want to wait until that is merged before proceeding.
Indeed. Let's hope that the next beta will have it.
Changed dependencies from #23572 to none
Branch pushed to git repo; I updated commit sha1. This was a forced push. New commits:
d3efc0b | Minor fixes for CDF powering |
This just fixes the doctest from french_book
. If #23572 gets merged, I'll rebase it then.
Reviewer: Travis Scrimshaw
Feel free to flip this to positive review if you think #23572 will take time. Although from what Volker wrote on #23572, it seems like that is easy to fix. Anyways, up to you on the order.
Replying to @tscrim:
Feel free to flip this to positive review if you think #23572 will take time.
I don't know whether it will take time. if this happens to be merged before #23572, fine. If it doesn't, also fine.
Changed branch from u/jdemeyer/proper_powering_in_cdf to d3efc0b
Changed branch from d3efc0b
to u/jdemeyer/proper_powering_in_cdf
This conflicts with #23572, per your request I'm giving that one priority
Dependencies: #23572
I merged #23572 and tests still pass on the patchbot.
I am confirming this is good.
Merge conflict...
Changed dependencies from #23572 to none
Branch pushed to git repo; I updated commit sha1. This was a forced push. New commits:
93c8f8e | Proper powering in CDF |
Sorry, one other test I just thought of. Can you add a doctest showing the behavior of something that masquerades as an element of ZZ
, such as 4/2
? Once added and a green patchbot, you can set a positive review.
Branch pushed to git repo; I updated commit sha1. New commits:
2803be7 | Add one more example |
Replying to @tscrim:
Sorry, one other test I just thought of. Can you add a doctest showing the behavior of something that masquerades as an element of
ZZ
, such as4/2
?
I did that. As expected, this is showing a worse result than using ZZ(2)
as exponent since powering by QQ
uses the coercion model to actually do a powering by CDF
.
Thank you.
Changed branch from u/jdemeyer/proper_powering_in_cdf to 2803be7
Fix the Cython warning
Component: basic arithmetic
Author: Jeroen Demeyer
Branch/Commit:
2803be7
Reviewer: Travis Scrimshaw
Issue created by migration from https://trac.sagemath.org/ticket/27222