New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
upgrade Singular to 4.1.1p2 #24735
Comments
comment:1
One immediate issue is Singular/Singular#855 |
comment:2
Replying to @dimpase:
Fixed upstream. |
Upstream: Fixed upstream, but not in a stable release. |
comment:4
A patch that makes Sage build with the API changes in 4.1.1 is available at [1] However, the removal of singclap_pmod and singclap_pdivide for polynomials with integer coefficients [2] causes some regressions:
|
Changed keywords from none to days94 |
comment:7
Singular no longer supports singclap_pmod and singclap_pdivide for polynomials with integer coefficients. This change
makes Sage not call Singular's singclap_pdivide for integer coefficients in quo_rem. With this, most test pass again (modulo some output format changes and generators ordering). There is another singclap_pdivide call in |
Commit: |
comment:10
Pushing WIP patch. Still needs a decision on what to do for Unfortunately starting from 8.3.beta7 there are many more test failures caused by the upgrade due to #23909 New commits:
|
This comment has been minimized.
This comment has been minimized.
comment:13
Okay in addition to the revert of #25313 it was also necessary to convert integral polynomials to rationals in I'm still running @antonio-rojas I don't want to "steal" your work, so feel free to pick those commits in your branch and rebase them if nobody objects. |
comment:14
Replying to @timokau:
What about gcd? AFAICS this is still calling singular's singclap_gcd for ZZ coefficients. Feel free to pick up my branch where I left it, I'm glad to see this move along. |
comment:15
Replying to @antonio-rojas:
The tests succeed, including the one over
|
comment:16
This gives a wrong answer
That is technically correct on Q[x,y] (although the coefficient is strange), but not on ZZ[x,y]. For the lcm, we should probably factor out the contents, take the lcm of the primitive parts over QQ and multiply by the lcm (on Z) of the contents. |
comment:18
There are still many test failures which are unrelated to the sinclap_pdivide issue. They seem caused by singular now returning an warning when computing groebner bases over CC, so Sage falls back to the toy implementation:
now returns
With singular 4.1.0 it gives
The problem is that the Singular warning
from singular.py so Sage (incorrectly) throws a SingularError. The test for errors in singular.py should be made more reliable. |
comment:20
Why? My understanding of the theory here is very limited, but shouldn't the lcm be
Yeah I didn't have time to finish the tests unfortunately. Thanks for continuing to work on this. What's the status? Are there still remaining failures? |
comment:21
Replying to @timokau:
The lcm is well defined only up to multiplication by a unit. Both
The ntl ones are going to need some serious debugging. |
comment:22
Replying to @antonio-rojas:
Ah I didn't know / forgot that. That makes sense.
But how does this solve the problem? E.g. let's determine the lcm of 2xy and 4xy using that technique. The contents would be 2 and 4, respectively. So we'd take Did I miss something? And do you know why Singular removed support for integer coefficients? I feel like we shouldn't have to hack around this. |
comment:23
Replying to @timokau:
Yes, of course you're right, this is all under the assumption that, when computing the gcd of two polynomials in QQ[] with integer coefficients and primitive, Singular will return a polynomial with the same properties. This sounds reasonable to me, but may very well not be true in all cases. An alternative would be to simply return
No idea, the commit message doesn't say much. But I wonder if it even makes sense at all to define |
comment:24
Or we can compute the gcd over ZZ and then coerce to QQ[] to perform the division with |
comment:79
I opened an issue on the Singular Trac about their versioning scheme: https://www.singular.uni-kl.de:8005/trac/ticket/836 |
Changed reviewer from Jeroen Demeyer to Jeroen Demeyer, Volker Braun |
Changed branch from u/jdemeyer/upgrade_singular_to_4_1_1 to |
Changed commit from |
This comment has been minimized.
This comment has been minimized.
comment:83
I noted a regression at #26023, though I might still unmerge this ticket... |
comment:86
I'm assuming that you reopened this because of #26023. |
Changed branch from |
Commit: |
Last 10 new commits:
|
comment:89
Thanks! |
comment:90
Replying to @timokau:
Thanks for the notice, and sorry for having taken so long to react. Unfortunately, I know very, very little about Singular. But I saw that, in |
comment:91
It may be, I don't recall the details of how it is used in that function. I don't think the |
Changed branch from u/jdemeyer/a9e5aed9215175f8cf49327c790ae153648a2513 to |
Singular 4.1.1 is out.
The changes for it can be found at:
https://www.singular.uni-kl.de/index.php/news/release-of-singular-4-1-1.html
The tarball:
http://www.mathematik.uni-kl.de/ftp/pub/Math/Singular/SOURCES/4-1-1/singular-4.1.1p2.tar.gz
Follow-up ticket for the upgrade to Singular 4.1.1p3: #25993.
Upstream: Fixed upstream, but not in a stable release.
CC: @kiwifb @antonio-rojas @timokau @mezzarobba
Component: packages: standard
Keywords: days94
Author: Antonio Rojas, Timo Kaufmann
Branch/Commit:
45ff371
Reviewer: Jeroen Demeyer, Volker Braun
Issue created by migration from https://trac.sagemath.org/ticket/24735
The text was updated successfully, but these errors were encountered: