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
Accept giac-1.7.x from the system #31594
Comments
Branch: u/mjo/ticket/31594 |
Commit: |
New commits:
|
comment:2
Sage development has entered the release candidate phase for 9.3. Setting a new milestone for this ticket based on a cursory review. |
comment:3
I am prepared to accept this, although I probably would want #31563 to be done first. |
comment:4
One of the main benefits of using system packages is that we don't all have to wait for someone to duplicate pointless work in the SPKG =) |
comment:5
Replying to @orlitzky:
I am impatiently waiting for removal of gcc and gfortran packages (along with the rest of Sage toolchain), to start with :-) |
comment:6
let me test on Gentoo. It seems I'd also need pari 2.13 from the system for this to work, though. |
Reviewer: Dima Pasechnik, Travis Scrimshaw |
comment:8
Beware that the latest 1.7.0.3 version causes some breakage. Some tests fail because of changes in expression expansion, but there are some other weird failures such as
|
comment:9
Seems that evalb is just broken in 1.7.0.3
Could someone with an account please report it upstream? |
comment:10
evalb bug fixed https://dev.geogebra.org/trac/changeset/69847/ |
comment:11
1.7.0-3 doesn't like the default |
comment:12
I'll change this to accept only 1.7.0.1 for now, since the simplification changes are likely to stick around. |
comment:13
Oh, they all return the same version number. |
comment:14
Why not just modify the tests so they pass with both versions? IMHO a slight change in output format shouldn't block using a system package. And some of these tests already have a very generic expected output:
|
comment:15
Replying to @antonio-rojas:
I came to this same conclusion a few seconds ago =) First I have to update our Gentoo package to 1.7.0.5 so that I can see the new output and fix the c++17 issue, but then I'll update this branch with tests that work with all versions. |
comment:16
The bump is trivial. c++17 is probably a can of worms. |
comment:17
By "fix" I mean |
comment:18
Yes, I did that in the overlay. A bit ham-fisted but it basically works. |
comment:19
Replying to @antonio-rojas:
This is not such a great example anyway:
I think I'll replace it with an integral that I can do in my head. |
comment:20
Replying to @orlitzky:
|
comment:21
If I understand correctly, you are evaluating an antiderivative at x=1 and you get an approximation. I do not understand how this test can be meaningful, since an antiderivative is defined up to a constant. One could argue it should stay the same for a given CAS, but that's not true, I made some improvements in symbolic integration recently in giac, and that means sometimes that the algorithm changes and the antiderivative expression too. |
comment:22
I'm not blaming giac for anything here. The existing test was a bad choice:
My use of |
comment:23
A more meaningful test would be F=integrate(f) then simplify(diff(F)-f). I have added some examples in the check subdirectory of giac, from the independent integrals testsuite from https://www.12000.org/my_notes/CAS_integration_tests/reports/rubi_4_16_1_graded/inch1.htm#x2-30001.2 |
Branch pushed to git repo; I updated commit sha1. This was a forced push. New commits:
|
comment:25
Replying to @sagetrac-parisse:
The call to |
comment:26
lgtm |
Changed branch from u/mjo/ticket/31594 to |
This works fine out-of-the-box, we just need to tweak the version bounds in spkg-configure.m4 to accept it.
CC: @mkoeppe @dimpase @isuruf @antonio-rojas @kiwifb
Component: build: configure
Author: Michael Orlitzky
Branch/Commit:
55e3613
Reviewer: Dima Pasechnik, Travis Scrimshaw
Issue created by migration from https://trac.sagemath.org/ticket/31594
The text was updated successfully, but these errors were encountered: