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
p_group_cohomology missing dependencies #27173
Comments
Commit: |
New commits:
|
Changed commit from |
Changed branch from u/jdemeyer/p_group_cohomology_depends_on_cython to none |
Replying to @jdemeyer:
That's actually good news. When you wrote "API change broke the spkg", I thought I need to change code again (by the way, I am not sure: What is the current official spkg version in Sage? v3.1?).
The current dependencies are
Really the package does depend on sagelib. Namely, it cimports various Sage types (including Element, Parent). In an ideal world, it would be possible to have a dependency list
|
comment:5
I hate the trac interface. Writing my comment destroyed your branch. Can you provide the branch again? |
New commits:
|
Commit: |
comment:7
Replying to @simon-king-jena:
No, what really happened is an incompatible change in |
comment:8
Replying to @simon-king-jena:
Actually, why don't we just do that? |
comment:9
Replying to @jdemeyer:
Well, when I wrote my comment, I thought that it isn't supported yet. So, are you saying that a dependency on specific parts of the Sage library is possible? Or would that mean to extend Sage's build system? |
Branch pushed to git repo; I updated commit sha1. This was a forced push. New commits:
|
comment:12
Note that this won't pick up transitive dependencies (the |
comment:13
Replying to @simon-king-jena:
It's a |
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
comment:15
Replying to @jdemeyer:
Nice! I thought the Also good that you add pip - indeed spkg-install uses pip. But how to test it? Pull the branch, build sage, touch element.pxd, do |
comment:16
I hate the trac interface. When I write a comment, I certainly don't want that trac changes the ticket description back to what it was when I started to write my comment! |
This comment has been minimized.
This comment has been minimized.
comment:18
Replying to @simon-king-jena:
Yes, something like that. |
comment:19
I think it would be possible to be a little more thorough. Here is the list of all cimports in the spkg that get stuff from outside the spkg (hence, that's dependencies):
So, the complete lists of dependencies is
So, I suggest to include all this in the list of dependencies. Which gives rise to the questions:
|
This comment has been minimized.
This comment has been minimized.
comment:20
Replying to @simon-king-jena:
Only the first line of the |
Changed keywords from none to dependencies |
Changed branch from u/jdemeyer/p_group_cohomology_depends_on_cython to u/SimonKing/p_group_cohomology_depends_on_cython |
comment:22
I have added all dependencies in sagelib, i.e., all pxd-files from which p_group_cohomology cimports. So, what's missing is libc and cysignals. How to name them in the dependencies? New commits:
|
comment:23
Replying to @simon-king-jena:
No, that's not the complete list of dependencies. Like I said, you should also consider implied dependencies: for example, For cysignals, just add a dependency on the |
comment:24
Replying to @jdemeyer:
I understood that. But I think to the very least one should name all pxd files from which the package cimports. Everything else would be too short.
Thanks. And for libc? |
Branch pushed to git repo; I updated commit sha1. New commits:
|
Changed author from Jeroen Demeyer to Jeroen Demeyer, Simon King |
comment:26
I have added |
comment:27
Replying to @simon-king-jena:
That's just the system C library on which virtually everything depends. It's assumed to remain ABI-compatible, so recompiling isn't needed. |
Reviewer: Simon King, Jeroen Demeyer |
Changed branch from u/SimonKing/p_group_cohomology_depends_on_cython to |
The API change of
coercion_model
(#27021) brokep_group_cohomology
.Rebuilding
p_group_cohomology
is sufficient (an analogous problem occurred withgiacpy_sage
on #27021).As experimental fix, we make
p_group_cohomology
depend onelement.pxd
and cython.CC: @simon-king-jena
Component: packages: optional
Keywords: dependencies
Author: Jeroen Demeyer, Simon King
Branch/Commit:
f39c9f1
Reviewer: Simon King, Jeroen Demeyer
Issue created by migration from https://trac.sagemath.org/ticket/27173
The text was updated successfully, but these errors were encountered: