configure --enable-system-site-packages via spkg-configure.m4 for cvxopt and other python packages #29665
Comments
|
spkg-configure.m4 to be copied under build/pkgs/cvxopt |
|
comment:1
Attachment: spkg-configure.m4.gz I don't see a branch |
|
comment:2
See comments in #29023 |
|
Commit: |
|
Changed author from gh-thierry-FreeBSD to none |
|
comment:5
There is per se nothing wrong with having spkg-configure.m4 for Python packages. |
|
comment:6
Not per se, only per the big picture. |
|
comment:7
Regardless of what we eventually wind up with... does this work? Does it hurt anything? The solution in #29023 could be years away. If we can avoid building e.g. numpy, scipy, matplotlib, and cvxopt with relatively trivial spkg-configures until then, why not do it? |
|
comment:8
No, it does not work. |
|
comment:9
Replying to @mkoeppe:
What goes wrong? I just did a quick test (without this branch):
Behavior is unchanged as far as I can tell. |
|
comment:10
Sound like you forgot to test with Sage runs in a venv that does not provide access to system site packages. |
|
comment:11
Replying to @mkoeppe:
Lucky guess! I forgot we made it possible to disable cvxopt even though it's type=standard. On the surface, setting |
|
comment:13
Matthias, does something like this fit with your master plan? Last 10 new commits:
|
|
Changed branch from u/gh-thierry-FreeBSD/spkg_configure_m4_for_cvxopt to u/mjo/ticket/29665 |
|
comment:14
I wouldn't object to doing something like this, as it is clearly marked experimental. However, be prepared for a long-term burden of maintenance regarding the declared version ranges. Lots of things can go wrong because the individual packages often have complicated version constraints. Also the version constraints shouldn't be hardcoded in the spkg-configure.m4 file; instead, they should really be taken from Also, note |
|
Branch pushed to git repo; I updated commit sha1. This was a forced push. Last 10 new commits:
|
Nothing, and I've rebased it at #36141 There's one test failure due to a deprecation warning in ipython somewhere, but I don't want to spend too much time chasing down minor test failures again. This is opt-in and I think it's a better use of time to get it merged and then deal with the issues rather than try to make it perfect before merging, given how many commits there are and how easily everything goes out of date. |
…e-packages
Rebased branch of sagemath#29665
Platform-specific testing:
- for example TARGETS_PRE="build-local" tox -e docker-ubuntu-focal-
standard-sitepackages -- build ptest
Fixes sagemath#29665
URL: sagemath#36141
Reported by: Michael Orlitzky
Reviewer(s): Dima Pasechnik, Matthias Köppe, Michael Orlitzky
…e-packages
Rebased branch of sagemath#29665
Platform-specific testing:
- for example TARGETS_PRE="build-local" tox -e docker-ubuntu-focal-
standard-sitepackages -- build ptest
Fixes sagemath#29665
URL: sagemath#36141
Reported by: Michael Orlitzky
Reviewer(s): Dima Pasechnik, Matthias Köppe, Michael Orlitzky
…e-packages
Rebased branch of sagemath#29665
Platform-specific testing:
- for example TARGETS_PRE="build-local" tox -e docker-ubuntu-focal-
standard-sitepackages -- build ptest
Fixes sagemath#29665
URL: sagemath#36141
Reported by: Michael Orlitzky
Reviewer(s): Dima Pasechnik, Matthias Köppe, Michael Orlitzky
sagemath#29665 Currently broken (since sagemath#29665): ``` $ build/bin/sage-get-system-packages install-requires zipp $ build/bin/sage-get-system-packages install-requires-toml zipp ``` This causes `pkgs/sagemath-standard/pyproject.toml` to be broken; as noted in sagemath#36964 (comment) The correct output: ``` $ build/bin/sage-get-system-packages install-requires zipp zipp >=0.5.2 $ build/bin/sage-get-system-packages install-requires-toml zipp 'zipp >=0.5.2', ``` <!-- ^^^^^ Please provide a concise, informative and self-explanatory title. Don't put issue numbers in there, do this in the PR body below. For example, instead of "Fixes sagemath#1234" use "Introduce new method to calculate 1+1" --> <!-- Describe your changes here in detail --> <!-- Why is this change required? What problem does it solve? --> <!-- If this PR resolves an open issue, please link to it here. For example "Fixes sagemath#12345". --> <!-- If your change requires a documentation PR, please link it appropriately. --> ### 📝 Checklist <!-- Put an `x` in all the boxes that apply. --> <!-- If your change requires a documentation PR, please link it appropriately --> <!-- If you're unsure about any of these, don't hesitate to ask. We're here to help! --> <!-- Feel free to remove irrelevant items. --> - [x] The title is concise, informative, and self-explanatory. - [x] The description explains in detail what this PR is about. - [x] I have linked a relevant issue or discussion. - [ ] I have created tests covering the changes. - [ ] I have updated the documentation accordingly. ### ⌛ Dependencies <!-- List all open PRs that this PR logically depends on - sagemath#12345: short description why this is a dependency - sagemath#34567: ... --> <!-- If you're unsure about any of these, don't hesitate to ask. We're here to help! --> URL: sagemath#36979 Reported by: Matthias Köppe Reviewer(s): Michael Orlitzky
sagemath#29665 Currently broken (since sagemath#29665): ``` $ build/bin/sage-get-system-packages install-requires zipp $ build/bin/sage-get-system-packages install-requires-toml zipp ``` This causes `pkgs/sagemath-standard/pyproject.toml` to be broken; as noted in sagemath#36964 (comment) The correct output: ``` $ build/bin/sage-get-system-packages install-requires zipp zipp >=0.5.2 $ build/bin/sage-get-system-packages install-requires-toml zipp 'zipp >=0.5.2', ``` <!-- ^^^^^ Please provide a concise, informative and self-explanatory title. Don't put issue numbers in there, do this in the PR body below. For example, instead of "Fixes sagemath#1234" use "Introduce new method to calculate 1+1" --> <!-- Describe your changes here in detail --> <!-- Why is this change required? What problem does it solve? --> <!-- If this PR resolves an open issue, please link to it here. For example "Fixes sagemath#12345". --> <!-- If your change requires a documentation PR, please link it appropriately. --> ### 📝 Checklist <!-- Put an `x` in all the boxes that apply. --> <!-- If your change requires a documentation PR, please link it appropriately --> <!-- If you're unsure about any of these, don't hesitate to ask. We're here to help! --> <!-- Feel free to remove irrelevant items. --> - [x] The title is concise, informative, and self-explanatory. - [x] The description explains in detail what this PR is about. - [x] I have linked a relevant issue or discussion. - [ ] I have created tests covering the changes. - [ ] I have updated the documentation accordingly. ### ⌛ Dependencies <!-- List all open PRs that this PR logically depends on - sagemath#12345: short description why this is a dependency - sagemath#34567: ... --> <!-- If you're unsure about any of these, don't hesitate to ask. We're here to help! --> URL: sagemath#36979 Reported by: Matthias Köppe Reviewer(s): Michael Orlitzky
sagemath#29665 Currently broken (since sagemath#29665): ``` $ build/bin/sage-get-system-packages install-requires zipp $ build/bin/sage-get-system-packages install-requires-toml zipp ``` This causes `pkgs/sagemath-standard/pyproject.toml` to be broken; as noted in sagemath#36964 (comment) The correct output: ``` $ build/bin/sage-get-system-packages install-requires zipp zipp >=0.5.2 $ build/bin/sage-get-system-packages install-requires-toml zipp 'zipp >=0.5.2', ``` <!-- ^^^^^ Please provide a concise, informative and self-explanatory title. Don't put issue numbers in there, do this in the PR body below. For example, instead of "Fixes sagemath#1234" use "Introduce new method to calculate 1+1" --> <!-- Describe your changes here in detail --> <!-- Why is this change required? What problem does it solve? --> <!-- If this PR resolves an open issue, please link to it here. For example "Fixes sagemath#12345". --> <!-- If your change requires a documentation PR, please link it appropriately. --> ### 📝 Checklist <!-- Put an `x` in all the boxes that apply. --> <!-- If your change requires a documentation PR, please link it appropriately --> <!-- If you're unsure about any of these, don't hesitate to ask. We're here to help! --> <!-- Feel free to remove irrelevant items. --> - [x] The title is concise, informative, and self-explanatory. - [x] The description explains in detail what this PR is about. - [x] I have linked a relevant issue or discussion. - [ ] I have created tests covering the changes. - [ ] I have updated the documentation accordingly. ### ⌛ Dependencies <!-- List all open PRs that this PR logically depends on - sagemath#12345: short description why this is a dependency - sagemath#34567: ... --> <!-- If you're unsure about any of these, don't hesitate to ask. We're here to help! --> URL: sagemath#36979 Reported by: Matthias Köppe Reviewer(s): Michael Orlitzky
This ticket adds an spkg-configure.m4 for cvxopt and various other python packages, to allow them to be used from the system (part of #27330).
These python packages are special because we currently use a python "venv" that is isolated from the system. Thus this ticket proceeds in several steps:
--enable-system-site-packagesflag to the top-level./configurescript. When present, this will tellsage-venvthat the venv it creates should not be isolated from the system.install-requires.txtare met.spkg-configure.m4files themselves, which should be relatively simple at that point.Platform-specific testing:
TARGETS_PRE="build-local" EXTRA_CONFIGURE_ARGS="--enable-system-site-packages" tox -e docker-ubuntu-focal-standard -- build ptestDepends on #32703
Depends on #33067
Component: build: configure
Keywords: cvxopt, system packages, Python library
Author: Michael Orlitzky
Branch/Commit: u/mjo/ticket/29665 @
baf6301Issue created by migration from https://trac.sagemath.org/ticket/29665
The text was updated successfully, but these errors were encountered: