Skip to content

so long and thanks for all the fish numpy x.x #415

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.

Sign up for GitHub

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

Merged
ocefpaf merged 1 commit into from
Jul 27, 2017
Merged

so long and thanks for all the fish numpy x.x #415

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
25 changes: 19 additions & 6 deletions src/meta.rst
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -73,10 +73,23 @@ A full description of selectors is

Building Against NumPy
----------------------
If you have a package which links against numpy you need to build and run against
the same version of numpy. Putting ``numpy x.x`` in the build and run requirements
ensures that a separate package will be built for each version of numpy that
conda-forge builds against.
If you have a package which links\* against numpy you need to build against the oldest possible version of numpy that is forwards compatible.
That can be achieved by pinning the build requirements and letting "free" the run requirements.
At the moment these are the oldest available numpy versions in conda-forge that you can use:
Copy link
Copy Markdown
Member

@isuruf isuruf Jul 21, 2017

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

These are not the oldest available versions in conda-forge right? They are pulled in from defaults, right?

Copy link
Copy Markdown
Member

@jakirkham jakirkham Jul 21, 2017

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

FWICT it looks like they have been added as branches to the feedstock. So they are coming from conda-forge.

ref: https://github.com/conda-forge/numpy-feedstock/tree/numpy17
ref: https://github.com/conda-forge/numpy-feedstock/tree/numpy19

Copy link
Copy Markdown
Member

@isuruf isuruf Jul 21, 2017

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Looks good to me.

Copy link
Copy Markdown
Member Author

@ocefpaf ocefpaf Jul 21, 2017

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

These are not the oldest available versions in conda-forge right? They are pulled in from defaults, right?

We added some LTS branches just for this.

Thanks for the review @isuruf!


.. code-block:: yaml

build:
- numpy 1.7.* # [py27]
- numpy 1.9.* # [py35]
- numpy 1.11.* # [py36]
run:
- numpy >=1.7 # [py27]
- numpy >=1.9 # [py35]
- numpy >=1.11 # [py36]

\* In order to know if your package links against numpy check for things like ``numpy.get_include()`` in your ``setup.py``,
or if the package uses ``cimport``.


Build Number
Expand All @@ -103,8 +116,8 @@ If you need additional source/data files for the build, download them using curl
and verify the checksum using openssl. Add curl and openssl to the build requirements and then you
can use curl to download and openssl to verify.

Example recipe is
Example recipe is
`here <https://github.com/conda-forge/pari-feedstock/blob/187bb3bdd0a5e35b2ecaa73ed2ceddc4ca0c2f5a/recipe/build.sh#L27-L35>`_.

Upstream issue for allowing multiple source is
Upstream issue for allowing multiple source is
`here <https://github.com/conda/conda-build/issues/1466>`_.