Skip to content
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

Version 0.16 release candidate #232

Merged
merged 4 commits into from
Oct 12, 2020
Merged

Conversation

pp-mo
Copy link
Member

@pp-mo pp-mo commented Oct 7, 2020

Now basing this against the release branch "0.16.x", which is branched from the feature branch "f__test-iris-master"

This PR now modifies that to specifically test against the Iris 3.0.0rc0 release-candidate from conda-forge/iris_rc
The changes here are specific to the 0.16 release and will not go back onto master.

Copy link
Contributor

@trexfeathers trexfeathers left a comment

Choose a reason for hiding this comment

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

Hi @pp-mo, I'd favour a slightly different implementation - see my single comment.

Also, shouldn't this now be pointing at an appropriately named branch in the same way that Iris 3.0.0rc0 comes from the v3.0.x branch?

.travis.yml Outdated Show resolved Hide resolved
@trexfeathers
Copy link
Contributor

... further to my review, the PIP requirement should also be changed to scitools-iris=3.0.0rc0 (although my Conda suggestion should mean this isn't necessary to fix these Travis failures anyway).

@pp-mo pp-mo changed the base branch from f__test-iris-master to v0.16.x October 12, 2020 10:42
@coveralls
Copy link

coveralls commented Oct 12, 2020

Coverage Status

Coverage remained the same at 92.482% when pulling bbac02d on pp-mo:rc_0v16 into 85a5f4b on SciTools:v0.16.x.

@pp-mo
Copy link
Member Author

pp-mo commented Oct 12, 2020

Rebased to fix conflicts.
I believe that the new versions of all these commits still make good sense.
Please re-check with care @trexfeathers

@pp-mo
Copy link
Member Author

pp-mo commented Oct 12, 2020

Regarding the general working process for an iris-dependent "iris option package" like this one,
I.E. the process to develop an RC of this against a pre-release of Iris.
(as we're still writing the book here)

I am not that happy that I needed to name the special-case "3.0.0rc0" release version in requirements/core.txt, but it seems unavoidable.
It would seem that, as an RC, this should still be happy to work (i.e. install) with the eventual 3.0 release as a dependency,
because "3.0.0 >= 3.0.0rc0".

@trexfeathers
Copy link
Contributor

I am not that happy that I needed to name the special-case "3.0.0rc0" release version in requirements/core.txt, but it seems unavoidable.
It would seem that, as an RC, this should still be happy to work (i.e. install) with the eventual 3.0 release as a dependency,
because "3.0.0 >= 3.0.0rc0".

Yes I've been similarly uncomfortable and similarly resigned!

@trexfeathers
Copy link
Contributor

Thanks so much @pp-mo , this exactly what I imagined and is hopefully the clearest and safest path forward 🥂

Next challenge: what to do with conda-forge labels...

@trexfeathers trexfeathers merged commit 9c6bea6 into SciTools:v0.16.x Oct 12, 2020
jamesp added a commit to jamesp/iris-grib that referenced this pull request Feb 23, 2021
* upstream/v0.16.x:
  Docstest 0v16 (SciTools#244)
  Remove eccodes bug workaround added in SciTools#208. (SciTools#224)
  Update requirements to pick up Iris 3. (SciTools#243)
  Add 'main' conda-forge channel, needed for docs builds. (SciTools#240)
  Fix RC date in release notes (about to cut). (SciTools#235)
  Version 0.16 release candidate (SciTools#232)
  Cosmetic change : rename the travis iris-test-version options (SciTools#234)
  Travis test with  both Iris latest-release and latest-master. (SciTools#231)
lbdreyer pushed a commit that referenced this pull request Feb 24, 2021
* Require Iris >=3 (just for the gdt90 changes).

* Whatsnew entry for requiring Iris 3.

* Set version string for release candidate.

* Test against Iris 3.0.0rc0 from conda-forge/rc_iris.
bjlittle pushed a commit that referenced this pull request Mar 2, 2021
* Travis test with  both Iris latest-release and latest-master. (#231)

* Travis test with  both Iris latest-release and latest-master.

* Modify test CMLs for latest Iris (Iris3.0 changes).

* Grib1 load fixes.

* Fix loading since units=None default for Iris3 coords

* Modify test to work with latest Iris (Iris3.0 changes).

* Test against latest Iris only.

* Review changes.

* Cosmetic change : rename the travis iris-test-version options (#234)

* Rename Iris test-version options, and enable all to check action.

* Review changes.

* Version 0.16 release candidate (#232)

* Require Iris >=3 (just for the gdt90 changes).

* Whatsnew entry for requiring Iris 3.

* Set version string for release candidate.

* Test against Iris 3.0.0rc0 from conda-forge/rc_iris.

* Fix RC date in release notes (about to cut). (#235)

* Add 'main' conda-forge channel, needed for docs builds. (#240)

* Add 'main' conda-forge channel, needed for docs builds.

* Use fixed spherical-earth-radius in GRIB1, ignoring change in gribapi default.

* Codestyle fix.

* Update requirements to pick up Iris 3. (#243)

* Remove eccodes bug workaround added in #208. (#224)

* Docstest 0v16 (#244)

* Document PR#240 in release notes.

* Fix version string and release-notes date.

* Added getting started .cirrus.yml

* Update cirrus to use miniconda image

* Update .cirrus.yml

* Update .cirrus.yml

Copied across cirrus file from iris-ugrid

* Added nox testing

Borrowing from iris and iris-ugrid, added nox testing and a test runner.

Tests currently fail on my local machine.

* Path fixes in cirrus.yml

* eccodes test added to noxfile

* Added config and coverage

* Trying to set SITE_CFG

* syntax error

* taking the IRIS_DIR from Travis CI config

* Add allow_failures to the linux task for now

* Dodgy r key...

* Update .cirrus.yml

* Allow failures in linting

* Force return error 0  for now

* Added eccodes test

* moved matrix

* Yaml syntax error & * wrong way around

* Typo

* Update iris version dependency

* Date fix

This resolves the same bash issue as SciTools/iris#4019

* Configure Iris in nox

* Updated CI config

* correct site-packages path

* Fix to iris test data path

* force cache update

* refactored writing iris config in noxfile

* Fix pep8 and license test fails

* Support for testing against packaged iris and building from source

* Adding yaml to list of cirrus container requirements

* invalidate cirrus cache

* really invalidate cirrus cache...

* yaml -> pyyaml

* missing iris_dir reference

* call write_iris_config

* docstrings and matrix testing

* git syntax error

* eccodes selfcheck in basic tests

* Removed py3.8 testing for now

* Fixed regression in setup.py test

* Removed unused imports

* Disabled lint checking

* Removed nox from test dependencies for python3.6 and python3.8

* Removed python3.8 from noxfile

* Removed nox from test requirements

Co-authored-by: Patrick Peglar <patrick.peglar@metoffice.gov.uk>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

3 participants