-
Notifications
You must be signed in to change notification settings - Fork 366
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
Fix support for Matplotlib 3.0.0 #1130
Conversation
1cee398
to
6fe3c4c
Compare
This is now okay with the |
Hmm, these new images should probably use the new default style since they will only be used by mpl>=2. I can add that, but we probably want to squash to remove the extra images. Edit: This is what that might look like: 973a07a |
.travis.yml
Outdated
@@ -11,10 +11,10 @@ env: | |||
# New FreeType causes changes in text output. | |||
- NAME="Latest everything." | |||
PYTHON_VERSION=3.6 | |||
PACKAGES="numpy matplotlib freetype<2.8 scipy" | |||
PACKAGES="numpy matplotlib scipy" |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
How come no freetype?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
The Matplotlib from the testing label uses a fixed copy of FreeType bundled into Matplotlib, so it will never change (at least until Matplotlib rebuilds all its images with a new version, which is unlikely unless some new comparison method is used.)
https://github.com/matplotlib/matplotlib/pull/1314 | ||
This function contains patches for Cartopy-specific behaviour, such as | ||
using the transform for limit checks, applying longitude wrapping, etc. | ||
See PATCH comments below. |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Nice detail. It is appreciated.
@@ -46,6 +46,7 @@ install: | |||
# Customise the testing environment | |||
# --------------------------------- | |||
- conda config --add channels conda-forge | |||
- conda config --add channels conda-forge/label/testing |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Can you send me a link to this please @QuLogic? I'm curious.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I see. That all makes a bit more sense now. Well This looks good to me, I guess we are just waiting for mpl 3.0.1.
Thanks for this @QuLogic, ping me when this can be merged and I'll press the button.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Pending addressing the .travis.yml
conflicts.
@QuLogic So this is waiting on me to get the matplotlib 3.0.1 conda-forge testing packages up? |
Yep; let me know when it's done and I'll push a rebase and some fixes. |
@QuLogic Packages should be up. |
This bundles a fixed version of FreeType, so we won't need to worry about those changes in test images.
_pcolormesh_patched is no longer used as a workaround for anything. Closes SciTools#992.
This is due to an indexing error in previous versions, cf. matplotlib/matplotlib#11455.
These changes are intended, with better selection of contour levels, cf. matplotlib/matplotlib#11412.
@corinnebosley @dopplershift OK, all set; we're just failing on #1171 now. |
I'll merge over CI and we can fix #1171 elsewhere. |
This fixes one straightforward issue wrt Matplotlib 3.0.0 compatibility. Two issues still exist in Matplotlib (cf matplotlib/matplotlib#12288), so I don't think we will be able to support 3.0.0, but 3.0.1 should be okay.
Once those issues are fixed, there are only 2 or 3 image failures that I need to go through to see if they're okay.