Update to version 1.0.3

[hmaarrfk]

I also modernized the script

[conda-forge-linter]

Hi! This is the friendly automated conda-forge-linting service.

I just wanted to let you know that I linted all conda-recipes in your PR (recipe) and found it was in an excellent condition.

[ocefpaf]

@conda-forge-admin, please rerender

[hmaarrfk]

@conda-forge-admin, please rerender

[hmaarrfk]

@jwiggins just a friendly ping about updating this package on conda-forge. If you would like, I could add myself as a maintainer to take some load off you.

[ocefpaf]

You can remove this pin and let the re-rendering take care of that.

[hmaarrfk]

freetype?

[ocefpaf]

Yep. On host and run.

[jwiggins]

@hmaarrfk Yes, I would appreciate the help.

(I'm now watching this repo. Sorry for not responding earlier)

[hmaarrfk]

@jwiggins cool, FYI, the windows build error is likely due to conda-forge/scikit-image-feedstock#30 (comment)

[jwiggins]

So we're basically waiting for fixes to happen upstream before this can be merged?

[ocefpaf]

So we're basically waiting for fixes to happen upstream before this can be merged?

If you want to "cash" in the OS X and Linux packages you can go ahead and merge this. Just open an issue to tracker the Windows problem there.

[jwiggins]

I'm not in any particular hurry

[hmaarrfk]

neither am I really, I got good at rebasing and rerending in case conda-smithy changes.

[hmaarrfk]

@ocefpaf any clue why this error is showing up https://ci.appveyor.com/project/conda-forge/celiagg-feedstock/build/1.0.19/job/j43k1cruydyeccxy#L632

[ocefpaf]

@ocefpaf any clue why this error is showing up ci.appveyor.com/project/conda-forge/celiagg-feedstock/build/1.0.19/job/j43k1cruydyeccxy#L632

No idea. Sorry... Windows is not my thing :-/

[hmaarrfk]

@jwiggins I guess Windows builds are also new for celiagg, but any idea why the following happens:

@isuruf @ 00:27 on gitter said:

It's a weird issue.
pip + numpy 1.12 works
pip + numpy 1.11 doesn't
python setup.py + numpy 1.12 works
python setup.py + numpy 1.11 works

[jwiggins]

Looks like it passed! (🎊)

Does 3.5 support need to be dropped? What do you mean by "pinning 2018.10.03" in your commit message?

[jwiggins]

And to answer your question, no, I don't know why the recipe fails for pip + numpy 1.11. I only remember having trouble with Python 3.4 back when I last worked on this recipe.

[hmaarrfk]

@jwiggins I think we can revert to an older version of conda-smithy and conda-forge-pinning and rerender the recipe.

[hmaarrfk]

@jwiggins this should build for python 3.5. We can rerender again after that.

[hmaarrfk]

@jwiggins friendly ping. I would merge this before something crazy happens with Travis making 3.5 compilation impossible again or the bot for 3.7 migration comes around.

[jwiggins]

Thanks for sticking with this!

[jwiggins]

The build was clean(-ish. Had to restart linux3.5): https://anaconda.org/conda-forge/celiagg/files?version=1.0.3