update to 8.6.5; add tix

[msarahan]

Should be merged into 8.6 branch, not master (unless that's what you want...)

Built Python 3.5 with this as an external library.

[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.

[msarahan]

Don't merge this yet. The dependence on Python is problematic. It will build OK, but once we build python with this, then we have a circular dependency, which has the end effect of creating empty packages here (because all files are already in place at build time, so nothing changes.)

Working on it...

[jakirkham]

Should be merged into 8.6 branch, not master (unless that's what you want...)

We decided master is always latest. The idea is the default workflow is to keep upgrading a package and only create version branches for older versions that need more maintenance.

[jakirkham]

Don't merge this yet. The dependence on Python is problematic. It will build OK, but once we build python with this, then we have a circular dependency, which has the end effect of creating empty packages here (because all files are already in place at build time, so nothing changes.)

How did we avoid this problem before? Could we try using the vc package or are there other reasons this needs python now?

[msarahan]

Python did not depend on tk before. Instead, Python downloaded and built
it's own tk. On Windows, we just copied tk from a Python.org build.

I made things better but worse. I am experimenting with vc package, but
still having trouble with activation of features. Really need to get that
build customization developed.

On Wed, Jun 8, 2016, 22:00 jakirkham notifications@github.com wrote:

Don't merge this yet. The dependence on Python is problematic. It will
build OK, but once we build python with this, then we have a circular
dependency, which has the end effect of creating empty packages here
(because all files are already in place at build time, so nothing changes.)

How did we avoid this problem before? Could we try using the vc package
or are there other reasons this needs python now?

—
You are receiving this because you authored the thread.
Reply to this email directly, view it on GitHub
#1 (comment),
or mute the thread
https://github.com/notifications/unsubscribe/AACV-Wmywl86GwcAclegE2wGoGjO66C6ks5qJ4G3gaJpZM4ItGlk
.

[jakirkham]

Python did not depend on tk before. Instead, Python downloaded and built it's own tk. On Windows, we just copied tk from a Python.org build.

That's right. I forgot about that mess.

I am experimenting with vc package, but still having trouble with activation of features.

What's the trouble?

Really need to get that build customization developed.

How does that relate here?

[msarahan]

The trouble is that the selectors don't seem to work correctly in the test section, and there are unresolvable conflicts.

The build customization work is relevant because it will make these feature messes easier to sort out. We'll figure them out once and then just template them.

[pelson]

We have a similar circular issue with the openssl feedstock. Once we have a new conda-smithy, I'm going to prioritise the VC stuff to try to remove the use of Python as a visual studio proxy. 😄

[jakirkham]

Should this be rebased and cleaned up for the current Tk version or should it be closed with/without taking parts into a new PR?

[msarahan]

tix is not essential, and I think that the 't' suffix stuff is already present elsewhere. I'm happy to close.