We haven't broken 3.6 compatibility so might as well release them. As I mentioned on the thread, as far as I'm concerned, when we find we want something that is 3.7+ only, we can drop 3.6 without hesitation. On Mon, 2 Nov 2020, at 7:07 AM, Mark Harfouche wrote:

Are we looking at the thread on the numpy mailing list to decide on releasing wheels for python 3.6?

On Mon, Nov 2, 2020 at 7:26 AM Juan Nunez-Iglesias wrote:

...

__ Hi everyone,

I intend to branch v0.18.x and release version 0.18.0 tomorrow or Wednesday. If you have any pull requests that you think are *critical* to get into 0.18, or that would be nice and you think are super close, please tag the appropriate PR with the 0.18 milestone.

Juan. _______________________________________________ scikit-image mailing list -- scikit-image@python.org To unsubscribe send an email to scikit-image-leave@python.org https://mail.python.org/mailman3/lists/scikit-image.python.org/ Member address: mark.harfouche@gmail.com

---

scikit-image mailing list -- scikit-image@python.org To unsubscribe send an email to scikit-image-leave@python.org https://mail.python.org/mailman3/lists/scikit-image.python.org/ Member address: jni@fastmail.com

Let me put it this way: if you make a PR dropping 3.6, I'd approve it. =) On Mon, 2 Nov 2020, at 7:33 AM, Mark Harfouche wrote:

I think being proactive about this stuff is a little better since it opens up the mind to more creativity.

Reducing CI matrix, enabling more builds for new architectures are all things that become possible when we decide to stop supporting things that are older.

Having to weigh decisions before even making the PR on the need for a huge compatibility matrix is draining.

I can almost guarantee that any PR that tries to bundle removing support for 3.6 will be held up on the count of a similar discussion.

As I mentioned in that thread, there is little point in us discussing this if NumPy won't follow their guidelines.

On Mon, Nov 2, 2020, 08:12 Juan Nunez-Iglesias wrote:

...

__ We haven't broken 3.6 compatibility so might as well release them. As I mentioned on the thread, as far as I'm concerned, when we find we want something that is 3.7+ only, we can drop 3.6 without hesitation.

On Mon, 2 Nov 2020, at 7:07 AM, Mark Harfouche wrote:

...

Are we looking at the thread on the numpy mailing list to decide on releasing wheels for python 3.6?

On Mon, Nov 2, 2020 at 7:26 AM Juan Nunez-Iglesias wrote:

...

__ Hi everyone,

I intend to branch v0.18.x and release version 0.18.0 tomorrow or Wednesday. If you have any pull requests that you think are *critical* to get into 0.18, or that would be nice and you think are super close, please tag the appropriate PR with the 0.18 milestone.

Juan. _______________________________________________ scikit-image mailing list -- scikit-image@python.org To unsubscribe send an email to scikit-image-leave@python.org https://mail.python.org/mailman3/lists/scikit-image.python.org/ Member address: mark.harfouche@gmail.com

---

scikit-image mailing list -- scikit-image@python.org To unsubscribe send an email to scikit-image-leave@python.org https://mail.python.org/mailman3/lists/scikit-image.python.org/ Member address: jni@fastmail.com

_______________________________________________ scikit-image mailing list -- scikit-image@python.org To unsubscribe send an email to scikit-image-leave@python.org https://mail.python.org/mailman3/lists/scikit-image.python.org/ Member address: mark.harfouche@gmail.com

---

scikit-image mailing list -- scikit-image@python.org To unsubscribe send an email to scikit-image-leave@python.org https://mail.python.org/mailman3/lists/scikit-image.python.org/ Member address: jni@fastmail.com