Yes, but we do not have any idea what people are actually using. For example the LIGO notebooks used some stuff from mlab (the power spectrum).

Well, won't a standard deprecation cycle be enough?

This stuff has been in matplotlib for avery long time. We would need a
few cycles for deprecation, and we should be providing clear replacements
in the deprecation notes.

On Mon, Jun 13, 2016 at 2:24 PM, Antony Leenotifications@github.com
wrote:

Well, won't a standard deprecation cycle be enough?

—
You are receiving this because you are subscribed to this thread.
Reply to this email directly, view it on GitHub
#6577 (comment),
or mute the thread
https://github.com/notifications/unsubscribe/AARy-OnE_tTXl7CfRMcgOvtA0WLSYpW0ks5qLaBMgaJpZM4Iz_V2
.

I certainly understand, and wouldn't have suggested it if I didn't find parts of the code that are broken. The alternative would be to improve the coverage, which is so-so right now (this is actually the kind of code for which high coverage could be achieved).

Indirectly backported to v2.x ascbec32b via#6868.