I just discovered that pywcs doesn't work anymore (CI tests fail) since
quite a while. Because it obsolete since a long time, I would propose to
drop it before stretch completely. I would even not create a "legacy"
package, since one just would need all references to "pywcs" with
The package has no reverse dependencies, so removing it wouldn't break
anything within Debian. However, it has still a popcon of ~144 (28
votes), these people would probably wake up at some day without pywcs.
Any hint to do better here? They *must* move at some point, and Debian 9
could give them the right push :-)
The same would in principle apply to pyfits, which is maintained by
I agree with your proposal: completely removing pywcs now.
It was moved to astropy.wcs 5 years ago and the last release of pywcs was in 2014:
In the meantime, astropy.wcs has seen several bug fixes and new features, so in my opinion dropping pywcs now is the way to go.