Re: python fails with nis.so missing error

[Greg Troxel <gdt%lexort.com@localhost>]

Ede Wolf <listac%nebelschwaden.de@localhost> writes:

> Not sure if there is any relation, I did rebuild world with: MKYP=no -
> but I am having troubles believing, that python really relies on nis?

You should believe!

$ /usr/pkgsrc/lang/python311 > egrep nis *
Makefile:PLIST_VARS+=   dbm nis
Makefile:PLIST.nis=     yes
PLIST:${PLIST.nis}lib/python${PY_VER_SUFFIX}/lib-dynload/nis.so
PLIST:lib/python${PY_VER_SUFFIX}/test/test_nis.py
PLIST:lib/python${PY_VER_SUFFIX}/test/test_nis.pyc
PLIST:lib/python${PY_VER_SUFFIX}/test/test_nis.pyo

Read Makefile, and note that it looks for the yp headers.  If you are
installing a non-MKYP build and leave the headers, you are in an
inconsistent state

In general, when you depart from the defaults, you bit off a greater
likelihood of trouble.

> => Generating post-install file lists
> pkg_create: can't stat
> `/import/build/pkgobj/lang/python311/work/.destdir/usr/pkg/lib/python3.11/lib-dynload/nis.so'
> => Checking file-check results for python311-3.11.6nb4
> ERROR: ************************************************************
> ERROR: The following files are in the PLIST but not in
> /import/build/pkgobj/lang/python311/work/.destdir/usr/pkg:
> ERROR:
> /import/build/pkgobj/lang/python311/work/.destdir/usr/pkg/lib/python3.11/lib-dynload/nis.so
> *** Error code 1

(BTW, your MUA appears to have wrapped this output, which is a bug.)

It looks like Makefile decided in favor of NIS but python chose not to
build it.