Hello Maxim,
thanks for having look at my patch.
Maxim Cournoyer <maxim.courno...@gmail.com> writes:
> We shouldn't hard code Python versions in the paths as it'd be too prone
> to break. You could probably make good use of the recently introduced
> search-input-directory procedure here :-).
I tried your suggestion, but unfortunately my attempt with modifying
the lines to as such:
--8<---------------cut here---------------start------------->8---
(talloc (string-append
(search-input-directory inputs
"/lib/python3.9/site-packages")))
(tdb (string-append
(search-input-directory inputs
"/lib/python3.9/site-packages")))
(python-dnspython (string-append
(search-input-directory inputs
"/lib/python3.9/site-packages")))
--8<---------------cut here---------------end--------------->8---
compiles to a package. But ‘samba-tools’ fails again.
It is not quite clear to me now why this method is better? Is hard
coding being avoid now? Should I avoid the ‘3.9’ in the search path?
Sorry that I have to ask for an explanation.
And while we are at it. I noticed that you have already replaced
heimdal with mit-krb5, which I had done personally already without
proposing it yet. Thanks for that.
Point is that I was trying to pull up an AD DC, which is why I run these
problems.
And then, can I ask you to add avahi to inputs? Since some time Samba
has disabled SMBv1, which means host discovery with NetBIOS doesn’t work
any more. Now it is best to have Avahi running alongside Samba hosts,
so that Samba can publish it’s service over multicast. No extra
configuration is needed for this.
This doesn’t help much for Windows hosts though. I’ve already packaged
wsdd [1], and have prepared a service, which I haven’t made public
either yet. I haven’t completed the documentation for info so far.
Should really finish it now.
I will open another issue for this later, and will not hijack this
thread for it.
Kind regards
Simon
[1] https://github.com/christgau/wsdd
