Kerberos raw prototol testing

Andrew Bartlett abartlet at
Mon May 3 00:33:03 UTC 2021

On Mon, 2021-05-03 at 01:16 +0200, Stefan Metzmacher wrote:
> Am 27.04.21 um 12:31 schrieb Andrew Bartlett:
> > 
> > Thanks for keeping this tree recently rebased, but how do we go
> > from
> > here?  
> > 
> > Should we just learn from the concepts and implement the narrow
> > case at
> > hand (FAST testing) and you will integrate it later, or is there a
> > better way?  How can I/we use your code?

> > There is clearly a lot of effort and value in between all the 'sq',
> > 'fixup' and reverts, but I don't know how to sift that gold out
> > properly and refine it into an 'upstream' state.

> Just look at the complete diff of that branch and
>, and work on top of that. If you're ok with it I can
> just
> squash almost everything and propose it as one large patch in order
> to
> make progress and get it to master. Or do you think it would be
> useful
> to have some of the preparation as individual commits?

Firstly, I would say I've come to like our pattern of small,
identifiable commits with single logical changes, because we learn and
can follow better when the patches tell as story, but the value
diminishes greatly for new testsuites.

In this case in particular I've checked the diff and other than of
course the pycredentials fix (thanks, obviously good!) this really is
quite fine squashed. 

I'm also happy if you keep some of the early preparation patches
distinct, eg "python/samba/tests/krb5/rfc4120: better etype dissection"
but I wouldn't spend much time beyond that. 

Please propose it squashed down and lets work to get it into master. 

I'll do a proper review but just one thing I noticed.  In:

commit ea514b738416efc7429943c7e06f7f52670cfbe9
Author: Stefan Metzmacher <metze at>
Date:   Tue Apr 14 14:05:03 2020 +0000


diff --git a/selftest/wscript b/selftest/wscript
index afb1fa936cd..3e68ec77d8b 100644
--- a/selftest/wscript
+++ b/selftest/wscript
@@ -144,6 +144,8 @@ def cmd_testonly(opt):
     if CONFIG_GET(opt, 'HAVE_MIT_KRB5_PRE_1_18'):
         env.FILTER_XFAIL += ' --expected-failures=${srcdir}/selftest/knownfail_mit_krb5_pre_1_18'
+        env.FILTER_XFAIL += ' --expected-failures=${srcdir}/selftest/knownfail_mit_krb5'
     if Options.options.FAIL_IMMEDIATELY:
         env.FILTER_XFAIL += ' --fail-immediately'

This needs to be dropped or consolidated as Gary already did the same
further down the file.

Anyway, a good example of why we need this merged: much better to all
be working on the same basis in master that than seeing duplicate work
done, we need to spend our time wisely!


Andrew Bartlett

Andrew Bartlett (he/him)
Samba Team Member (since 2001)
Samba Team Lead, Catalyst IT

Samba Development and Support, Catalyst IT - Expert Open Source

More information about the samba-technical mailing list