Comment # 9 on bug 1036283 from Noel Power

(In reply to Noel Power from comment #7)
> (In reply to Peter Simons from comment #6)
> > (In reply to Panagiotis Georgiadis from comment #5)
> > > Do we have a reproducer? In case of 'yes', could you please provide the
> > > required steps?
> > 
> > The respective first message of
> > https://bugzilla.samba.org/show_bug.cgi?id=12721 and
> > https://bugzilla.redhat.com/show_bug.cgi?id=1436145 describe how users
> > produced the issue on their servers. Unfortunately, the information given is
> > not very detailed and I'm not sure it's obvious to everyone how to translate
> > into concrete steps. It's certainly not obvious to me. Maybe the package
> > maintainer can help?
> 
> No, it's not clear to me either, however the patches ported contain a test
> (only runnable from a dev build) and I ran that test succsessfully. I'll
> have a play with it again tomorrow to see if I can get a simple reporducer

sorry for the delay, I had difficulty getting the reproducer to work due to a
typo which resulted in much time wasting :-(.
It's actually simple enough to test this 

without patch/fix but with up to date samba containing fix for CVE-2017-2619;

  a) set up a simple standalone fileserver
  b) create a local user to access the share (e.g. smbpasswd -a localuser)
  c) create a share, create a subdirectiory (eg. dir1) in the share (make sure
the permissions are such that other users can create directories)
  d) add 'follow symlinks = now' param (restart smbd)
  e) try and create a sub dir via

smbclient -Ulocaluser%passwd //localhost/sharename -c 'mkdir dir1\dir2'

it should fail with 'NT_STATUS_NOT_SUPPORTED making remote directory dir1\dir2'

update with new samba with regression fix, repeat step 'e)' above and 

this should succeed