Samba Tested Changes Policy?

[Rowland Penny]

On Fri, 28 Apr 2017 08:33:45 +1200
Andrew Bartlett via samba-technical <samba-technical at lists.samba.org>
wrote:

> G'Day,
> 
> First, I do wish to express my hearty appreciation to those who built
> the infrastructure that allows me to even suggest a statement like the
> below.  To those who built, revamped and extended whole frameworks to
> ensure that software testing is even possible, I salute you!
> 
> I recently went to find our policy on testing code, and could only
> find "test your code, run make test" on our Contribute page on the
> wiki.
> 
> Therefore, I propose:
> 
> -
> 
> "Untested code is broken code"
> 
> Therefore all changes made to Samba should include either a specific
> automated test, or be clearly covered by an existing testsuite.  
> 
> An automated test is one that is called from our autobuild.py script.
> 
> Exceptions to this rule should be requested explicitly on samba-
> technical, with justification. 
> 
> -
> 
> I propose that we include it in our Contribute page on our wiki, and
> in the source tree in README.code-testing-policy.
> 
> This is what most of us apply most of the time anyway, and so I do
> realise that this is hard work!  
> 
> 'Clearly correct' patches still need tests, perhaps even more than
> others, and sometimes we just don't have the infrastructure to make
> such a test easy.  I particularly admire the work of those who when
> presented with this challenge go away and build more test
> infrastructure!
> 
> I also expect we will issue a lot of exceptions, and on a code-base
> our size that is OK.  
> 
> What do folks think? 
> 
> Thanks,
> 
> Andrew Bartlett
> 

I have patches for 'samba-user create' that will obtain the NETBios
domain name automatically, I haven't sent them to list because I cannot
get a test to work because NONE of the test environments seem to load
ypServ30.ldif. Do I attempt to modify the perl scripts (something I
know nothing about) and possibly break something else or what ?

Can anybody explain why this was never noticed before ? I supposed this
means that ypServ30.ldif is broken ;-)

Rowland