mx2927 at gmail.com
Mon Apr 1 18:34:27 UTC 2019
Thanks for the reply,
Besides the problem with
Good to hear that selftest is actively used, then do I understand it
right that 'make test' should succeed?
My bigger problem is that it failed with lots of errors.
This must be a problem with my build, then, but since this is a fresh
tarball I am a bit puzzled.
I would appreciate some hint on where to start looking for any mistake.
All run as the ordinary user - I understand from the 2012 talk that
'root' privileges should not be needed. Do I get this wrong?
Thanks in advance,
On 4/1/2019 8:03 PM, Andrew Bartlett wrote:
> On Mon, 2019-04-01 at 17:08 +0200, Manfred via samba wrote:
>> Hi all,
>> I just downloaded a fresh samba-4.10.0 tarball to verify the result of
>> 'make test'
>> What I did is:
>> ./configure --enable-selftest
>> make test
>> All run as the ordinary user.
>> I encountered a build problem with 'make' in
>> due to the following conditional being placed /before/ #include "includes.h"
>> #if defined(DEVELOPER) || defined(ENABLE_SELFTEST)
>> I fixed this by moving the conditional after the #include, and the build
>> completed fine.
>> This is the only change I made in the code.
>> 'make test' failed miserably with tons of errors.
>> The presentation from Andrew Bartlett says that selftest should be
>> enforced as part of the regular release workflow, which is great, but
>> since that talk is from 2012, I wonder: what is the current status of
>> selftest ?
> This looks like https://bugzilla.samba.org/show_bug.cgi?id=13854
> This should be in a new Samba 4.10.x shortly.
> I don't know why this only seems to happen on some platforms or what
> exactly is going on, but yes, selftest and CI is absolutely critical to
> our workflow.
> Andrew Bartlett
More information about the samba