[PATCH] DRAFT: Add new test which covers force user bug

[Robin Hack]

Hi Uri and Andreas.

Thank you for your feedback. I must admit that I didn't explore samba
testsuite fully.

I agree that source3/script/tests/test_smbclient_auth.sh is much
better place to live :).

However.

I don't think that add one line to
samba3.blackbox.smbclient_auth.plain would work. I really need to use
group with exactly same name as user have. I created one user and
group named: force_user.
Should I hardcode user name and group name to test-shell script?
Or extend parameters of script?

Uri, what do you mean by: "it has to run on all envs."?

Patch need rebase because new tests were added :).

Thank you and have nice day
Robin Hack

On Mon, Feb 1, 2016 at 7:17 PM, Uri Simchoni <uri at samba.org> wrote:
> Yeah, OK.
> samba3.blackbox.smbclient_auth.plain seemed like a place where adding the
> test would be just adding a line, but it won't work - it has to run on all
> envs. samba3.blackbox.valid_users is not where we want it - it should be
> with other "force user" tests eventually.
>
> The patch does need the other fixes I pointed out, and re-basing (I tried
> doing the fixes myself but due to the rebase need, I cannot apply it with
> git am and applying with "patch" would lose all the gitness of the patch :))
>
> Thanks,
> Uri.
>
>
> On 02/01/2016 12:18 PM, Andreas Schneider wrote:
>>
>> On Wednesday 27 January 2016 13:17:04 Uri Simchoni wrote:
>>>
>>> It appears to me that the essence of this new test suite is to perform a
>>> successful tree-connect to the force_user_valid_users share, using
>>> $USERNAME/$PASSWORD credentials. Can't it be just added as another test
>>> to an existing test suite that tests similar things using similar
>>> techniques? Good candidates are:
>>> - samba3.blackbox.valid_users
>>> - samba3.blackbox.smbclient_auth.plain
>>>
>>> (actually I think the two test suites should be merged, or that all
>>> "force user" test should be combined in a separate test suite, where it
>>> can be verified not only that the tree-connect succeeds, but also that
>>> the force-user works, e.g. by creating a file and doing smbcacls on it -
>>> but that can be done later. In the meantime I don't see a reason for
>>> adding another test suite)
>>
>> I think we should move all 'force user' test to a new test. Could we push
>> the
>> patch upstream and then work on merging them?
>>
>>
>>         -- andreas
>>
>