bloodhound-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Olemis Lang <ole...@gmail.com>
Subject Re: The state of trunk
Date Fri, 20 Sep 2013 18:16:53 GMT
On Fri, Sep 20, 2013 at 11:31 AM, Olemis Lang <olemis@gmail.com> wrote:

> Hi !
> :)
>
> On 9/20/13, Anze Staric <anze.staric@gmail.com> wrote:
> > Security suite is the only suite that does multiproduct testing (it
> > upgrades the environment to multiproduct). Since BEP-0010, we modify
> > ticket table to set UID as autoincrement, and increment ID manually.
> > After each test, reset_db is called, but it does not recreate the
> > tables, it just removes their content, so autoincrement field remains
> > UID.
> >
> > If "single product" tests are run afterwards, they are run in an
> > upgraded database, but since the version table has been wiped, they do
> > not know that. Sql translation unaware of the upgraded database does
> > not fill the id field in the ticket table so NULL gets inserted as
> > ticket id, causing the errors you mentioned.
> >
> > Unless anyone is aware of a simple way to recreate the tables when
> > needed, I'd go with reordering the tests.
> >
>
> All tests most (if not all) tests


i.e. most tests (if not all)


> in bloodhound_multiproduct will
> perform multi-product assertions as long as they invoke
> MultiproductTestCase._upgrade_mp class method (... in my local copy
> after applying patches for functional test suite , in /trunk this
> might still be an instance method ...)
>
>
it's a class method in trunk as well

http://bhnet.blood-hound.net/browser/bloodhound/bloodhound_multiproduct/tests/env.py?marks=224-240#L224

[...]

-- 
Regards,

Olemis - @olemislc

Mime
  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message