Re: tenant problem in framework can somebody (Adam) help?

[Pierre Smits <pi...@gmail.com>]

Adam,

Pleaase elaborate on what the unit test(s) should achieve? Creating new
tenants and loading the tenant instantiation with demo data? Or?

Regards,

Pierre

2012/5/7 Adam Heath <do...@brainfood.com>

> On 05/07/2012 09:18 AM, Adam Heath wrote:
> > I'll fix it when I get to work.
> >
> > ps: there are really no test cases for tenants?
>
> ps: I have a plan for testing the tenant code.  Do not start it by
> just 'cloning' the existing entitymodels for the tenant.  That will
> take *way* to long.  Instead, create a completely brand new base
> delegator, that only has one or two entity definitions in it, then add
> tenants for that.
>

Re: tenant problem in framework can somebody (Adam) help?

[Adam Heath <do...@brainfood.com>]

On 05/07/2012 03:40 PM, Pierre Smits wrote:
> Given that the tenants use the same codebase and only use a different
> delegator than main, you could say that performing unit tests as in a
> tenant delegator should get you the same results as doing them in the main
> delegator.
> 
> But, then there would always be an uncertainty that might itch...
> 
> Testing it al in a tenant environment could/would entail:
> 
>    1. create the tenant
>    2. load the demo data into the environment (db) of the tenant through
>    its delegator
>    3. trigger test suites and/or individual test in the environment of the
>    tent through its delegator
> 
> We can and could use current ant targets for items 1 and 2 mentioned above.
> But for item 3  a new ant target (in style of current testing targets in
> build.xml) should be created.

There's no reason to run *all* tests in a particular tenant.  That
would be useless extra work.  Only test what is nescessary to verify
that tenant code actually works.  Use code-coverage to help with this.

ps: You do realize who you are talking to?  Someone who isn't a newbie
with ofbiz.  The tone of your responses seem to lean in that direction.

Re: tenant problem in framework can somebody (Adam) help?

[Pierre Smits <pi...@gmail.com>]

Adam,

Given that the tenants use the same codebase and only use a different
delegator than main, you could say that performing unit tests as in a
tenant delegator should get you the same results as doing them in the main
delegator.

But, then there would always be an uncertainty that might itch...

Testing it al in a tenant environment could/would entail:

   1. create the tenant
   2. load the demo data into the environment (db) of the tenant through
   its delegator
   3. trigger test suites and/or individual test in the environment of the
   tent through its delegator

We can and could use current ant targets for items 1 and 2 mentioned above.
But for item 3  a new ant target (in style of current testing targets in
build.xml) should be created.

Then we would be covered.

Regards,

Pierre


2012/5/7 Adam Heath <do...@brainfood.com>

> On 05/07/2012 02:46 PM, Pierre Smits wrote:
> > Adam,
> >
> > Pleaase elaborate on what the unit test(s) should achieve? Creating new
> > tenants and loading the tenant instantiation with demo data? Or?
>
> All of the above, anything, something.  There are no automated test
> cases *at all* for tenant anything.
>
> Every single one of the commits I just checked in recently had full
> tests cases run against it about 3 times.  I use git-svn, rebase -i, I
> reorder commits radically as I get a feature set working.  I'll run
> all test cases against a series of commits many times.  Takes about 7
> minutes per commit.  This is why I have tools/git-rebase-runner.sh
> checked in.
>

Re: tenant problem in framework can somebody (Adam) help?

[Adam Heath <do...@brainfood.com>]

On 05/07/2012 02:46 PM, Pierre Smits wrote:
> Adam,
> 
> Pleaase elaborate on what the unit test(s) should achieve? Creating new
> tenants and loading the tenant instantiation with demo data? Or?

All of the above, anything, something.  There are no automated test
cases *at all* for tenant anything.

Every single one of the commits I just checked in recently had full
tests cases run against it about 3 times.  I use git-svn, rebase -i, I
reorder commits radically as I get a feature set working.  I'll run
all test cases against a series of commits many times.  Takes about 7
minutes per commit.  This is why I have tools/git-rebase-runner.sh
checked in.