On 5/2/12 2:59 PM, Alexey Loubyansky wrote:
On 05/02/2012 07:19 PM, Brian Stansberry wrote:
> On 5/2/12 10:47 AM, Alexey Loubyansky wrote:
>> I'll continue looking into minimizing and removing the CLIWrapper in
>> favour of the CLI public API.
>>
>
> Good. We need to rationalize things in the testsuite a bit. E.g. once we
> adequately test that the CLI public API provides the same input to and
> output from the core CLI as reading/writing from a separate process'
> stdio does, there no need for other tests to use a separate CLI process.
There is no duplication of parsing or DMR composition code, it's the
same code that is used internally and now exposed through the public api.
Yep.
> Similarly, if we establish that the CLI parses strings
representing
> low-level ops into a proper ModelNode structure for use in the native
> management API, then there's no need to test a bunch of low level ops
> using both the native management API and the CLI. Prove in some tests
> that both mechanisms work in general, and then if the real goal of a
> test is to test server side behavior of a particular op (e.g. a
> datasource add), pick either the CLI public API or a native management
> op to test that, not both.
The tests are already there.
Yes, I knew that. ;)
I wasn't trying to lay out a set of steps we need to take, do A then B,
then we can do C. I just want to state publicly that if A and B are
adequately tested, there's no reason to incur major overhead to test
them again in a of tests of C, D, E, ..Z. So people will understand why
things are changing when we change them.
BTW, I want to be clear that the CLI public API is fairly new, so what
I'm talking about here is taking advantage of something new to improve
things. I'm not complaining about how things were done in the past.
Alexey
>
>> Alexey
>>
>> On 04/28/2012 12:10 PM, Darran Lofthouse wrote:
>>> For me this has been happening after a failed OSGI test that looks as if
>>> it may use the CLI as well so not sure if that is related.
>>>
>>> Regards,
>>> Darran Lofthouse.
>>>
>>>
>>> On 04/28/2012 10:14 AM, Alexey Loubyansky wrote:
>>>> It doesn't seem to be a CLI testsuite related issue.
>>>> I simply removed all the CLI tests (cli local and the main testsuite),
>>>> './build.sh clean' and './build.sh' and I still see it.
Not after every
>>>> run though.
>>>>
>>>> Alexey
>>>>
>>>> On 04/27/2012 08:01 PM, Brian Stansberry wrote:
>>>>> A number of people are reporting that after building the current
>>>>> upstream master and running the testsuite, the terminal used to do
that
>>>>> no longer echos characters you type back to the screen.
>>>>>
>>>>> Just wanted to let you know we're aware of the issue and are
working it,
>>>>> with a pretty good idea as to the root cause.
>>>>>
>>>>> Some solutions people have used to get a normally functioning
terminal
>>>>> after this happens:
>>>>>
>>>>> 1) The Reset command from the Terminal menu.
>>>>> 2) Running 'stty sane' from the command line
>>>>> 3) Open a new terminal. ;)
>>>>>
>>>> _______________________________________________
>>>> jboss-as7-dev mailing list
>>>> jboss-as7-dev(a)lists.jboss.org
>>>>
https://lists.jboss.org/mailman/listinfo/jboss-as7-dev
>>> _______________________________________________
>>> jboss-as7-dev mailing list
>>> jboss-as7-dev(a)lists.jboss.org
>>>
https://lists.jboss.org/mailman/listinfo/jboss-as7-dev
>> _______________________________________________
>> jboss-as7-dev mailing list
>> jboss-as7-dev(a)lists.jboss.org
>>
https://lists.jboss.org/mailman/listinfo/jboss-as7-dev
>
>
_______________________________________________
jboss-as7-dev mailing list
jboss-as7-dev(a)lists.jboss.org
https://lists.jboss.org/mailman/listinfo/jboss-as7-dev
--
Brian Stansberry
Principal Software Engineer
JBoss by Red Hat