pivot-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Andrei Pozolotin <andrei.pozolo...@gmail.com>
Subject Re: Pivot unit testing
Date Wed, 29 Jun 2011 17:47:51 GMT
could also borrow some ideas form this:
http://code.google.com/javadevtools/wintester/html/index.html

-------- Original Message  --------
Subject: Re: Pivot unit testing
From: Chris Bartlett <cbartlett.x@gmail.com>
To: dev@pivot.apache.org
Date: Wed 29 Jun 2011 06:57:39 AM CDT
> Thanks for the info, Greg.
>
> I did have a *very* quick look at the SwingDemo, but I wasn't clear about
> exactly what is required of an ApplicationContext.
> My initial thought was that it might be simple to create a new
> UnitTestApplicationContext, but then the size of DesktopApplicationContext
> suggested otherwise.
>
> My dirty hack works for now, but I will create a JIRA wish for
> UnitTestApplicationContext, and try to follow up later.
>
> Chris
>
> On 29 June 2011 18:43, Greg Brown <gk_brown@verizon.net> wrote:
>
>   
>>> 1) Is there an 'official' minimum version of JUnit expected for Pivot's
>>> build?
>>>       
>> I believe it is 4.7.
>>
>>     
>>> 2a) Has there ever been a discussion on unit testing Pivot GUIs, using
>>>       
>> real
>>     
>>> Components and a real Display?
>>>
>>> 2b) If so, did the idea of a UnitTestApplicationContext come up, or is it
>>> worth considering?
>>>       
>> This is an interesting idea, as is the concept of a headless Display.
>>
>>     
>>> I hacked together a quick 'unit test Display provider' by instantiating a
>>> custom Application with DesktopApplicationContext and waiting until a
>>>       
>> static
>>     
>>> Display field was set.
>>>       
>> Take a look at the Swing demo for an example of creating a custom
>> application context.
>>
>>     
>>> 3) Are unit tests that require a Display deliberately avoided so that the
>>> build can run in headless environments?
>>>       
>> Yes, that's part of it. But you are right - there are definitely ways to
>> make this work.
>>
>> G
>>
>>
>>     
>
>   


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