I get it.  

So I will try to change the code so that it supports both the old behavior (ie 
by subclassing) and new behavior (ie by styling).

Maurice 

-----Message d'origine-----
De : Alex Harui [mailto:aha...@adobe.com] 
Envoyé : jeudi 10 octobre 2013 16:56
À : Maurice Amsellem; dev@flex.apache.org
Objet : Re: Build failed in Jenkins: flex-sdk_mustella #460 (fixed)



On 10/9/13 11:51 PM, "Maurice Amsellem" <maurice.amsel...@systar.com>
wrote:

>Understood.
>
>In this case, I' am changing the test code so that it matches the 
>existing baseline, not changing the baseline.
>So it should not have any impact.
Maybe I'm not understanding you, but IMO, the test code represents a user's 
application.  Every time you change the test code, it means that some number of 
our customers will also have to change their application code.  For sure, 
sometimes that number is 0 because the test is incorrectly written, and 
sometimes the number is small and we decide the fix is more important (and in 
the case of a security issue, it doesn't matter how many people it breaks).

If I understand this thread, some number of folks with custom skins may have to 
tweak their skins, otherwise they may notice a one pixel shift in the callout 
position.  Again, I'm not saying don't do that, but consider whether that is an 
acceptable result.  You may have the option of using a version flag to allow 
folks to get back to old behavior.

-Alex

Reply via email to