My only thought is whether the new class creation wizard is actually API or not. It’s only API if people use it that way. Otherwise, the change shouldn’t affect
anyone. But I’m not clear on the details of this case.
BTW, my moto is “Do the right thing”, blindly following process isn’t always the right thing, especially if it fixes things without breaking things.
Doug.
From: cdt-dev-bounces@xxxxxxxxxxx [mailto:cdt-dev-bounces@xxxxxxxxxxx]
On Behalf Of James Blackburn
Sent: Friday, May 13, 2011 2:49 AM
To: CDT General developers list.
Subject: Re: [cdt-dev] Backward incompatible changes in NewClassCreationWizardPage.java
On 13 May 2011 03:49, Sergey Prigogin <eclipse.sprigogin@xxxxxxxxx> wrote:
NewClassCreationWizardPage is an interesting class. It doesn't have any privacy but lives in a externally visible package. I made few improvements to the New Class wizard that inevitably resulted in backward incompatible API changes. If
your code extends NewClassCreationWizardPage, you may want to check that these changes didn't break your code.
I don't want to be a stick in the mud, but I can't see how we can do this at this stage:
You say that you've made backwards incompatible change. However there's no Bug # on the commit and no indication of what the breakage is, and what the risk is for clients / extenders. Who's likely to be relying on this API?
I've been holding back on any non-bug-fix changes until Juno. I'm not saying everyone should do this, but I think we need to be _very_ careful about breaking API at this stage -- especially as it's clear integrators have been testing the
M7 build. They may well miss an API change at this stage.
_______________________________________________
cdt-dev mailing list
cdt-dev@xxxxxxxxxxx
https://dev.eclipse.org/mailman/listinfo/cdt-dev
