Re: request for testing AWT/SWING

[Paul Jenner]

Hi Andreas, list.

On Wed, 2005-10-05 at 22:08 +0200, Andreas Tobler wrote:
> may I ask for a minimal test sequence when you dear AWT/SWING coders 
> submit/commit a patch for AWT/SWING?

Or more automated and requiring less manual effort, a periodic
regression check against mauve?

For example daily mauve regression run gave the following on 30/09/2005:

Regressions:
+FAIL: gnu.testlet.javax.swing.AbstractButton.constructor (number 2)
+FAIL: gnu.testlet.javax.swing.AbstractButton.init: AbstractButton.text
is "" per default (number 1)


followed on 04/10/2005:

Fixes:
-FAIL: gnu.testlet.javax.swing.AbstractButton.constructor (number 2)
-FAIL: gnu.testlet.javax.swing.AbstractButton.init: AbstractButton.text
is "" per default (number 1)

(Picked one instance at random as example - no other reason for picking
this above any other recent regression and subsequent fix)

This shows a fix went in quickly and the nature of active development is
(sometimes necessary) breakage followed by better implementation to fix
before next stable release.

However I wonder if cross reference against daily mauve regressions
would help identify and fix these more quickly?

Paul

-- 
Paul Jenner <address@hidden>