[Gambas-user] Wizard question
adamnt42 at ...626...
adamnt42 at ...626...
Wed Dec 2 01:30:14 CET 2015
On Wed, 2 Dec 2015 00:12:16 +0100
Tobias Boege <taboege at ...626...> wrote:
> On Wed, 02 Dec 2015, Moviga Technologies wrote:
> > Can't you just find the .children[] index and hide it?
>
> Whenever I see somebody doing this, part of me wants to commit a silly
> change (like changing the name of a control, or whatever bit of information
> their algorithm uses to find the correct child) and break their application
> with the next release.
>
> Regards,
> Tobi
>
Ha-ha, you evil child!
:-)
(But then, I could have used you on some testing contracts we had some years ago!)
On the other hand I could question the wisdom of implementing a non-cancel-able wizard? But then again, there is a case for some further study here...
One presumes that the reason for preventing a cancel is that having reached step X, the subject of the wizard may be left, at best, in an indeterminate state and at worst in an unstable state. There is a fairly good example in the IDE. The packaging wizard can be cancelled at any step, but changes made in the early steps are "committed" to the .project file. Now, to me, this is a good thing as it let's me set up some things deliberately in the .project file (such as vendor information and licensing) at the early stage of development and these will be used later on when I really want to package the project. On the other hand, if I select the "vendor-prefixing" option and then cancel, then it seems to always be set whether I want it or not.....
N.B, This is just an observation, not a criticism.
In the back of my mind there is a little voice that keeps saying, "If you are going to provide a wizard, then don't commit anything until the final step" and another that says "... but you'd better provide a way to set / configure values for the attributes covered by the wizard elsewhere!"
cheers
bruce
--
B Bruen <adamnt42 at ...3379... (sort of)>
More information about the User
mailing list