[Lazarus] Opinions required on what the ReadOnly property should do for LCL controls.

[Inoussa OUEDRAOGO]

2008/2/22, Inoussa OUEDRAOGO <inoussa12 at gmail.com>:
> 2008/2/22, John <jszcmpr at netspace.net.au>:
>
> >
>  >  svaa wrote:
>  >  Hello:
>  >
>  > For me there is no doubt. When I set a control as readonly, I expect
>  > that the user won't be able to change it at all.
>  >
>  > In fact, years ago I hacked rxlib to avoid pop a list when they where
>  > readonly. Particulary if they are attached to a database, it's confusing
>  > for the user to change it clicking the button but updating nothing. The
>  > downside is that they won't be able to show the calendar, but I think it
>  > is better than confusing user. If the popped calendar could show clearly
>  > that is not changeable, it would be the best, but a low priority
>  > enhacement. By the way, when I say TDateTimeEdit, I mean any control
>  > with a combo etc.
>  >
>  > Anyhow, if you want people see that a control is not usable, it is
>  > better to assign enable:=false. The pity of disabled controls it that
>  > they are not easily readable.
>  >
>  > Sometimes I think that there should be three colors to inform the user
>  > of the state of the control: enabled, disabled and readonly. Long time
>  > ago I played with the idea of changing the Dataware controls to have a
>  > yellow background when they where not editable (because they were
>  > readonly or because the TDatasource was not autoedit or because the
>  > underlaying table was readonly.)
>  >
>  > Just my 0.25 cents.
>  >
>  > Reagards
>  > Santiago A.
>  >
>  >  +1
>  >
>  >  I usually change the colour or background colour of a control if it
>  > *always* readonly, it would be great to have an easy way to do it.  I
>  > probably tend to disable a control if its mode changes, but it can still be
>  > useful to distinguish between "valid, but read only" and "invalid in this
>  > context".
>  >
>  >  I think the difference between the two views depends on whether you see the
>  > TDateEdit as two controls ( in which case I agree with Vincent) or a single
>  > control, in which case I agree with Gerard.
>  >
>  >  If it were two controls, you make the edit box readonly (to keystrokes),
>  > but the button is changing it programatically.  But if it is a single
>  > control, readonly = readonly = readonly - no user changes of any kind.
>  >
>  >  As it appears to only a single control - you can't disable the button
>  > without disabling the whole control - I think I have to agree with Gerard
>  >
>  >  cheers,
>  >  John Sunderland
>  >
>  >
>  >
>  >
>  >
>  >  There's a discussion going on about TDateTimeEdit.ReadOnly in Mantis
>  > report 10861.
>  > It's not a technical discussion, it's more about what ReadOnly should
>  > and shouldn't do.
>  >
>  > See. http://bugs.freepascal.org/view.php?id=10861
>  >
>  > The list is probably a better place to discuss this, so please read
>  > the report and give your opinion here.
>
>
> IMHO, whenever "ReadOnly=True", user should not be able to modify the
>  value, be it by
>   the text edit or by the button. I use code like
>  [code]
>  for i := 0 to Pred(ContrloCount) do begin
>   ct := Control[i];
>   if IsPublishedProp(ct,'ReadOnly') then
>     SetProp(ct,'ReadnOnly',False);
>  end;
>  [/code]
>  to be sure the user cannot modify any value.
>  Example  : On a Edit form, I want the user to be able to modify
>  certain properties ( "color of your car") depending on the value of
>  other properties ( "Own a car" ).
So in edit style : "Readonly =False", but "Enable" depends on the
state of the object.
In view style, "ReadOnly = True", the user can just view the item;

Best regards.
-- 
Inoussa O.