Made "Feedback" inherit from parent if empty

[richardwilliamson]

Not fully there but shows the principle..

Only does the FeedbackPath for now, basically if that’s empty then it pulls the value from the Grid (which is now enabled). Couple of issues though..

• Haven’t looked at persistence
• Probably need to split the Getter on feedbackPath() so that it doesn’t change the local value when editPanel is loaded (as that also calls GetFeedbackPath()
• Haven’t looked at wildcards yet, but am thinking that you could have %c = column, %r = row, %n = cell number

[richardwilliamson]

Also, not sure but I think the pairing may need updating once the parent is changed as I think the messages are cached.

Let me know what you think - I'm happy to potter on with this bit if it seems useful!

[MizPlusPlus]

Oh, this is a fantastic idea! Will review the changes when I get some time...

[richardwilliamson]

Not sure if you've seen my various commits but I think I have got this feature pretty much there (if you like it!) A couple of notes/bugs/todos..

• Currently, in order to update the (non OSC) label with wildcards, this is checked every time the paint event is called as I don't really understand the whole QT slots thing - I think there would be a better way to do this when the page up/down is changed..
• Currently the Label is inherited, however this isn't great as it is unlikely (I think) that you would want a grid's label on all the widgets within it - there probably needs to be an extra field in the editGrid
• Currently the up/down buttons appear on every widget, I think that something like the command widget wouldn't want them - with the other ones I suppose the question is should they always be there, or should the user be able to turn on/off (or, they could only appear if the user has used the {px} wildcard?
• There is a worry that page up/down could confuse the user into thinking that they can set a manual value to every widget across all the pages.. That is probably just a documentation thing

I have only done basic testing, so there probably are some bugs!

Let me know any thoughts..

[richardwilliamson]

Oh, for reference these are the wildcards

     //{r} = the row in the grid (0 based)
    //{r1} = the row in the grid (1 based)
    //{c} = the column in the grid (0 based)
    //{c1} = the column in the grid (1 based)
    //{n} = the number of the item in the grid (0 based)
    //{n1} = the number of the item in the grid (1 based)

    //if any of the above preceded by p (ie{pr}) then the offset version is used

I wonder if the default should be the 1 based option, as this is probably more logical to the user, and there could be an {x0} for each one which sets it to 0 based?