U++ forum: Welcome to the forum

Re: large ArrayCtrl [message #842 is a reply to message #840]

Sat, 04 February 2006 13:05

hojtsy
Messages: 241
Registered: January 2006
Location: Budapest, Hungary

Experienced Member

This was part of the code that is now written in a shorter form.

array.AddRowNumColumn("number");
array.AddRowNumColumn("text");
array.HeaderTab(0).SetRatio(20);
array.HeaderTab(1).SetRatio(80);
array.ColumnAt(1).SetConvert(numberToText);

I see that the current internal logic of the ArrayCtrl makes it necessary to provide two different methods ColumnAt and HeaderTab which return a different types. But it is hard to remember which method (ColumnAt or HeaderTab) and type (ArrayCtrl::Column or HeaderCtrl::Column) can be used to set/modify/query a specific property of the column. It would be better for the clients to have a single method returning a an object which could be used to manipuate all aspects of the column. I think the naming of these two methods just add to the confusion - both of them are returning a different Column type, but only one is called ColumnAt. Maybe the ArrayCtrl::Column could be extended with all the methods of HeaderCtrl::Column, and those would just forward the call to the HeaderCtrl.

During the development of this example, I found a bug: Assist was not working correctly in the inline constructor of BigArray. It only offered the code completions that are valid in the namespace scope. I believe you can reproduce this if you create a new project with my original example.

Report message to a moderator