Design of register/memory viewers

[Gunther Piez]

Am Donnerstag 01 September 2005 08:40 schrieb Vladimir Prus:

> Can you explain? The amount per line should ideally depend on window width.

If you use a memory view, it's often because you want to watch some big array 
(at least for me) which is to clumsy to handle in a watch window. If the 
array has more than on dimension, lets say int[][16], the natural data 
display width would be 16 ints, a not depending on the width of the window. 
But if no value for the data width is given (the default), make it whatever 
fits in the window.
Also it happens, that an array is sparsely filled. Maybe its an array of ints, 
but the values in the array have only a range of 0-100. So watching an array 
of bytes with every other three bytes left out would be sufficient and 
conserve some space. Or you have an int[][16], of which only the values int[]
[0..9] are used (Choosing powers of 2 as a sub index happens quite often for 
perfomance reasons). I this case you want to show 10 ints, and hide 6. Or an 
array of a struct, which is 9 bytes long and 16 byte aligned for each 
element.
One would need two more numbers to make it possible to configure this flexible 
enough, one for the number of bytes to show and one for the number of bytes 
to hide.
So for example, you could have a display width 64 bytes width, but only the 
one byte seen and every other three bytes hidden. This would display 16 bytes 
per line, eating up an amount of 48 chars width.

Gunther

-
to unsubscribe from this list send an email to kdevelop-request at kdevelop.org with the following body:
unsubscribe »your-email-address«