Check to see whether any DataType we hold references to is included in the Stack (which holds all DataType instances that directly or indirectly reference this instance, including this instance itself).
Sets the units to use for the comparison. This is a little
complicated because common usage has created standards that play havoc with capitalization rules. Thus, some people will use "K" for indicating 1000's, when the SI standard calls for "k". Others have tried to introduce "K" as a multiple of 1024, but that falls down when you reach "M", since "m" is already defined as 0.001. <p>
To get around this complexity, a number of standards bodies have proposed the 2^10 standard, and at least one has adopted it. But we are still left with a populace that isn't clear on how capitalization should work. <p>
We therefore ignore capitalization as much as possible. Completely mixed case is not possible, but all upper and lower forms are accepted for all long and short forms. Since we have no need to work with the 0.001 case, this practice works here. <p>
This function translates all the long and short forms that a unit prefix can occur in and translates them into a single multiplier.
<p>Checks to make sure all settings are kosher. In this case, it
means that the size attribute has been set (to a positive value), that the multiplier has a valid setting, and that the size limit is valid. Since the latter is a calculated value, this can only fail due to a programming error. </p> <p>If a problem is detected, the setError() method is called. </p>