>devblog_

There are thousands of classes, structure, functions and constants in the Windows library and it’s often hard to remember the right spelling. Here’s a simple trick to use the usertype.dat as a very simple spell checker.

Usertype.dat?
First, usertype.dat is a file which can be configured in the MS IDE (see help file, because it differs with the versions). Simply activate it and use my usertype.dat as a start. It contains almost 4000 key words from the libraries.

When you now type one of these key words right, it will be displayed in a different color (I prefer gray – to emphasize that I don’t have to pay attention here). Here’s how it can look like.

Example:

Note that all the common things like HICON, LoadImage or LR_DEFAULTCOLOR are written in gray because they’re spelled correct. And other words (variables in this case) like hIcon and IconId1 a printed in white because they’re not predefined. If I had written Load_Image (instead of LoadImage), it would have been printed in white, too, and I know that it might be incorrect.

When it comes to sorting, QuickSort is an all-time-favourite (so I though) and I chose it for implementing a sortable list with data. In some cases (when there were on few changes), it didn’t perform well, to I switched to an old-fashioned bubble-sort in such cases.

What I had to learn now, is that if the data that has to be sorted, is arranged in the right / unfortunate way, then QuickSort can easily crash your program by consuming excessively stack space and finally causing a stack overflow. You won’t need much data for this. 10,000 items is enough to perform terribly slow and finally crashing the software.

I switched to IntroSort, which seams to work better (so far) and (according to Wikipedia) behaves better in worst-scenarios than QuickSort. We’ll see. Sample implementations for IntroSort are rare, but I found one linked in this blog.