Linux本机配置文件格式的读取效率

Question

Is there a common practice on reading linux config files in a native format efficiently? Files are handled as streams, hence I guess the standard way to lookup by parsing is in O(n*m) (Let n lines, m avg line length). Is it common practice to build search trees from config files, or has this already been implemented, by eg QSettings?

Answer 1

Why is this a problem? QSettings usually has very few kb of data, which can be loaded and parsed in a fraction of time. You don't really need to care about it. Reading kb s in a desktop is something which cannot be measured.

The recommendation is usually to create a QSettings object when needed (in stack) and use it to read when needed. This is what I see in some Qt applications. The Qt documentation mentions this is light weight. I usually maintain a global variable in my QMainWindow.