Matthew Davison m.davison@virgin.net wrote:
During my work to clean up the Unicode -> ascii cross-calls I have noticed some A:W pairs where both are implemented in essentially the same way. Would it be sensible in these case to make the A function into a call to the W one, to reduce code size, or would I be wasting my time.
I have noticed the same, and looking through it I have for myself come to the conclusion that if the function is rather simple, such as only calling a few A/W specific WinAPIs there is no gain in trying to make the A function call the W function other than an additional possible overhead in execution. Just make sure both functions remain at the same position in the source so that a prospective developer who patches one of the functions can't miss the other one. I for myself handle the rule: If the 'A calls W function with Unicode translation' results in more lines of source code than the direct A implementation I won't spend time to try to convert them.
If the function however is a little more complex there is always the issue of maintaining two different functions in parallel and then making the A function call the W function is the most sensible approach indeed.
Rolf Kalbermatter