On Jan 10, 2006, at 4:43 AM, Ralf Stephan wrote:
Please bear with the side effects. I firmly believe that they are worth it all. A log file of the removal is available on request. All in all, a 1.5 GHz computer with 2.5" hard disk needed 12 hours full work for it.
Well... that's not quite the more important aspect. Aside from the CPU spent reducing LOC, are there other metrics you have?
A reduction in build times might be helpful, but an improvement in code quality would probably be much more important. Do you happen to have anything checking standard code metrics before and after?
Also...
One general guideline for improving #includes is to remove #include's from inside of .h files and instead count on including needed ones in the using .cpp files. Among other things, this reduces dependencies, trims hardcoding, improves flexibility and maintainability (with the last one pretty important).
Did you do any passes to address #includes in .h files before removing them from .cpp files? If not, then we can probably just work on that as a next phase.