@john_ohno wrote:
When taking over as a primary maintainer for some piece of code, I have been known to experimentally determine the number of spaces the fool who originally wrote it used for indentation and then use sed to convert the entire codebase to tabs -- sometimes producing commits in the millions of lines. Because if you are using spaces for indentation, literally nobody can tell how many spaces you are using (because using spaces for indentation is indistinguishable from having inconsistent indentation, outside of languages like python that autodetect and throw errors on inconsistent indentation), and it's extremely likely that someone later miscalculated the number of spaces you are using and introduced inconsistent indentation, even if you happened to remember.
(It doesn't help that I work with people who legitimately support both mixing tabs and spaces in indentation and using half-indents in certain places. The person who sits next to me at work prefers two-space tabs and will half-indent line continuations, hard-wrap on 40 characters, and right-justify debug prints. And prefers to use noops for comments over actual comments.)