On Sat, 28 Dec 2019, Segher Boessenkool wrote: > > This is about extracting attributions from changelogs when unambiguous > > there, and then correcting mistakes or otherwise making minor variants > > more uniform. > > Yes, and I'm saying you probably shouldn't do that.
Extracting attributions from ChangeLog files is clearly useful to attribute changes committed by someone other than their author, and to provide more precise attributions in the cases where someone was committing under multiple email addresses to distinguish the affiliation relevant for each commit. > Note that these errors did not exist in the changelog in the commit > message, for example. Mostly the changelogs in the commit message don't give the author email at all (or if they do, it's because they've cut-and-pasted the complete ChangeLog entry from one place to another, so with the same typos). > Since people very often typo their own name (as the evidence shows), the > heuristic for deriving it should be robust against that. I think Richard's heuristic for finding cases of typos and proposing fixes for them (then to be applied automatically, after adjusting the preferred version in cases where comments such as Jakub's indicate something other than the most common version is to be preferred) looks robust to me. -- Joseph S. Myers j...@polyomino.org.uk
