Shawn Bohrer wrote: > Agreed I'll send an updated patch later tonight. One additional thought > though. 2 is MATCHED_FNMATCH which worries me a little because I think > this would mean 'git clean -f *' will also remove directories (I haven't > tried though). Perhaps this should really be 3 MATCHED_EXACTLY just to > be safe. Does anyone have opinions either way? I don't have strong opinions on this since I don't use this form of the command, but still: I think that the best option would be to never remove a directory, even if given explicitly, unless -d is given. Because my gut feeling is that when a directory name is specified, it is most often meant as "clean inside the given directory", ie. as a path delimiter. Indeed, if the directory has tracked files inside of it, git clean dir and git clean dir/ have the same effect. If there are no tracked files inside, the current patch gives the path-delimiting effect on this form git clean dir/ but removes the whole directory irrespective of "-d" for this form git clean dir I think that a "honor (lack of) -d even if pathspec matches" would reduce the consequences of this particular kind of user error (by deleting too little instead of too much). -j. -- To unsubscribe from this list: send the line "unsubscribe git" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
