Pádraig Brady <[email protected]> writes: > Is this gaining you anything though. > Just because the file extended in the narrow window while install(1) > was comparing doesn't really matter. > If it extends just after that we also miss it. > > So I don't think it's worth complicating the code for that.
Upon second thought, I think I agree. I was missing the forest for the trees or something. My understanding was that RHEL patched GNU 'install' with this feature to improve install times by avoiding updating the mtime of files [1]. In that case 'install' would be run by a privileged user on files that cannot be modified by others. In cases where the source and/or destination are modifiable by others, the behavior doesn't seem great, but it isn't avoidable. And 'install' is unlikely to be used on those files anyways. Perhaps it is worth a mention in the manual, if I can think of something that isn't too verbose. Collin [1] https://lists.gnu.org/archive/html/bug-coreutils/2009-01/msg00122.html
