> On 2023-02-11, at 8:19 PM, Paul Eggert <egg...@cs.ucla.edu> wrote: > > On 2023-02-10 17:27, George Valkov wrote: >> I’d prefer a timestamp consistent with source. > > You can get that with 'cp -p' (or with 'cp --preserve=timestamps' etc.). So > this is not an issue of whether you can get what you prefer. It's merely an > issue about what cp's default behavior is, when the user doesn't specify > options.
Thanks! > POSIX is reasonably clear that if you use plain cp to create a file, the > file's last-modified time should be the current time. And it's not just a > question of standards conformance: lots of real-world uses of 'cp' expect > this behavior. For example, people use plain 'cp' in Makefiles, and 'make' > relies on the last-modified time being the current time. If cp changed this > behavior, it'd break a lot of builds. Ah, yes, make relays on timestamps. So we should update timestamp after clone by default. Let me know when you have a patch without CLONE_ACL and I will test it. Georgi Valkov httpstorm.com nano RTOS
