Re: tar --mtime

2025-03-03 Thread Paul Eggert
On 2025-03-03 03:54, Bruno Haible via Gnulib discussion list wrote: I don't think we have any users on HP-UX any more, for several years already. still seems to be active. It has a 64-bit Itanium 2 copy of Coreutils 9.6, released in January. Insta

Re: tar --mtime

2025-03-03 Thread Simon Josefsson via Gnulib discussion list
Bruno Haible via Gnulib discussion list writes: >> > - a POSIX file system (not the Haiku file system, which assigns slightly >> > different mtimes at random when the tarball is unpacked). >> >> There is the idea to store modification times of files in a text file >> suitable for parsing a

Re: tar --mtime

2025-03-03 Thread Bruno Haible via Gnulib discussion list
Simon Josefsson wrote: > > - a 'make' program that considers equal time stamps as "up-to-date" > > (a condition fulfilled by all 'make' programs except HP-UX make), > > Yes you are right. But for users, advice should be practical -- is > there any reasonable way to solve the problem on HP-U

Re: tar --mtime

2025-03-03 Thread Simon Josefsson via Gnulib discussion list
Bruno Haible writes: > Simon Josefsson wrote: >> Personally I use tar --mtime to the latest git commit time instead of >> the complexity of the vc-mtime approach. I've done two software >> releases with this approach (libtasn1, inetutils) and no reports of >>

Re: tar --mtime

2025-03-03 Thread Bruno Haible via Gnulib discussion list
Simon Josefsson wrote: > Personally I use tar --mtime to the latest git commit time instead of > the complexity of the vc-mtime approach. I've done two software > releases with this approach (libtasn1, inetutils) and no reports of > problems so far, but I'm open to the i