On Mon, 2010-02-01 at 21:08 +1030, Ty John wrote:
On Mon, 1 Feb 2010 15:08:33 +0800 Ray Rashif <schivmeister@gmail.com> wrote:
On 01/02/2010, fons@kokkinizita.net <fons@kokkinizita.net> wrote:
On Sun, Jan 31, 2010 at 11:55:57PM +0100, Giovanni Scafora wrote:
2010/1/31, fons@kokkinizita.net <fons@kokkinizita.net>:
that means that cdrkit has been renamed to cdrtools ? :-)
Of course, it means that the software has benn renamed or replaced by another one.
So it can mean two very different things.
Which means that the exact background of the question 'Replace kernel-headers by api-headers ?' is unclear, and that the OP had good reason to ask what it meant. Pacman did *not* tell him this was just a rename.
Oh nono, $replaces isn't used like that. When for instance you have deleted a package and brought in a new one with a different name, often due to a name change (upstream or not), you need to make sure pacman will know and seamlessly "update" to the new package. Sometimes, projects go defunct and forks become active.
Asking the user to answer the question resolves one big thing:
1) He will not complain later; he won't be freaked out when he finds one of his packages is missing and/or the system has something he can't recall installing.
-- GPG/PGP ID: B42DDCAD
I understand what you are saying but it comes back to KISS ideals. The Arch user should know exactly what's happening to their system and not just let everything happen automagically.
Does that preclude informing them? Not everyone is subscribed to [arch-dev-public], and that's probably the only place I heard of the switch from kernel-headers to linux-api-headers before it actually happened, both in [testing] and [core]. I see a distinction between 'knowing what's happening to your system' and 'having to find out the hard way what needs changing'.