[pacman-dev] [PATCH] Rework of check-for-new-pacman-version

Xavier shiningxc at gmail.com
Thu Feb 14 17:10:03 EST 2008

Chantry Xavier wrote:
> This patch is primarily a fix for FS#9228.
> The -Sy operation was moved to its own transaction, which allows us to do
> the following steps :
> 1) refresh the database if asked
> 2) check if a new pacman version if available
> if yes, initialize a "-S pacman" transaction.
> If no, initialize the original -S or -Su transaction.
> Note that the newversion check was extended to the "-S target" operation,
> which is helpful in case of syncdb syntax change (see versioned provisions
> for example)
> Original-work-by: Nagy Gabor <ngaba at bibl.u-szeged.hu>
> Signed-off-by: Chantry Xavier <shiningxc at gmail.com>
> ---
>  src/pacman/sync.c |  129 +++++++++++++++++++++++++++--------------------------
>  1 files changed, 66 insertions(+), 63 deletions(-)

Here is a comment from Nagy in bug 9228 :
 > Well, maybe a bit off here, this check-for-new-version could be more
 > general, with simple "-S foo" can be useful imho (can be
 > enabled/disabled in pacman.conf).
 > So we could implement a check_for_new_version() function to the
 > back-end [with parameters repo (or NULL for all), and pkgname]; and
 > this could be called by the front-end before transaction start.

Dan seemed to think this was a good idea:
19:09  toofishes >> i guess i just don't like hardcoded things like this
19:10  toofishes >> and what if we (or someone else) has a libalpm 
package and a pacman package
19:10  toofishes >> and only libalpm gets updated
19:10  toofishes >> then it wouldn't stop and say "do this alone"
19:10  toofishes >> like...maybe this should be a config file thing.
19:10  toofishes >> AlonePkg or some nonsense :P

Well, if we can find a decent config option name for that, I am willing 
to implement it (by making the above patch more general).

More information about the pacman-dev mailing list