Dan McGee wrote: > On Fri, May 16, 2008 at 8:06 AM, Allan McRae <[EMAIL PROTECTED]> wrote: > >> 1. makepkg - Reduces the missing arch error to a warning when only >> generating intergity checks (-g or --geninteg flag) >> > Probably smart, this seems good. > > >> 2. libalpm - remove unused handle->uid from pmhandle_t. The need to >> check permissions should be determined by the frontend (and is in pacman). >> > Hmm- I almost think the backend should do something (such as verifying > we have read/write perms on the relevant dbs if necessary), but you > are right, it is unused now so it should go. Any headers that can be > dropped in handle.c because of this change? > >
Probably. I'll look into it. >> 3. libalpm - fix comment on noextract in pmhandle_t. >> > So we definitely use this and it still works? Heh. We might need to > beef up pactests in this area. > Well, a quick search showed it is still used in the libalpm code. The option is documented in the pacman.conf man page so I assume it still works! > >> 4. pacman - only ask for removal confirmation when the recusre or >> cascade options add packages to the removal list >> > I thought this came up a few months ago, and I believe I at least > thought it wasn't the greatest of ideas. I wanted behavior to be > consistent with the option, as some people might use these flags as "I > know it will ask me for confirmation" step. I know it would confuse me > (and scripts) if it asked sometimes and other times did not, and it is > not immediately apparent to the end user why this is happening. > Fair enough. I just found it annoying to be asked when no additional packages were flagged. I was thinking this was consistent with the -R option no longer asking for confirmation. Anyway, it is a small annoyance and I will get over it. _______________________________________________ pacman-dev mailing list [email protected] http://archlinux.org/mailman/listinfo/pacman-dev
